AIRO IDLab Ghent University - imec

2025

’Values in social robots: Implementing inclusive, value-aware human-robot interactions

Abbo, Giulio Antonio, and Belpaeme, Tony

In HRI2025, the ACM/IEEE International Conference on Human-Robot Interaction 2025

Biblio
Cauliflower centre detection and 3-dimensional tracking for robotic intrarow weeding

Willekens, Axel, Callens, Bert, wyffels, Francis, Pieters, Jan, and Cool, Simon R.

PRECISION AGRICULTURE 2025

Abs Biblio

Mechanical weeding is an important part of integrated weed management. It destroys weeds between (interrow) and in (intrarow) crop rows. Preventing crop damage requires precise detection and tracking of the plants. In this work, a detection and tracking algorithm was developed and integrated on an intrarow hoeing prototype. The algorithm was developed and validated on 12 rows of 950 cauliflower plants. Therefore, a methodology was provided to automatically generate a label based on the crop plants’ Global Navigation Satellite System (GNSS) position during data collection with a robot platform. A CenterNet architecture was adjusted for plant centre detection by comparing different encoder networks and selecting the optimal hyperparameters. The monocular camera projection error of the plant centre detections in pixel to 3D coordinates was evaluated and used in a position- and velocity-based tracking algorithm to determine the timing for intrarow hoeing knife actuation. A dataset of 53k labelled images was created. The best CenterNet model resulted in an F1 score on the test set of 0.986 for detecting cauliflower centres. The position tracking had an average variation of 1.62 cm. Velocity tracking had a standard deviation of 0.008 with respect to the robot’s operational target velocity. Overall, the entire integration showed effective actuation of the prototype in field conditions. Only one false positive detection occurred during operation in two test rows of 135 cauliflowers.
Conveying Emotions to Robots Through Touch and Sound

Ren, Qiaoqiao, Proesmans, Remko, Bossuyt, Frederick, Vanfleteren, Jan, wyffels, Francis, and Belpaeme, Tony

2025

Biblio
Enabling high-throughput quantitative wood anatomy through a dedicated pipeline

Bulcke, Jan, Verschuren, Louis, De Blaere, Ruben, Vansuyt, Simon, Dekegeleer, Maxime, Kibleur, Pierre, Pieters, Olivier, De Mil, Tom, Hubau, Wannes, Beeckman, Hans, Van Acker, Joris, and wyffels, Francis

PLANT METHODS 2025

Abs Biblio

Throughout their lifetime, trees store valuable environmental information within their wood. Unlocking this information requires quantitative analysis, in most cases of the surface of wood. The conventional pathway for high-resolution digitization of wood surfaces and segmentation of wood features requires several manual and time consuming steps. We present a semi-automated high-throughput pipeline for sample preparation, gigapixel imaging, and analysis of the anatomy of the end-grain surfaces of discs and increment cores. The pipeline consists of a collaborative robot (Cobot) with sander for surface preparation, a custom-built open-source robot for gigapixel imaging (Gigapixel Woodbot), and a Python routine for deep-learning analysis of gigapixel images. The robotic sander allows to obtain high-quality surfaces with minimal sanding or polishing artefacts. It is designed for precise and consistent sanding and polishing of wood surfaces, revealing detailed wood anatomical structures by applying consecutively finer grits of sandpaper. Multiple samples can be processed autonomously at once. The custom-built open-source Gigapixel Woodbot is a modular imaging system that enables automated scanning of large wood surfaces. The frame of the robot is a CNC (Computer Numerical Control) machine to position a camera above the objects. Images are taken at different focus points, with a small overlap between consecutive images in the X-Y plane, and merged by mosaic stitching, into a gigapixel image. Multiple scans can be initiated through the graphical application, allowing the system to autonomously image several objects and large surfaces. Finally, a Python routine using a trained YOLOv8 deep learning network allows for fully automated analysis of the gigapixel images, here shown as a proof-of-concept for the quantification of vessels and rays on full disc surfaces and increment cores. We present fully digitized beech discs of 30–35 cm diameter at a resolution of 2.25 μm, for which we automatically quantified the number of vessels (up to 13 million) and rays. We showcase the same process for five 30 cm length beech increment cores also digitized at a resolution of 2.25 μm, and generated pith-to-bark profiles of vessel density. This pipeline allows researchers to perform high-detail analysis of anatomical features on large surfaces, test fundamental hypotheses in ecophysiology, ecology, dendroclimatology, and many more with sufficient sample replication.
Insights for robotic cloth manipulation : a comprehensive analysis of a competition-winning system

De Gusseme, Victor-Louis, Proesmans, Remko, Lips, Thomas, Verleysen, Andreas, and wyffels, Francis

INTERNATIONAL JOURNAL OF ADVANCED ROBOTIC SYSTEMS 2025

Abs Biblio

Robotic cloth manipulation will be important for assistive robots. To thoroughly evaluate progress in this field, a Cloth Manipulation and Perception Competition was organised at IROS 2022 and ICRA 2023. In this article, we present the system that won the folding track at IROS 2022 and the folding and unfolding tracks at ICRA 2023. By combining visual and tactile information with engineered motions, we built a system that can generalise to a range of patterned towels made from various materials, as required for the competition. We describe our system and its limitations, which we relate to future work with the goal of creating systems that can deal with any cloth, robot, or surface.
Large Language Models Cover for Speech Recognition Mistakes: Evaluating Conversational AI for Second Language Learners

Verhelst, Eva, and Belpaeme, Tony

In Proceedings of the 2025 ACM/IEEE International Conference on Human-Robot Interaction 2025

Abs Biblio

Automatic Speech Recognition (ASR) technology has been reported to reach near-human performance in recent years, yet it continues to struggle with atypical speakers, particularly second language learners. This limitation has hindered progress in leveraging social robots for second language education, a field with significant promise. Recent advancements in Large Language Models (LLMs), which demonstrate capabilities in context understanding, common sense reasoning, and pragmatics, offer a potential solution by compensating for transcription errors introduced by ASR. This study examines whether ASR combined with an LLM can produce flowing conversation. Particularly, we look at its application in learning French as a second language by Dutch-speaking students. Through task-based interactions, where successful task completion depends on the accurate interpretation of user speech, the study evaluates the impact of LLMs on conversational outcomes. Results confirm that the performance of ASR degrades significantly for both speakers with limited proficiency and a non-English language. Nonetheless, LLMs demonstrate the ability to interpret context and sustain meaningful conversations despite suboptimal ASR outputs, highlighting a promising path forward for the integration of these technologies in second-language education.
Machine translation from signed to spoken languages : state of the art and challenges (01 Apr, 10.1007/s10209-023-00992-1, 2023)

De Coster, Mathieu, Shterionov, Dimitar, Van Herreweghe, Mieke, and Dambre, Joni

2025

Biblio
Teachers’ computational thinking content knowledge : development of a measurement instrument

Monteyne, Sara, Struyve, Charlotte, Gesquière, Natacha, Neutens, Tom, wyffels, Francis, Braak, Johan, and Aesaert, Koen

COMPUTERS & EDUCATION 2025

Abs Biblio

Computational thinking has become an integral component of curricula worldwide, necessitating teachers to develop this competence in their students. To effectively meet these curricular requirements, teachers themselves need a solid foundation of computational thinking content knowledge, which refers to the understanding and skills they possess in this area. However, despite widespread recognition of this need, few studies have rigorously examined teachers’ content knowledge in this domain. Addressing this gap requires the development of high-quality measurement tools. This study details the development of an instrument, created as part of the International Computer and Information Literacy Study (ICILS) 2023 in Flanders, to measure lower secondary school teachers’ computational thinking content knowledge in a valid and reliable way. The article first outlines the construction process of the instrument, which involved close collaboration with experts in the field and drew upon the framework of Fraillon and colleagues (2023). Following this, the instrument’s psychometric properties are presented, which include both item-level and overall instrument characteristics. These properties were evaluated using data from a sample of 352 participants, applying both Classical Test Theory and Item Response Theory. The final tool consists of 16 multiple-choice and short constructed response questions. The results indicate favorable item and overall instrument characteristics, thereby affirming its potential to measure the intended construct in a valid and accurate way.
Touched by ChatGPT: Using an LLM to drive affective tactile interaction

Ren, Qiaoqiao, and Belpaeme, Tony

In HRI2025, the ACM/IEEE International Conference on Human-Robot Interaction 2025

Biblio
Trained YOLOv8 model and traning data accompanying the paper "Enabling high-throughput quantitative wood anatomy through a dedicated pipeline"

Bulcke, Jan, Verschuren, Louis, and wyffels, Francis

2025

Abs Biblio

Trained YOLOv8 model for vessel and ray segmentation on (gigapixel) RGB TIFF images of beech. Training and annotation data in COCO format are included. This model is needed for running the source code of which releases can be found on https://doi.org/10.5281/zenodo.14637855. The full images of the increment cores can be found on https://doi.org/10.5281/zenodo.14627909. The full images of the disks (see paper for more details) can be found on https://doi.org/10.6019/S-BIAD1574. This is part of an entire sample preparation, imaging and analysis pipeline available in the paper "Enabling high-throughput quantitative wood anatomy through a dedicated pipeline" by Van den Bulcke and co-authors: https://doi.org/10.1186/s13007-025-01330-7. Cite our paper (when accepted) when using these data and/or model.

2024

A photonics perspective on computing with physical substrates

Abreu, S., Boikov, I., Goldmann, M., Jonuzi, T., Lupo, A., Masaad, Sarah, Nguyen, L., Picco, E., Pourcel, G., Skalli, A., Talandier, L., Vettelschoss, Benedikt, Vlieg, E.A., Argyris, A., Bienstman, Peter, Brunner, D., Dambre, Joni, Daudet, L., Domenech, J.D., Fischer, I., Horst, F., Massar, S., Mirasso, C.R., Offrein, B.J., Rossi, A., Soriano, M.C., Sygletos, S., and Turitsyn, S.K.

REVIEWS IN PHYSICS 2024

Abs Biblio

We provide a perspective on the fundamental relationship between physics and computation, exploring the conditions under which a physical system can be harnessed for computation and the practical means to achieve this. Unlike traditional digital computers that impose discreteness on continuous substrates, unconventional computing embraces the inherent properties of physical systems. Exploring simultaneously the intricacies of physical implementations and applied computational paradigms, we discuss the interdisciplinary developments of unconventional computing. Here, we focus on the potential of photonic substrates for unconventional computing, implementing artificial neural networks to solve data-driven machine learning tasks. Several photonic neural network implementations are discussed, highlighting their potential advantages over electronic counterparts in terms of speed and energy efficiency. Finally, we address the challenges of achieving learning and programmability within physical substrates, outlining key strategies for future research.
Adapting multimodal foundation models for text-to-3D retrieval with domain-specific vocabulary

Wiedler, Jensen, Herrewegen, Jarne, Tourwé, Tom, Demeester, Thomas, and wyffels, Francis

In CV4Metaverse 2024, 3rd Computer Vision for Metaverse Workshop, Abstracts 2024

Abs Biblio

This report addresses the challenge of applying multimodal foundation models such as Uni3D for text-to-shape retrieval in a domain-specific context. While the recent multimodal models show strong performance on zero-shot 3D classification tasks, we found that their out-of-the-box performance on our dental manufacturing dataset is underwhelming, achieving only 13.21% zero-shot test accuracy. With appropriate fine-tuning however, we managed to improve the text-to-3D classification performance to 86.02% accuracy – a fine demonstration for the transfer learning capabilities of recent multimodal foundation models.
Adaptive second language tutoring using generative AI and a social robot

Verhelst, Eva, Janssens, Ruben, Demeester, Thomas, Belpaeme, Tony, and Assoc Computing Machinery, [missing]

In COMPANION OF THE 2024 ACM/IEEE INTERNATIONAL CONFERENCE ON HUMAN-ROBOT INTERACTION, HRI 2024 COMPANION 2024

Abs Biblio

The most effective second language learning occurs through extensive interpersonal interaction and tutoring. However, limited funding and a lack of language teachers often prevent students from engaging in individualised practice, a lack which can be addressed using AI and social robots. We present a system that leverages generative AI to provide customized educational content in real-time, adapting to students’ skills through an engaging, visually-grounded game played alongside a social robot. To test effectiveness, we conducted a study in which Dutch high school students learned Spanish vocabulary either with or without the robot present. Results showed significant vocabulary gains regardless of robot presence, indicating the game itself, not the social embodiment, drove learning. While further refinements are needed, these findings highlight the potential for generative AI to deliver personalized language tutoring and circumvent the constraints posed by limited resources and staffing in schools. Ongoing work aims to enhance social presence and better align generative content with individuals’ abilities and pacing.
Automatic calibration for an open-source magnetic tactile sensor

Stockt, Lowiek, Proesmans, Remko, and wyffels, Francis

In International Conference on Robotics and Automation 2024, Proceedings 2024

Abs Biblio

Tactile sensing can enable robots to perform complex, contact-rich tasks. Magnetic sensors offer accurate three-axis force measurements while using affordable materials. Calibrating such a sensor involves either manual data collection, or automated procedures with precise mounting of the sensor relative to an actuator. We present an open-source magnetic tactile sensor with an automatic, in situ, gripper-agnostic calibration method, after which the sensor is immediately ready for use. Our goal is to lower the barrier to entry for tactile sensing, fostering collaboration in robotics. Design files and readout code can be found at https://github.com/LowiekVDS/Open-source-Magnetic-Tactile-Sensor.
AwarePrompt : using diffusion models to create methods for measuring value-aware AI architectures

Ciupinska, Kinga, Marchesi, Serena, Abbo, Giulio Antonio, Belpaeme, Tony, and Wykowska, Agnieszka

In Proceedings of the 16th International Conference on Agents and Artificial Intelligence : volume 3 2024

Biblio
Challenges with sign language datasets

Vandeghinste, Vincent, De Sisto, Mirella, Gómez, Santiago Egea, and De Coster, Mathieu

2024

Abs Biblio

Sign Languages are the primary means of communication more than half a million people in Europe alone. However, the development of sign language recognition and translation tools is slowed down by a series of obstacles concerning resource scarcity and, when data is available, in standardisation issues in the available data. The former challenge relates to the volume and quality of data available for machine learning as well as the time required to collect and process new data. The latter obstacle is linked to the variety of the data, i.e., annotation formats are not unified and vary amongst different resources. The available data formats are often not suitable for machine learning, obstructing the provision of automatic tools based on neural models. This chapter provides an overview of such challenges by comparing various sign language corpora and sign language machine learning datasets. Furthermore, it proposes a framework to address the lack of standardisation at format level, unify the available resources and facilitate sign language research for different languages. The framework takes ELAN files as inputs and returns textual and visual data ready to train sign language recognition and translation models. We present a proof of concept, training neural translation models on the data produced by the proposed framework.
Child speech recognition in human-robot interaction : problem solved?

Janssens, Ruben, Verhelst, Eva, Abbo, Giulio Antonio, Ren, Qiaoqiao, Pinto Bernal, Maria Jose, and Belpaeme, Tony

In TAHRI ’24 : Proceedings of the 2024 International Symposium on Technological Advances in Human-Robot Interaction 2024

Abs Biblio

Automated Speech Recognition shows superhuman performance for adult English speech on a range of benchmarks, but disappoints when fed children’s speech. This has long sat in the way of child-robot interaction. Recent evolutions in data-driven speech recognition, including the availability of Transformer architectures and unprecedented volumes of training data, might mean a breakthrough for child speech recognition and social robot applications aimed at children. We revisit a study on child speech recognition from 2017 and show that indeed performance has increased, with newcomer OpenAI Whisper doing markedly better than leading commercial cloud services. While transcription is not perfect yet, the best model recognises 60.3% of sentences correctly barring small grammatical differences, with sub-second transcription time running on a local GPU, showing potential for usable autonomous child-robot speech interactions.
Co-speech gestures : evaluation and generation

Wolfert, Pieter

2024

Abs Biblio

Dit proefschrift onderzoekt subjectieve evaluaties van gegenereerde co-spraakgebaren voor belichaamde gespreksagenten (ECA’s) en het genereren van non-verbaal spraak- en luistergedrag voor ECA’s. Dagelijkse communicatie omvat zowel verbale als non-verbale aspecten, waarbij non-verbaal gedrag, zoals gezichtsuitdrukkingen en gebaren, cruciaal is. Het proefschrift bekijkt data-gedreven benaderingen voor gebarengeneratie en legt de nadruk op subjectieve evaluatiemethoden. Verschillende evaluatiemethoden worden vergeleken, waaronder rangschikking, beoordelingsschalen en een vragenlijst voor het beoordelen van menselijkheid en geschiktheid. Beoordelingsschalen blijken meer informatie te verschaffen en gebruiksvriendelijker te zijn. Een vragenlijst voor het beoordelen van menselijkheid en geschiktheid geeft echter geen significante resultaten. De auteur rapporteert ook over twee georganiseerde challenges voor het genereren van co-spraak gebaren, waarbij de resultaten tussen mono- en dyadische interacties verschillen. Het proefschrift sluit af met het genereren van non-verbaal luister- en spreekgedrag, waarbij aangepaste modellen worden vergeleken met een basismodel. Hoewel gelijkenis met menselijk gedrag vergelijkbaar is, toont de aanpassing positieve resultaten voor de beoogde taken. Het onderzoek draagt bij aan het begrip en de evaluatie van computergegenereerde non-verbale communicatie.
Compliant robust control for robotic insertion of soft bodies

Liu, Yi, Verleysen, Andreas, and wyffels, Francis

IEEE ROBOTICS AND AUTOMATION LETTERS 2024

Abs Biblio

This letter proposes a novel framework for insertion-type tasks with soft bodies, such as cleaning a bottle with a soft brush. First, a multimodal model based on vision and force perception is trained. Domain randomization is used for the soft body’s properties to overcome the simulation-to- reality gap. Second, we propose a dynamic safety lock method based on force perception, which is embedded in the training model to make sure that the tool explores and traverses the hole’s path in a compliant way. This result in a higher success rate without damaging the tools/holes. Finally, we perform experiments in simulation and the real world, and the success rate of our proposed method reaches 85.14% in simulation and 83.45% in the real world. Ablation experiments in the real world demonstrate that our method is effective for complex paths and soft bodies with varying deformation intensities.
Computational thinking competencies of Flemish college students : vision on data collection

Lapage, Willem, wyffels, Francis, and Neutens, Tom

In Colloque Didapro 10 sur la Didactique de l’informatique et des STIC 2024

Abs Biblio

Computational thinking has become an increasingly vital competence in our technologically driven world. As a problem-solving methodology, it can be considered a competence that transcends disciplines and plays an important part in multiple diverse fields. It has also gained a more prominent role in the Flemish education system. Therefore, assessing computational thinking and collecting the necessary data to do so has become increasingly important during students’ education. This paper describes how the computational thinking competencies of college students can be monitored in a controlled environment. By combining a literature study as well as knowledge of the context wherein the data will be collected, a subset of data sources has been selected that show potential for a multimodal assessment of computational thinking. This paper outlines an envisioned data collection method to gauge computational thinking competencies among second-year computer science engineering students at Ghent University. The desired end result is a collection of data that can be managed and processed as an input source to assess computational thinking and affect educational practices. This paper describes a way of collecting data that shows potential for a multimodal assessment of computational thinking. It also opens the door for future research exploring the potential of AI-driven methods for automatic assessment and the development of interactive visualisation of said assessments.
Computational thinking integrated within a social robot project

Gesquière, Natacha, and Hermans, Seppe

In Colloque Didapro 10 sur la Didactique de l’informatique et des STIC 2024

Abs Biblio

This case study explores the integration of computationalthinking into the mandatory STEM curriculum in Flanders. It investigatesthe execution of the ’Social Robot’ project in three separate classeswithin the Flemish education system, each representing a distinct studytrack. These tracks include general education with a strong emphasis onSTEM, vocational education with a STEM focus, and vocational educationwithout a specific STEM focus. Initial findings reveal variations instudent interest, autonomous motivation, and self-efficacy among thesethree classes. In contrast to students in vocational education classes,those in the general education class display a heightened interest, perceivingthe project as both more challenging and engaging. This casestudy provides valuable insights into the intricacies of integrating computationalthinking within STEM education, emphasizing the necessityfor customized computational thinking initiatives, and underlining theimportance of infusing creativity and real-world relevance to enhancestudent motivation and learning outcomes.
Data-driven communicative behaviour generation : a survey

Oralbayeva, Nurziya, Aly, Amir, Sandygulova, Anara, and Belpaeme, Tony

ACM TRANSACTIONS ON HUMAN-ROBOT INTERACTION 2024

Abs Biblio

The development of data-driven behaviour generating systems has recently become the focus of considerable attention in the fields of human-agent interaction and human-robot interaction. Although rule-based approaches were dominant for years, these proved inflexible and expensive to develop. The difficulty of developing production rules, as well as the need for manual configuration to generate artificial behaviours, places a limit on how complex and diverse rule-based behaviours can be. In contrast, actual human-human interaction data collected using tracking and recording devices makes humanlike multimodal co-speech behaviour generation possible using machine learning and specifically, in recent years, deep learning. This survey provides an overview of the state of the art of deep learning-based co-speech behaviour generation models and offers an outlook for future research in this area.
Editorial : plant sensing and computing - PlantComp 2022

Stock, Michiel, De Swaef, Tom, and wyffels, Francis

2024

Biblio
Een heldere kijk op artificiële intelligentie, machinelearning en Large Language Models

Dambre, Joni

2024

Biblio
Empowering vocational students : a research-based framework for computational thinking integration

Hermans, Seppe, Neutens, Tom, wyffels, Francis, and Van Petegem, Peter

EDUCATION SCIENCES 2024

Abs Biblio

Vocational Education and Training (VET) faces significant challenges in equipping individuals for modern workplaces, which increasingly require digital literacy and Computational Thinking (CT) skills. This paper addresses the imperative of integrating CT into VET programs and outlines key research questions. Our methodology primarily involves a systematic literature review, resulting in the identification of 29 relevant papers. Through qualitative content analysis, we develop a CT integration framework that connects CT practices and integration elements to the engineering design process, while highlighting the VET context. Arguably, the innovative aspect of this framework lies in its core dimensions of harnessing computational power for enhanced efficiency. Raising the question of whether computers can optimize the efficiency and effectiveness of specific tasks is paramount for addressing challenges in technology-rich environments. Therefore, this inquiry merits unwavering attention at every stage of the process. The proposed framework provides educators with a structured approach to identify integration opportunities and help prepare students for multifaceted vocational careers. Furthermore, other key findings underscore the inherently interdisciplinary nature of VET, the growing demand for STEM competencies, and the transformative potential of CT integration. Implications emphasize the need for further research, supportive policies, and practical CT integration. Despite limitations, this study strongly advocates for CT integration, empowering VET students for success in the contemporary workforce.
Evaluating gesture generation in a large-scale open challenge : the GENEA Challenge 2022

Kucherenko, Taras, Wolfert, Pieter, Yoon, Youngwoo, Viegas, Carla, Nikolov, Teodor, Tsakov, Mihail, and Henter, Gustav Eje

ACM TRANSACTIONS ON GRAPHICS 2024

Abs Biblio

This paper reports on the second GENEA Challenge to benchmark data-driven automatic co-speech gesture generation. Participating teams used the same speech and motion dataset to build gesture-generation systems. Motion generated by all these systems was rendered to video using a standardised visualisation pipeline and evaluated in several large, crowdsourced user studies. Unlike when comparing different research papers, differences in results are here only due to differences between methods, enabling direct comparison between systems. The dataset was based on 18 hours of full-body motion capture, including fingers, of different persons engaging in a dyadic conversation. Ten teams participated in the challenge across two tiers: full-body and upper-body gesticulation. For each tier, we evaluated both the human-likeness of the gesture motion and its appropriateness for the specific speech signal. Our evaluations decouple human-likeness from gesture appropriateness, which has been a difficult problem in the field. The evaluation results show some synthetic gesture conditions being rated as significantly more human-like than 3D human motion capture. To the best of our knowledge, this has not been demonstrated before. On the other hand, all synthetic motion is found to be vastly less appropriate for the speech than the original motion-capture recordings. We also find that conventional objective metrics do not correlate well with subjective human-likeness ratings in this large evaluation. The one exception is the Fréchet gesture distance (FGD), which achieves a Kendall’s tau rank correlation of around -0.5. Based on the challenge results we formulate numerous recommendations for system building and evaluation.
Exploiting signal propagation delays to match task memory requirements in reservoir computing

Iacob, Stefan-Teodor, and Dambre, Joni

BIOMIMETICS 2024

Abs Biblio

Recurrent neural networks (RNNs) transmit information over time through recurrent connections. In contrast, biological neural networks use many other temporal processing mechanisms. One of these mechanisms is the inter-neuron delays caused by varying axon properties. Recently, this feature was implemented in echo state networks (ESNs), a type of RNN, by assigning spatial locations to neurons and introducing distance-dependent inter-neuron delays. These delays were shown to significantly improve ESN task performance. However, thus far, it is still unclear why distance-based delay networks (DDNs) perform better than ESNs. In this paper, we show that by optimizing inter-node delays, the memory capacity of the network matches the memory requirements of the task. As such, networks concentrate their memory capabilities to the points in the past which contain the most information for the task at hand. Moreover, we show that DDNs have a greater total linear memory capacity, with the same amount of non-linear processing power.
Exploring the effectiveness of evaluation practices for computer-generated nonverbal behaviour

Wolfert, Pieter, Henter, Gustav Eje, and Belpaeme, Tony

APPLIED SCIENCES-BASEL 2024

Abs Biblio

This paper compares three methods for evaluating computer-generated motion behaviour for animated characters: two commonly used direct rating methods and a newly designed questionnaire. The questionnaire is specifically designed to measure the human-likeness, appropriateness, and intelligibility of the generated motion. Furthermore, this study investigates the suitability of these evaluation tools for assessing subtle forms of human behaviour, such as the subdued motion cues shown when listening to someone. This paper reports six user studies, namely studies that directly rate the appropriateness and human-likeness of a computer character’s motion, along with studies that instead rely on a questionnaire to measure the quality of the motion. As test data, we used the motion generated by two generative models and recorded human gestures, which served as a gold standard. Our findings indicate that when evaluating gesturing motion, the direct rating of human-likeness and appropriateness is to be preferred over a questionnaire. However, when assessing the subtle motion of a computer character, even the direct rating method yields less conclusive results. Despite demonstrating high internal consistency, our questionnaire proves to be less sensitive than directly rating the quality of the motion. The results provide insights into the evaluation of human motion behaviour and highlight the complexities involved in capturing subtle nuances in nonverbal communication. These findings have implications for the development and improvement of motion generation models and can guide researchers in selecting appropriate evaluation methodologies for specific aspects of human behaviour.
Fine-tuning 3D foundation models for geometric object retrieval

Herrewegen, Jarne, Tourwé, Tom, Ovsjanikov, Maks, and wyffels, Francis

COMPUTERS & GRAPHICS-UK 2024

Abs Biblio

Foundation models, such as ULIP-2 (Xue et al., 2023) recently projected forward the field of 3D deep learning. These models are trained with significantly more data and show superior representation learning capacity in many downstream tasks like 3D shape classification and few-shot part segmentation. A particular characteristic of the recent 3D foundation models is that they are typically multi-modal, and involve image (2D) as well as caption (text) branches. This leads to an intricate interplay that benefits all modalities. At the same time, the nature of the 3D encoders alone, involved in these foundation models is not well-understood. Specifically, there is little analysis on the utility of both pre-trained 3D features provided by these models, or their capacity to adapt to new downstream 3D data. Furthermore, existing studies typically focus on label-oriented downstream tasks, such as shape classification, and ignore other critical applications, such as 3D content-based object retrieval. In this paper, we fill this gap and show, for the first time, how 3D foundation models can be leveraged for strong 3D-to-3D retrieval performance on seven different datasets, on par with state-of-the-art view-based architectures. We evaluate both the pre-trained foundation models, as well as their fine-tuned versions using downstream data. We compare supervised fine-tuning using classification labels against two self-supervised label-free fine-tuning methods. Importantly, we introduce and describe a methodology for fine-tuning, as we found this to be crucial to make transfer learning from 3D foundation models work in a stable manner.
Generative Fractional Diffusion Models

Nobis, Gabriel, Springenberg, Maximilian, Aversa, Marco, Detzel, Michael, Daems, Rembert, Murray-Smith, Roderick, Nakajima, Shinichi, Lapuschkin, Sebastian, Ermon, Stefano, Birdal, Tolga, Opper, Manfred, Knochenhauer, Christoph, Oala, Luis, and Samek, Wojciech

In Advances in Neural Information Processing Systems 2024

Biblio
KeyCLD : learning constrained Lagrangian dynamics in keypoint coordinates from images

Daems, Rembert, Taets, Jeroen, wyffels, Francis, and Crevecoeur, Guillaume

NEUROCOMPUTING 2024

Abs Biblio

We present KeyCLD, a framework to learn Lagrangian dynamics from images. Learned keypoints represent semantic landmarks in images and can directly represent state dynamics. We show that interpreting this state as Cartesian coordinates, coupled with explicit holonomic constraints, allows expressing the dynamics with a constrained Lagrangian. KeyCLD is trained unsupervised end-to-end on sequences of images. Our method explicitly models the mass matrix, potential energy and the input matrix, thus allowing energy based control. We demonstrate learning of Lagrangian dynamics from images on the cl_bnmsqnk pendulum, cartpole and acrobot environments. KeyCLD can be learned on these systems, whether they are unactuated, underactuated or fully actuated. Trained models are able to produce long-term video predictions, showing that the dynamics are accurately learned. We compare with Lag-VAE, Lag-caVAE and HGN, and investigate the benefit of the Lagrangian prior and the constraint function. KeyCLD achieves the highest valid prediction time on all benchmarks. Additionally, a very straightforward energy shaping controller is successfully applied on the fully actuated systems.
Language resources for European sign languages

Vandeghinste, Vincent, De Sisto, Mirella, Kopf, Maria, Schulder, Marc, Brosens, Caro, Soetemans, Lien, Omardeen, Rehana, Picron, Frankie, Van Landuyt, Davy, Murtagh, Irene, Avramidis, Elefterios, and De Coster, Mathieu

2024

Abs Biblio

This chapter on Language resources for Europe’s Sign Languages is an adaptation of a deliverable that is part of a series of language deliverables developed within the framework of the European Language Equality (ELE) project. The series seeks to not only delineate the current state of affairs for each European language, but to additionally identify the gaps and factors that hinder further development in research and technology. The survey presented here focuses on the condition of Language Technology (LT) with regard to Europe’s sign languages, a set of languages often forgotten in the context of European language equality. With the rise of the deep learning paradigm in artificial intelligence, sign language technologies become technologically feasible, provided that enough data is available to feed this data-hungry paradigm. It is exactly the quality and quantity of data that is the main bottleneck in development of well performing and useful technologies. In the past, there have been several projects aimed at developing sign language technologies and methodologies that have been deemed of little value by the deaf communities. Co-creation and involvement of deaf communities throughout projects and development of technologies ensures that this does not happen again.
Learning keypoints for robotic cloth manipulation using synthetic data

Lips, Thomas, De Gusseme, Victor-Louis, and wyffels, Francis

IEEE ROBOTICS AND AUTOMATION LETTERS 2024

Abs Biblio

Assistive robots should be able to wash, fold or iron clothes. However, due to the variety, deformability and self-occlusions of clothes, creating robot systems for cloth manipulation is challenging. Synthetic data is a promising direction to improve generalization, but the sim-to-real gap limits its effectiveness. To advance the use of synthetic data for cloth manipulation tasks such as robotic folding, we present a synthetic data pipeline to train keypoint detectors for almost-flattened cloth items. To evaluate its performance, we have also collected a real-world dataset. We train detectors for both T-shirts, towels and shorts and obtain an average precision of 64% and an average keypoint distance of 18 pixels. Fine-tuning on real-world data improves performance to 74% mAP and an average distance of only 9 pixels. Furthermore, we describe failure modes of the keypoint detectors and compare different approaches to obtain cloth meshes and materials. We also quantify the remaining sim-to-real gap and argue that further improvements to the fidelity of cloth assets will be required to further reduce this gap. The code, dataset and trained models are available here.
Machine translation from signed to spoken languages : state of the art and challenges

De Coster, Mathieu, Shterionov, Dimitar, Van Herreweghe, Mieke, and Dambre, Joni

UNIVERSAL ACCESS IN THE INFORMATION SOCIETY 2024

Abs Biblio

Automatic translation from signed to spoken languages is an interdisciplinary research domain on the intersection of computer vision, machine translation (MT), and linguistics. While the domain is growing in terms of popularity-the majority of scientific papers on sign language (SL) translation have been published in the past five years-research in this domain is performed mostly by computer scientists in isolation. This article presents an extensive and cross-domain overview of the work on SL translation. We first give a high level introduction to SL linguistics and MT to illustrate the requirements of automatic SL translation. Then, we present a systematic literature review of the state of the art in the domain. Finally, we outline important challenges for future research. We find that significant advances have been made on the shoulders of spoken language MT research. However, current approaches often lack linguistic motivation or are not adapted to the different characteristics of SLs. We explore challenges related to the representation of SL data, the collection of datasets and the evaluation of SL translation models. We advocate for interdisciplinary research and for grounding future research in linguistic analysis of SLs. Furthermore, the inclusion of deaf and hearing end users of SL translation applications in use case identification, data collection, and evaluation, is of utmost importance in the creation of useful SL translation models.
Memory-non-linearity trade-off in distance-based delay networks

Iacob, Stefan-Teodor, and Dambre, Joni

BIOMIMETICS 2024

Abs Biblio

The performance of echo state networks (ESNs) in temporal pattern learning tasks depends both on their memory capacity (MC) and their non-linear processing. It has been shown that linear memory capacity is maximized when ESN neurons have linear activation, and that a trade-off between non-linearity and linear memory capacity is required for temporal pattern learning tasks. The more recent distance-based delay networks (DDNs) have shown improved memory capacity over ESNs in several benchmark temporal pattern learning tasks. However, it has not thus far been studied whether this increased memory capacity comes at the cost of reduced non-linear processing. In this paper, we advance the hypothesis that DDNs in fact achieve a better trade-off between linear MC and non-linearity than ESNs, by showing that DDNs can have strong non-linearity with large memory spans. We tested this hypothesis using the NARMA-30 task and the bitwise delayed XOR task, two commonly used reservoir benchmark tasks that require a high degree of both non-linearity and memory.
Minimizing torque requirements in robotic manipulation through elastic elements optimization in a physics engine

Marchal, Maxime, Marzougui, Dries, Furnémont, Raphaël, Verstraten, Tom, and wyffels, Francis

INTERNATIONAL JOURNAL OF ADVANCED ROBOTIC SYSTEMS 2024

Abs Biblio

The increasing number of robots and the rising cost of electricity have spurred research into energy-reducing concepts in robotics. One such concept, elastic actuation, introduces compliant elements such as springs into the robot structure. This article presents a comparative analysis between two types of elastic actuation, namely, monoarticular parallel elastic actuation and biarticular parallel elastic actuation, and demonstrates an end-to-end pipeline for their optimization. Starting from the real-world system identification of an RRR robotic arm, we calibrate a simulation model in a general-purpose physics engine and employ in silico evolutionary optimization to co-optimize spring configurations and trajectories for a pick-and-place task. Finally, we successfully transfer the in silico optimized elastic elements and trajectory to the real-world prototype. Our results substantiate the ability of elastic actuation to reduce the actuators’ torque requirements heavily. In contrast to previous work, we highlight the superior performance of the biarticular variant over the monoarticular configuration. Furthermore, we show that a combination of both proves most effective. This work provides valuable insights into the torque-reducing use of elastic actuation and demonstrates an actuator-invariant in silico optimization methodology capable of bridging the sim2real gap.
Modelling and measuring forest microclimate at high spatiotemporal resolution

Walle, Emma, De Hertog, Steven, Meunier, Félicien, Calders, Kim, De Frenne, Pieter, Yang, Zhizhi, Li, Yanlu, Stock, Michiel, wyffels, Francis, Terryn, Louise, Sanczuk, Pieter, Verhelst, Tom, and Verbeeck, Hans

In Microclimate Ecology and Biogeography (ME&B), Abstracts 2024

Biblio
Multi-modal language learning : explorations on learning Japanese vocabulary

Wolfert, Pieter, De Gersem, Lisa, Janssens, Ruben, and Belpaeme, Tony

In COMPANION OF THE 2024 ACM/IEEE INTERNATIONAL CONFERENCE ON HUMAN-ROBOT INTERACTION, HRI 2024 COMPANION 2024

Abs Biblio

We explore robot-assisted language learning with a social robot, in which the robot teaches Japanese vocabulary. Specifically, we study if the mode of presentation of referents of nouns influences learning outcomes, and hypothesise that multimodal presentation of referents leads to improved learning outcomes. Three conditions are tested: referents are either presented as Japanese audio only, referents are visually presented, or referents are presented as actual objects that learners could pick up and manipulate. The learners were taught 4 words per condition and were distracted between the conditions with general questions related to the robot. There was a significant difference in the number of learned words between the audio-only and visual conditions, as well as between the audio-only and tactile conditions. No significant difference was found between the visual and tactile conditions. However, from our study, it follows that both these conditions are preferred over learning through only audio.
Multi-temporal 3D virtual forest reconstruction using terrestrial laser scanning in a temperate forest

Terryn, Louise, Calders, Kim, De Frenne, Pieter, Kuyken, Bart, Li, Yanlu, Sanczuk, Pieter, Stock, Michiel, Walle, Emma, wyffels, Francis, and Verbeeck, Hans

In ForestSAT 2024 Conference, Abstracts 2024

Abs Biblio

Multi-temporal 3D virtual forest reconstruction using terrestrial laser scanning in a temperate forest Introduction: European forests are currently undergoing large-scale changes in both structure and species composition, primarily driven by climate change, and various disturbances affecting the forest canopy. The canopies of forests across Europe are currently opening up due to tree mortality caused by factors such as drought, pests, storms, and fire. Understanding the implications of these changes for forest functionality is crucial for effective forest management. Therefore, it is essential to accurately quantify the spatial and temporal relationships among forest structure, light availability, and microclimate. Methods: To address this need, we have established a novel edge-to-core transect within a temperate deciduous forest near Ghent (Aelmoeseneiebos, Gontrode, Belgium). This 30m-wide transect spans from the forest edge to 135 meters deep into the forest interior, encompassing both an oak-beech-dominated zone and an ash-dominated zone characterized by ash dieback. Along the transect, we have deployed a densely spaced network of light and microclimate sensors at 15-meter intervals. Additionally, a fiber optic sensing cable for distributed temperature sensing (both in air and soil) runs along the entire transect. A 35-meter-high measuring tower is part of the setup, allowing for measurements of light and microclimate along a vertical transect from the ground till above the canopy. To quantify the temporal and spatial variations in forest structure, we have collected terrestrial laser scanning (TLS) data on a monthly basis, since March 2023. This data is acquired using a RIEGL VZ400i laser scanner at a pulse repetition rate of 600 kHz using a 15 by 15 m grid. Using this multi-temporal TLS data we will reconstruct a 3D virtual forest transect throughout time. To construct this 3D virtual forest transect, the transect point cloud is first fully segmented to individual tree point clouds in RIEGL’s RiSCAN PRO software using a combination of the software’s tree segmentation plug-in and manual corrections. Next, the tree point clouds undergo a leaf-wood separation using the GBSeparation algorithm of Tian et al. (2022). This is followed by reconstructing the woody points to the finest detail using cylinders (so called QSMs, quantitative structure models) for each individual tree applying the treeQSM version 2.0 workflow of Calders et al. (2015) which builds upon Raumonen et al. (2013). Leaves are added to the tree QSM structures using the Foliage and Needles Naïve Insertion (FaNNI) algorithm (Åkerblom et al., 218). Outlook: This virtual forest transect will serve as input for radiative transfer modeling (RTM), a simulation method that simulates the interaction between light and forest structure. Utilizing highly detailed forest structure obtained from TLS data, 3D RTMs provide an effective means to accurately model light transmission within forests at a high resolution. Using this approach we aim to (i) validate 3D light measurements conducted along the transect and (ii) implement virtual light sensors. The former enables the assessment of uncertainty in the collected time series data, while the latter offers a comprehensive understanding of the light conditions in the canopy. This information is crucial for evaluating the impact of canopy structure on light penetration and its subsequent effects on the microclimate. References: Åkerblom, M., Raumonen, P., Casella, E., Disney, M. I., Danson, F. M., Gaulton, R., ... & Kaasalainen, M. (2018). Non-intersecting leaf insertion algorithm for tree structure models. Interface Focus, 8(2), 20170045. Calders, K., Newnham, G., Burt, A., Murphy, S., Raumonen, P., Herold, M., ... & Kaasalainen, M. (2015). Nondestructive estimates of above‐ground biomass using terrestrial laser scanning. Methods in Ecology and Evolution, 6(2), 198-208. Raumonen, P., Kaasalainen, M., Åkerblom, M., Kaasalainen, S., Kaartinen, H., Vastaranta, M., ... & Lewis, P. (2013). Fast automatic precision tree models from terrestrial laser scanner data. Remote Sensing, 5(2), 491-520. Tian, Z., & Li, S. (2022). Graph-Based Leaf–Wood Separation Method for Individual Trees Using Terrestrial Lidar Point Clouds. IEEE Transactions on Geoscience and Remote Sensing, 60, 1-11.
Neural sign language recognition and translation : towards sign language applications powered by artificial intelligence

De Coster, Mathieu

2024

Abs Biblio

Artificiële intelligentie (AI) is sinds enkele jaren alomtegenwoordig. We bedienen toestellen met AI-gestuurde spraakherkenning en vertalen geschreven teksten met AI. Gebarentaalherkenning en -vertaling zijn echter minder geavanceerd. De twee voornaamste redenen hiervoor zijn een gebrek aan data en het feit dat deze uitdagende taken een interdisciplinaire aanpak vragen, met samenwerking tussen taalkundigen en computerwetenschappers. Dit proefschrift gaat aan de slag met gebarentaaldata, in het bijzonder voor de Vlaamse Gebarentaal (VGT). Dit doet het vanuit het perspectief van vier AI-gebaseerde toepassingen die in de toekomst kunnen ontwikkeld worden. Deze zijn: automatische transcriptie van gebarentaaldata, een zoekfunctie voor het vertalend woordenboek VGT-Nederlands, automatische vertaling van een gebarentaal naar geschreven taal en applicaties die toelaten individuele gebaren te leren en oefenen. De AI-modellen die beschreven zijn in dit onderzoek, zijn in staat om handvormen en bewegingen te herkennen. Op basis van deze bouwblokken kunnen ze ook gebaren herkennen, individueel of in een zin. De technieken die hiervoor gebruikt worden, zijn ook onderliggend aan de op dit moment meest geavanceerde AI-modellen. Op basis van het onderzoek zullen op de korte tot middellange termijn toepassingen voor gebarentalen kunnen ontwikkeld worden, gebruik makend van AI.
No more mumbles : enhancing robot intelligibility through speech adaptation

Ren, Qiaoqiao, Hou, Yuanbo, Botteldooren, Dick, and Belpaeme, Tony

IEEE ROBOTICS AND AUTOMATION LETTERS 2024

Abs Biblio

Spoken language interaction is at the heart of interpersonal communication, and people flexibly adapt their speech to different individuals and environments. It is surprising that robots, and by extension other digital devices, are not equipped to adapt their speech and instead rely on fixed speech parameters, which often hinder comprehension by the user. We conducted a speech comprehension study involving 39 participants who were exposed to different environmental and contextual conditions. During the experiment, the robot articulated words using different vocal parameters, and the participants were tasked with both recognising the spoken words and rating their subjective impression of the robot’s speech. The experiment’s primary outcome shows that spaces with good acoustic quality positively correlate with intelligibility and user experience. However, increasing the distance between the user and the robot exacerbated the user experience, while distracting background sounds significantly reduced speech recognition accuracy and user satisfaction. We next built an adaptive voice for the robot. For this, the robot needs to know how difficult it is for a user to understand spoken language in a particular setting. We present a prediction model that rates how annoying the ambient acoustic environment is and, consequentially, how hard it is to understand someone in this setting. Then, we develop a convolutional neural network model to adapt the robot’s speech parameters to different users and spaces, while taking into account the influence of ambient acoustics on intelligibility. Finally, we present an evaluation with 27 users, demonstrating superior intelligibility and user experience with adaptive voice parameters compared to fixed voice.
Plant science in the age of simulation intelligence

Stock, Michiel, Pieters, Olivier, De Swaef, Tom, and wyffels, Francis

FRONTIERS IN PLANT SCIENCE 2024

Abs Biblio

Historically, plant and crop sciences have been quantitative fields that intensively use measurements and modeling. Traditionally, researchers choose between two dominant modeling approaches: mechanistic plant growth models or data-driven, statistical methodologies. At the intersection of both paradigms, a novel approach referred to as simulation intelligence, has emerged as a powerful tool for comprehending and controlling complex systems, including plants and crops. This work explores the transformative potential for the plant science community of the nine simulation intelligence motifs, from understanding molecular plant processes to optimizing greenhouse control. Many of these concepts, such as surrogate models and agent-based modeling, have gained prominence in plant and crop sciences. In contrast, some motifs, such as open-ended optimization or program synthesis, still need to be explored further. The motifs of simulation intelligence can potentially revolutionize breeding and precision farming towards more sustainable food production.
Predicting errors and failures in human-robot interaction from multi-modal temporal data

Janssens, Ruben, Verhelst, Eva, and De Coster, Mathieu

In PROCEEDINGS OF THE 26TH INTERNATIONAL CONFERENCE ON MULTIMODAL INTERACTION, ICMI 2024 2024

Abs Biblio

Social robots should be able to detect social signals sent by their user, such as when the robot made a mistake or the user feels awkward. As our submission to the ERR@HRI challenge, we present a number of neural and traditional machine learning models to predict when this occurs in a human-robot conversation, based on facial expressions, body pose, and non-verbal speech characteristics. The small size of the dataset, imbalance in the label distribution, and low-frequency label annotations provided significant challenges. However, three of our approaches show promising results: modifying the training of a gated recurrent unit (GRU) model to predict at lower frequency than that of the input features, using an embedding layer and convolutional neural network to pre-process temporal data before feeding it to the GRU, and using traditional random forests.
Predictive turn-taking : leveraging language models to anticipate turn transitions in human-robot dialogue

Pinto Bernal, Maria Jose, and Belpaeme, Tony

In 2024 33RD IEEE INTERNATIONAL CONFERENCE ON ROBOT AND HUMAN INTERACTIVE COMMUNICATION, ROMAN 2024 2024

Abs Biblio

Natural and engaging spoken dialogue systems require seamless turn-taking coordination to avoid awkward interruptions and unnatural pauses. Traditional systems often rely on simplistic silence thresholds, relinquishing the turn after a predetermined period of silence, which invariably leads to a suboptimal interaction experience. This work explores the potential of Large Language Models (LLMs) for improved turn-taking prediction. Building upon research that uses linguistic cues, we investigate how LLMs, with their rich contextual knowledge and semantic encoding of language, can be used for this task. We hypothesize that by analysing dialogue context, syntactic structure, and pragmatic cues within the user’s utterance, LLMs can offer more accurate turn-completion predictions. This research evaluates the capabilities of recent LLMs such as Gemini, OpenAI’s API, Anthropic’s Claude2, and Meta AI’s Llama 2 to predict turn-ending points solely based on textual information, and demonstrates how the conversation between elderly users and companion robots can be enhanced by LLM-powered end-of-turn prediction.
Quantifying and modelling feedbacks between forest structure, light, microclimate and carbon cycling in temperate forests

Walle, Emma, De Hertog, Steven, Meunier, Félicien, Calders, Kim, De Frenne, Pieter, Yang, Zhizhi, Stock, Michiel, wyffels, Francis, Terryn, Louise, Sanczuk, Pieter, Verhelst, Tom, and Verbeeck, Hans

In EGU General Assembly 2024, Abstracts 2024

Abs Biblio

Studying the feedback between forest structure and the environment, particularly below canopies, is crucial for sustainable forest management, biodiversity conservation, and climate mitigation. Advanced vegetation models play a key role in unraveling the complex interaction between forest composition and environmental conditions, as these allow to understand the dynamics of ecosystems by simulating the interactions between plant species and their environment. An essential aspect necessitating refinement in these models is understanding how radiation interacts with intricate structures like forest canopies. In this study, we employ advanced terrestrial laser scanning techniques, distributed fiber-optic, and microclimate sensors to investigate the relationships between light, microclimate, carbon cycling, and forest structure in temperate forests. In a temperate forest in Belgium, we implemented a sensor setup since March 2023. It comprises a Distributed Temperature Sensing (DTS) fiber, Temperature and Moisture Sensor (TMS) microclimate loggers, SurveyTag microclimate loggers, Photosynthetic Active Radiation (PAR) sensors, and pyranometer (direct/diffuse) light sensors along a 135 m long transect from forest edge to core. Monthly 3D terrestrial laser scanning (TLS) of the transect allowed us to quantify forest structure with high spatiotemporal resolution. Preliminary results reveal distinct microclimate gradients along the transect and seasonal changes in forest structure in 3D space, including budding and changes in canopy volume. These findings will be used to calibrate and improve existing radiative transfer models (RTMs) to be further implemented in vegetation models. Integrating observations and model parameters in a common framework will provide breakthrough insights into the feedbacks between light, forest structure, microclimate, and their impact on the carbon cycle in temperate forests.
Robotic grasping and manipulation competition at the 2024 IEEE/RAS International Conference on Robotics and Automation

Sun, Yu, Calli, Berk, Kimble, Kenny, wyffels, Francis, De Gusseme, Victor-Louis, Hang, Kaiyu, D’Avella, Salvatore, Xompero, Alessio, Cavallaro, Andrea, Roa, Maximo A., Avendano, Jose, and Mavrommati, Anastasia

2024

Abs Biblio

The Ninth Robotic Grasping and Manipulation Competition (RGMC) took place in Yokohama, Japan, during the 2024 IEEE/RAS International Conference on Robotics and Automation (ICRA). The series of RGMC events started in 2016 at the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) with strong support from the conference’s organization committee, and since then they have been held each year at ICRA or IROS [1]. Across the editions, RGMC engaged the community in solving the open challenges associated with various robotic grasping and manipulation tasks for manufacturing, service robots, and logistics, and in advancing research and technology towards more realistic scenarios that can be encountered in daily activities at home or in warehouses. These tasks include assembling and disassembling boards, hand-in-hand grasping, picking and placing various objects, pouring liquids into a cup, bin picking, rearranging and setting formal tables, folding and unfolding cloths, and receiving objects handed over by a person. The goal of RGMC across these tasks is to assess the autonomous manipulation capabilities of a robotic arm when dealing with unknown or novel objects with varying physical properties and when handling scenarios with various degrees of uncertainty caused by a cluttered scene, random initial configurations, or human behaviors when interacting with the robot. For example, objects can vary in their shapes, appearances, transparency, deformability, and weight.
Sign languages as source language for machine translation : historical overview and challenges

Dambre, Joni, and De Coster, Mathieu

2024

Abs Biblio

Sign language machine translation (SLMT) is machine translation (MT) in which at least the source or the target language is a sign language, but the combination of both is also possible. In general, today’s approaches to SLMT heavily build on traditional (i.e., text-to-text) MT systems and focus on solutions to adapt those to having a sign language at the input side or at the output side. Both cases lead to very different challenges and technological solutions, and this chapter focuses only on having a sign language as the source language. It covers the differences between written and sign languages that must be taken into account, the historical evolutions we have seen in the field and the way they have been impacted by the amount and quality of the available data.
Sim-to-real dataset of industrial metal objects

De Roovere, Peter, Moonen, Steven, Michiels, Nick, and wyffels, Francis

MACHINES 2024

Abs Biblio

We present a diverse dataset of industrial metal objects with unique characteristics such as symmetry, texturelessness, and high reflectiveness. These features introduce challenging conditions that are not captured in existing datasets. Our dataset comprises both real-world and synthetic multi-view RGB images with 6D object pose labels. Real-world data were obtained by recording multi-view images of scenes with varying object shapes, materials, carriers, compositions, and lighting conditions. This resulted in over 30,000 real-world images. We introduce a new public tool that enables the quick annotation of 6D object pose labels in multi-view images. This tool was used to provide 6D object pose labels for all real-world images. Synthetic data were generated by carefully simulating real-world conditions and varying them in a controlled and realistic way. This resulted in over 500,000 synthetic images. The close correspondence between synthetic and real-world data and controlled variations will facilitate sim-to-real research. Our focus on industrial conditions and objects will facilitate research on computer vision tasks, such as 6D object pose estimation, which are relevant for many industrial applications, such as machine tending. The dataset and accompanying resources are available on the project website.
Social value alignment in large language models

Abbo, Giulio Antonio, Marchesi, Serena, Wykowska, Agnieszka, and Belpaeme, Tony

In VALUE ENGINEERING IN ARTIFICIAL INTELLIGENCE, VALE 2023 2024

Abs Biblio

Large Language Models (LLMs) have demonstrated remarkable proficiency in text generation and display an apparent understanding of both physical and social aspects of the world. In this study, we look into the capabilities of LLMs to generate responses that align with human values. We focus on five prominent LLMs - GPT-3, GPT-4, PaLM-2, LLaMA-2 and BLOOM - and compare their generated responses with those provided by human participants. To evaluate the value alignment of LLMs, we presented domestic scenarios to the model and elicited a response with minimal prompting instructions. Human raters judged the responses on appropriateness and value alignment. The results revealed that GPT-3, 4 and PaLM-2 performed on par with human participants, displaying a notable level of value alignment in their generated responses. However, LLaMA-2 and BLOOM fell short in this aspect, indicating a possible divergence from human values. Furthermore, our findings indicate that the raters faced difficulty in distinguishing between responses generated by LLMs and those by humans, with raters exhibiting a preference for machine-generated responses in certain cases. These findings shed light on the capabilities of state-of-the-art LLMs to align with human values, but also allow us to speculate on whether these models could be value-aware. This research contributes to the ongoing exploration of LLMs’ understanding of ethical considerations and provides insights into their potential for engaging in value-driven interactions.
Tactile interaction with social robots influences attitudes and behaviour

Ren, Qiaoqiao, and Belpaeme, Tony

INTERNATIONAL JOURNAL OF SOCIAL ROBOTICS 2024

Abs Biblio

Tactile interaction plays an essential role in human-to-human interaction. People gain comfort and support from tactile interactions with others and touch is an important predictor for trust. While touch has been explored as a communicative modality in HCI and HRI, we here report on two studies in which touching a social robot is used to regulate people’s stress levels and consequently their actions. In the first study, we look at whether different intensities of tactile interaction result in a physiological response related to stress, and whether the interaction impacts risk-taking behaviour and trust. We let 38 participants complete a balloon analogue risk task (BART), a computer-based game that serves as a proxy for risk-taking behaviour. In our study, participants are supported by a robot during the BART task. The robot builds trust and encourages participants to take more risk. The results show that affective tactile interaction with the robot increases participants’ risk-taking behaviour, but gentle affective tactile interaction increases comfort and lowers stress whereas high-intensity touch does not. We also find that male participants exhibit more risk-taking behaviour than females while being less stressed. Based on this experiment, a second study is used to ascertain whether these effects are caused by the social nature of tactile interaction or by the physical interaction alone. For this, instead of a social robot, participants now have a tactile interaction with a non-social device. The non-social interaction does not result in any effect, leading us to conclude that tactile interaction with humanoid robots is a social phenomenon rather than a mere physical phenomenon.
The ICRA 2024 cloth competition : benchmarking robotic cloth unfolding

De Gusseme, Victor-Louis, Proesmans, Remko, and wyffels, Francis

In 40th Anniversary of the IEEE International Conference on Robotics and Automation (ICRA@40), Abstracts 2024

Abs Biblio

Unfolding cloth in the air is a fundamental task in robotic cloth manipulation, essential for various subsequent tasks like folding, hanging, or even assisted dressing. It has been extensively studied, yet, until now a standardized performance comparison across many different labs and methods has never been done. The ICRA 2024 Cloth Competition marks a significant milestone by offering a head-to-head evaluation of 11 diverse teams on a shared real-world robotic platform. This unprecedented event not only sets a new benchmark for the field but also fosters collaboration and innovation, while providing a comprehensive dataset to drive future research towards more robust and generalizable solutions in robotic cloth manipulation.
Towards a Definition of Awareness for Embodied AI

Abbo, Giulio Antonio, Marchesi, Serena, Ciupinska, Kinga, Wykowska, Agnieszka, and Belpaeme, Tony

In Proceedings of the 16th International Conference on Agents and Artificial Intelligence : volume 3 2024

Biblio
Variational inference for SDEs driven by fractional noise

Daems, Rembert, Opper, Manfred, Crevecoeur, Guillaume, and Birdal, Tolga

In ICLR 2024 : the Twelfth International Conference on Learning Representations, Proceedings 2024

Biblio

2023

’Am I listening?’ : evaluating the quality of generated data-driven listening motion

Wolfert, Pieter, Henter, Gustav Eje, and Belpaeme, Tony

In ICMI ’23 Companion : Companion Publication of the 25th International Conference on Multimodal Interaction 2023

Abs Biblio

This paper asks if recent models for generating co-speech gesticulation also may learn to exhibit listening behaviour as well. We consider two models from recent gesture-generation challenges and train them on a dataset of audio and 3D motion capture from dyadic conversations. One model is driven by information from both sides of the conversation, whereas the other only uses the character’s own speech. Several user studies are performed to assess the motion generated when the character is speaking actively, versus when the character is the listener in the conversation. We find that participants are reliably able to discern motion associated with listening, whether from motion capture or generated by the models. Both models are thus able to produce distinctive listening behaviour, even though only one model is truly a listener, in the sense that it has access to information from the other party in the conversation. Additional experiments on both natural and model-generated motion finds motion associated with listening to be rated as less human-like than motion associated with active speaking.
’Robots for good’: ten defining questions

Sabanovic, Selma, Charisi, Vicky, Belpaeme, Tony, Bethel, Cindy L., Mataric, Maja, Murphy, Robin, and Levy-Tzedek, Shelly

2023

Abs Biblio

Ten questions to guide reflection and assessment of the "good" in robotics projects are suggested. Ten questions to guide reflection and assessment of the "good" in robotics projects are suggested.
A bio-inspired model for audio processing

Cazalets, Tanguy, and Dambre, Joni

In 2023 RIVF International Conference on Computing and Communication Technologies (RIVF) 2023

Abs Biblio

Homeostatic Activity Dependant Structural Plasticity (HADSP) is a recently introduced technique to generate network using structural plasticity. The algorithm use only homeostatic plasticity but let emerge principles of Hebbian learning. A previous study suggested that HADSP was able to generate networks that effectively leverage the inter-relationships between correlated time series but the idea was tested only on simple benchmarks. This paper examines HADSP’s performance in speech recognition, its first application on a realistic dataset. Mimicking human hearing, a single-variable recording is transformed into a multi-variable time series through audio processing. The bio-inspired HADSP algorithm then creates a reservoir computing architecture, enhancing data representation and improving performance of the reservoir. Our principal results are that using spectral representation of the audio signal greatly improves the performance of speech recognition for echo state networks (ESNs). HADSP generated architectures show improvements in performance, corroborating the algorithm capacity to generate better reservoir connectivity.
A robust and safe strategy for robotic assembly

Liu, Yi, Verleysen, Andreas, and wyffels, Francis

In ICRA2023 : Embracing Contacts workshop : IEEE International Conference on Robotics and Automation, Proceedings 2023

Biblio
AI in de zorg : beslissingsbomen in de zorgsector : leerlingencursus : finaliteit doorstroom

Gesquière, Natacha, Neutens, Tom, and wyffels, Francis

2023

Abs Biblio

In het project ‘AI in de Zorg’ maken leerlingen kennis met een beslissingsboom, een techniek uit machinaal leren, die veel gebruikt wordt in de zorgsector. De principes van zo’n beslissingsboom zijn al te begrijpen met de leerstof van de tweede graad van het secundair onderwijs. De Early Warning Score (EWS) is een richtlijn die internationaal wordt toegepast in ziekenhuizen om de gezondheid van een patiënt te kunnen inschatten. De EWS is gebaseerd op de vitale functies: bloeddruk, lichaamstemperatuur, hartslag, ademhalingsfrequentie, alertheid en zuurstofsaturatie. In het project ‘AI in de Zorg’ bekijken leerlingen hoe de EWS gebruikt wordt in de praktijk. Bovendien gaan ze na hoe een computer op basis van kunstmatige intelligentie automatisch het risiconiveau van een patiënt kan inschatten. Beslissingsbomen uit de zorgsector zijn maatschappelijk relevant en vormen de ideale context voor een STEM-project dat leerlingen concepten van artificiële intelligentie bijbrengt.
An homeostatic activity-dependent structural plasticity algorithm for richer input combination

Cazalets, Tanguy, and Dambre, Joni

In 2023 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, IJCNN 2023

Abs Biblio

This paper introduces a novel rate-based variant of homeostatic activity-dependent structural plasticity (HADSP) for echo state networks. Despite its importance in brain development, structural plasticity has been largely overlooked in artificial neural networks. Our algorithm, although using only homeostatic plasticity, let emerge principles of Hebbian learning. Our analysis sheds light on the information processing capabilities of HADSP-powered echo state networks and suggests that HADSP effectively leverages the inter-relationships of the network’s inputs. The study highlights the potential for rate-based HADSP to contribute to the field of computational neuroscience and plasticity in echo state networks. Furthermore, our findings highlight the crucial role of structural plasticity in influencing network function and organization and contribute significantly to the ongoing research on leveraging plasticity for the advancement of reservoir computing techniques.
Augmenting off-the-shelf grippers with tactile sensing

Proesmans, Remko, and wyffels, Francis

In ICRA 2023 : the International Conference on Robotics and Automation, Proceedings 2023

Abs Biblio

The development of tactile sensing and its fusion with computer vision is expected to enhance robotic systems in handling complex tasks like deformable object manipulation. However, readily available industrial grippers typically lack tactile feedback, which has led researchers to develop and integrate their own tactile sensors. This has resulted in a wide range of sensor hardware, making it difficult to compare performance between different systems. We highlight the value of accessible open-source sensors and present a set of fingertips specifically designed for fine object manipulation, with readily interpretable data outputs. The fingertips are validated through two difficult tasks: cloth edge tracing and cable tracing. Videos of these demonstrations, as well as design files and readout code can be found at https://github.com/RemkoPr/icra-2023-workshop-tactile-fingertips.
Behavioural models of risk-taking in human-robot tactile interactions

Ren, Qiaoqiao, Hou, Yuanbo, Botteldooren, Dick, and Belpaeme, Tony

SENSORS 2023

Abs Biblio

Touch can have a strong effect on interactions between people, and as such, it is expected to be important to the interactions people have with robots. In an earlier work, we showed that the intensity of tactile interaction with a robot can change how much people are willing to take risks. This study further develops our understanding of the relationship between human risk-taking behaviour, the physiological responses by the user, and the intensity of the tactile interaction with a social robot. We used data collected with physiological sensors during the playing of a risk-taking game (the Balloon Analogue Risk Task, or BART). The results of a mixed-effects model were used as a baseline to predict risk-taking propensity from physiological measures, and these results were further improved through the use of two machine learning techniques-support vector regression (SVR) and multi-input convolutional multihead attention (MCMA)-to achieve low-latency risk-taking behaviour prediction during human-robot tactile interaction. The performance of the models was evaluated based on mean absolute error (MAE), root mean squared error (RMSE), and R squared score (R-2), which obtained the optimal result with MCMA yielding an MAE of 3.17, an RMSE of 4.38, and an R-2 of 0.93 compared with the baseline of 10.97 MAE, 14.73 RMSE, and 0.30 R-2. The results of this study offer new insights into the interplay between physiological data and the intensity of risk-taking behaviour in predicting human risk-taking behaviour during human-robot tactile interactions. This work illustrates that physiological activation and the intensity of tactile interaction play a prominent role in risk processing during human-robot tactile interaction and demonstrates that it is feasible to use human physiological data and behavioural data to predict risk-taking behaviour in human-robot tactile interaction.
CenDerNet : center and curvature representations for render-and-compare 6D pose estimation

De Roovere, Peter, Daems, Rembert, Croenen, Jonathan, Bourgana, Taoufik, Hoog, Joris, and wyffels, Francis

In Computer Vision : ECCV 2022 Workshops : proceedings, part VIII 2023

Abs Biblio

We introduce CenDerNet, a framework for 6D pose estimation from multi-view images based on center and curvature representations. Finding precise poses for reflective, textureless objects is a key challenge for industrial robotics. Our approach consists of three stages: First, a fully convolutional neural network predicts center and curvature heatmaps for each view; Second, center heatmaps are used to detect object instances and find their 3D centers; Third, 6D object poses are estimated using 3D centers and curvature heatmaps. By jointly optimizing poses across views using a render-and-compare approach, our method naturally handles occlusions and object symmetries. We show that CenDerNet outperforms previous methods on two industry-relevant datasets: DIMO and T-LESS .
Delay-sensitive local plasticity in echo state networks

Iacob, Stefan-Teodor, Chavlis, Spyridon, Poirazi, Panayiota, and Dambre, Joni

In 2023 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, IJCNN 2023

Abs Biblio

Time delays are inherently present in any physical or biological network. However, the role of delays in echo state networks (ESNs) has only been touched upon. In recent years, the use of local plasticity has been explored in the field of reservoir computing, and specifically in ESNs. In this paper, we investigate the role of distance dependent inter-neuron delays in adaptive reservoirs. We introduce a novel ESN design called adaptive distance-based delay network (ADDN), that combines inter-neuron delays with local synaptic plasticity in the reservoir weights using a delay sensitive version of the Bienenstock-Cooper-Munro (BCM) rule. We show that ADDNs perform better on prediction tasks compared to ESNs, regular distance-based delay networks, and ESNs with conventional BCM connections. We optimized the hyperparameters of ADDNs and each of the baseline models using covariance matrix adaptation evolution strategy (CMA-ES). We prove that with ADDNs, we can evolve a single set of hyperparameters that can generate networks which, after unsupervised adaptation, can obtain good performance on different Mackey-Glass sequences with a range of different time constants. By adapting its reservoir weights to the dynamics of the input data, ADDNs can generalize between versions of the same “class” of tasks.
Do different robot appearances change emotion recognition in children with ASD?

Pinto Bernal, Maria Jose, Sierra M., Sergio D., Munera, Marcela, Casas, Diego, Villa-Moreno, Adriana, Frizera-Neto, Anselmo, Stoelen, Martin F., Belpaeme, Tony, and Cifuentes, Carlos A.

FRONTIERS IN NEUROROBOTICS 2023

Abs Biblio

IntroductionSocially Assistive Robotics has emerged as a potential tool for rehabilitating cognitive and developmental disorders in children with autism. Social robots found in the literature are often able to teach critical social skills, such as emotion recognition and physical interaction. Even though there are promising results in clinical studies, there is a lack of guidelines on selecting the appropriate robot and how to design and implement the child-robot interaction. MethodsThis work aims to evaluate the impacts of a social robot designed with three different appearances according to the results of a participatory design (PD) process with the community. A validation study in the emotion recognition task was carried out with 21 children with autism. ResultsSpectrum disorder results showed that robot-like appearances reached a higher percentage of children’s attention and that participants performed better when recognizing simple emotions, such as happiness and sadness. DiscussionThis study offers empirical support for continuing research on using SAR to promote social interaction with children with ASD. Further long-term research will help to identify the differences between high and low-functioning children.
Evaluation of skin prick location on the forearm using a novel skin prick automated test device

Seys, Sven F., Gorris, Senne, Uyttebroek, Saartje, Backaert, Wout, Jorissen, Mark, Schrijvers, Rik, Daems, Rembert, Loeckx, Dirk, Van Gerven, Laura, and Hellings, Peter

FRONTIERS IN ALLERGY 2023

Abs Biblio

BackgroundThe skin prick test (SPT) is the gold standard for identifying allergic sensitization in individuals suspected of having an inhalant allergy. Recently, it was demonstrated that SPT using a novel skin prick automated test (SPAT) device showed increased reproducibility and tolerability compared to the conventional SPT, among other benefits.ObjectiveThis study aimed to evaluate prick location bias using the novel SPAT device.MethodsA total of 118 volunteers were enrolled in this study and underwent SPATs with histamine (nine pricks) and glycerol control (one prick) solutions on the volar side of their forearms. Imaging of the skin reactions was performed using the SPAT device, and the physician determined the longest wheal diameter by visually inspecting the images using a web interface. Prick location bias was assessed along the medial vs. lateral and proximal vs. distal axes of the forearm.ResultsIn total, 944 histamine pricks were analyzed. Four medial and four lateral histamine pricks were grouped, and wheal sizes were compared. The longest wheal diameters were not significantly different between the medial and lateral prick locations (p = 0.41). Furthermore, the pricks were grouped by two based on their position on the proximal-distal axis of the forearm. No significant difference was observed among the four groups of analyzed prick locations (p = 0.73).ConclusionThe prick location on the volar side of the forearm did not influence wheal size in SPAT-pricked individuals.
Experimental results on nonlinear distortion compensation using photonic reservoir computing with a single set of weights for different wavelengths

Gooskens, Emmanuel, Sackesyn, Stijn, Dambre, Joni, and Bienstman, Peter

SCIENTIFIC REPORTS 2023

Abs Biblio

Photonics-based computing approaches in combination with wavelength division multiplexing offer a potential solution to modern data and bandwidth needs. This paper experimentally takes an important step towards wavelength division multiplexing in an integrated waveguide-based photonic reservoir computing platform by using a single set of readout weights for up to at least 3 ITU-T channels to efficiently scale the data bandwidth when processing a nonlinear signal equalization task on a 28 Gbps modulated on-off keying signal. Using multiple-wavelength training, we obtain bit error rates well below that of the 1.5 x 10(-2) forward error correction limit at high fiber input powers of 18 dBm, which result in high nonlinear distortion. The results of the reservoir chip are compared to a tapped delay line filter and clearly show that the system performs nonlinear equalization. This was achieved using only limited post processing which in future work can be implemented in optical hardware as well.
From scarcity to understanding : transfer learning for the extremely low resource Irish sign language

Holmes, R., Rushe, E., De Coster, Mathieu, Bonnaerens, Maxim, Satoh, S., Sugimoto, A., and Ventresque, A.

In 2023 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION WORKSHOPS, ICCVW 2023

Abs Biblio

One of the most significant challenges to sign language recognition (SLR) today is the low resource nature of sign language datasets, with many datasets being extremely low resource. Transfer learning is therefore a promising, and likely indispensable, method of increasing recognition performance. The use of pose estimation models, which are typically trained on a large and diverse population, can also aid generalization for extremely low resource sign languages. However, research on transfer learning for pose estimation keypoints as inputs has been limited. In this work, we explore transfer learning as a means to improve SLR classification performance for the extremely low resource Irish Sign Language (ISL). We show that transfer learning on larger datasets containing secondary sign languages significantly improves performance on our target sign language, ISL. To understand these results and the attributes that make one dataset better than another for pre-training, we analyse the linguistic relationships between these datasets. We find that certain attributes of datasets are associated with better transfer learning performance. We hope that our findings will not only motivate further research into transfer learning for pose keypoint-based SLR but also act as a practical guide to researchers on choosing the most suitable datasets with which to pre-train models.
Homeostatic activity-dependent structural plasticity in rate based neural networks for reacher input combination

Cazalets, Tanguy, and Dambre, Joni

In Bernstein Conference 2023, Abstracts 2023

Biblio
Improving the classification accuracy in label-free flow cytometry using event-based vision and simple logistic regression

Gouda, Muhammed Gouda Ahmed, Lugnan, Alessio, Dambre, Joni, Branden, Gerd, Posch, Christoph, and Bienstman, Peter

IEEE JOURNAL OF SELECTED TOPICS IN QUANTUM ELECTRONICS 2023

Abs Biblio

Event-based cameras are novel bio-inspired vision sensors that do not follow the mechanism of traditional frame-based cameras. In the field of data acquisition, the replacement of CMOS cameras with event-based cameras has proved to enhance the accuracy of machine learning methods in situations where critical lighting conditions and rapid dynamics are paramount. In this paper, we investigate for the first time the use of extreme learning machines on data coming from event-based cameras in the context of flow cytometry. Except for the image sensor, the experimental setup is similar to a setup we used in (Lugnan et al., 2020) where we showed that a simple linear classifier can achieve around 10% error rate on background-subtracted cell frames. Here, we show that the the error rate of this simple imaging flow cytometer could be decreased to less than 2% just by making use of the capabilities of an event camera. Moreover, additional benefits like more sensitivity and efficient memory usage are gained. Finally, we suggest further possible improvements to the experimental setup used to record events from flowing micro-particles allowing for more accurate and stable cellx sorting.
Integrated STEM professional development in interdisciplinary teacher design teams : teacher self-efficacy profiles using cluster analysis

Hermans, Seppe, Gesquière, Natacha, wyffels, Francis, and Van Petegem, Peter

In STEM & Open Schooling for Sustainability Education : proceedings of the 4th Educating the Educators Conference 2023

Biblio
Integrated photonic reservoir computing with an all-optical readout

Ma, Chonghuai, Van Kerrebrouck, Joris, Deng, Hong, Sackesyn, Stijn, Gooskens, Emmanuel, Bai, Bing, Dambre, Joni, and Bienstman, Peter

OPTICS EXPRESS 2023

Abs Biblio

Integrated photonic reservoir computing has been demonstrated to be able to tackle different problems because of its neural network nature. A key advantage of photonic reservoir computing over other neuromorphic paradigms is its straightforward readout system, which facilitates both rapid training and robust, fabrication variation-insensitive photonic integrated hardware implementation for real-time processing. We present our recent development of a fully-optical, coherent photonic reservoir chip integrated with an optical readout system, capitalizing on these benefits. Alongside the integrated system, we also demonstrate a weight update strategy that is suitable for the integrated optical readout hardware. Using this online training scheme, we successfully solved 3-bit header recognition and delayed XOR tasks at 20 Gbps in real-time, all within the optical domain without excess delays.(c) 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
Is the autonomous social robot within reach?

Belpaeme, Tony

In PROCEEDINGS OF THE 11TH CONFERENCE ON HUMAN-AGENT INTERACTION, HAI 2023 2023

Biblio
KeyCLD : learning constrained Lagrangian dynamics in keypoint coordinates from images

Daems, Rembert, Taets, Jeroen, wyffels, Francis, and Crevecoeur, Guillaume

In Machine Learning and the Physical Sciences Workshop, NeurIPS 2023, Proceedings 2023

Biblio
Learned thresholds token merging and pruning for vision transformers

Bonnaerens, Maxim, and Dambre, Joni

TRANSACTIONS ON MACHINE LEARNING RESEARCH 2023

Abs Biblio

Vision transformers have demonstrated remarkable success in a wide range of computer vision tasks over the last years. However, their high computational costs remain a significant barrier to their practical deployment. In particular, the complexity of transformer models is quadratic with respect to the number of input tokens. Therefore techniques that reduce the number of input tokens that need to be processed have been proposed. This paper introduces Learned Thresholds token Merging and Pruning (LTMP), a novel approach that leverages the strengths of both token merging and token pruning. LTMP uses learned threshold masking modules that dynamically determine which tokens to merge and which to prune. We demonstrate our approach with extensive experiments on vision transformers on the ImageNet classification task. Our results demonstrate that LTMP achieves state-of-the-art accuracy across reduction rates while requiring only a single fine-tuning epoch, which is an order of magnitude faster than previous methods.
Learning self-supervised task progression metrics : a case of cloth folding

Verleysen, Andreas, Biondina, Matthijs, and wyffels, Francis

APPLIED INTELLIGENCE 2023

Abs Biblio

An important challenge for smart manufacturing systems is finding relevant metrics that capture task quality and progression for process monitoring to ensure process reliability and safety. Data-driven process metrics construct features and labels from abundant raw process data, which incurs costs and inaccuracies due to the labelling process. In this work, we circumvent expensive process data labelling by distilling the task intent from video demonstrations. We present a method to express the task intent in the form of a scalar value by aligning a self-supervised learned embedding to a small set of high-quality task demonstrations. We evaluate our method on the challenging case of monitoring the progress of people folding clothing. We demonstrate that our approach effectively learns to represent task progression without manually labelling sub-steps or progress in the videos. Using case-based experiments, we find that our method learns task-relevant features and useful invariances, making it robust to noise, distractors and variations in the task and shirts. The experimental results show that the proposed method can monitor processes in domains where state representation is inherently challenging.
Leveraging FSPMs for Unconventional Computing with Plants

Pieters, Olivier, De Swaef, Tom, Stock, Michiel, and wyffels, Francis

2023

Abs Biblio

This software contains the code and presentation from the conference paper “ Comparing FSPMs using Unconventional Computing Methods”, presented at FSPM2023 in Berlin. The software to run the analysis can be found on [Github](https://github.com/opieters/fspm2023). The YAML-files (‘hydroshoot_environment.yml‘, ‘wheatfspm_environment.yml‘) should be used to create the anaonda environments and reproduce the output CSV files. The files are also included for convenience (‘hydroshoot.zip‘ and ‘WheatFspm.zip‘). The source code is also included here in case the original repositories are no longer available on GitHub. The code for the grass leaf model is not yet available because the research paper has not yet been published. The input files (‘*_meteo.csv‘) is the input meteorological data. The output files all end with ‘_data.csv‘. Import these into the ‘data‘ directory from the GitHub code and you should be able to reproduce the results.
Limitations of audiovisual speech on robots for second language pronunciation learning

Amioka, Saya, Janssens, Ruben, Wolfert, Pieter, Ren, Qiaoqiao, Pinto Bernal, Maria Jose, and Belpaeme, Tony

In HRI ’23 : proceedings of the 2023 ACM/IEEE International Conference on Human-Robot Interaction 2023

Abs Biblio

The perception of audiovisual speech plays an important role in infants’ first language acquisition and continues to be important for language understanding beyond infancy. Beyond that, the perception of speech and congruent lip motion supports language understanding for adults, and it has been suggested that second language learning benefits from audiovisual speech, as it helps learners distinguish speech sounds in the target language. In this paper, we study whether congruent audiovisual speech on a robot facilitates the learning of Japanese pronunciation. 27 native-Dutch speaking participants were trained in Japanese pronunciation by a social robot. The robot demonstrated 30 Japanese words of varying complexity using either congruent audiovisual speech, incongruent visual speech, or computer-generated audiovisual speech. Participants were asked to imitate the robot’s pronunciation, recordings of which were rated by native Japanese speakers. Against expectation, the results showed that congruent audiovisual speech resulted in lower pronunciation performance than low-fidelity or incongruent speech. We show that our learners, being native Dutch speakers, are only very weakly sensitive to audiovisual Japanese speech which possibly explains why learning performance does not seem to benefit from audiovisual speech.
Low-latency classification of social haptic gestures using transformers

Ren, Qiaoqiao, Hou, Yuanbo, and Belpaeme, Tony

In COMPANION OF THE ACM/IEEE INTERNATIONAL CONFERENCE ON HUMAN-ROBOT INTERACTION, HRI 2023 2023

Abs Biblio

Social touch, and its recognition and classification, is increasingly important in human-robot interaction. We present a Transformer-based model trained and evaluated on an open-source dataset. The dataset, the Human-Animal Affective Robot Touch (HAART) dataset, was collected for the 2015 Recognition of Touch Gesture Challenge (RTGC 2015) and contains different haptic actions directed at a robotic animal. The actions are recorded using a multi-resolution pressure sensor. We feed the output, containing the touch type to the Nao robot to make the robot sense the touch type. The proposed transformer-based gesture classification model achieved 72.8% classification accuracy in 2.67 seconds, which outperforms the best-submitted algorithm of the RTGC 2015 which has a test classification accuracy of 70.9 % and needed 8 seconds.
New insights on homeostatic activity-dependent structural plasticity in rate based neural networks

Cazalets, Tanguy, and Dambre, Joni

In 2023 International Conference on Neuromorphic, Natural and Physical Computing, Proceedings 2023

Abs Biblio

We introduce a novel rate-based homeostatic activity-dependent structural plasticity (HADSP) algorithm for echo state networks (ESNs). The algorithm employs solely homeostatic plasticity, yet enables the emergence of principles of Hebbian learning. Our analysis suggest that HADSP is able to generate networks that effectively recombine redundant inputs to improve performances and highlight the role of structural plasticity in influencing network function and organization.
Novel skin prick automated test device contributes to standardization of allergy diagnosis

Seys, S., Gorris, S., Uyttebroek, S., Backaert, W., Jorissen, M., Schrijvers, R., Daems, Rembert, Loeckx, D., Van Gerven, L., and Hellings, Peter

In ALLERGY 2023

Biblio
Personalised socially assistive robot for cardiac rehabilitation : critical reflections on long-term interactions in the real world

Irfan, Bahar, Cespedes, Nathalia, Casas, Jonathan, Senft, Emmanuel, Gutierrez, Luisa F., Rincon-Roncancio, Monica, Cifuentes, Carlos A., Belpaeme, Tony, and Munera, Marcela

USER MODELING AND USER-ADAPTED INTERACTION 2023

Abs Biblio

Lack of motivation and low adherence rates are critical concerns of long-term rehabilitation programmes, such as cardiac rehabilitation. Socially assistive robots are known to be effective in improving motivation in therapy. However, over longer durations, generic and repetitive behaviours by the robot often result in a decrease in motivation and engagement, which can be overcome by personalising the interaction, such as recognising users, addressing them with their name, and providing feedback on their progress and adherence. We carried out a real-world clinical study, lasting 2.5 years with 43 patients to evaluate the effects of using a robot and personalisation in cardiac rehabilitation. Due to dropouts and other factors, 26 patients completed the programme. The results derived from these patients suggest that robots facilitate motivation and adherence, enable prompt detection of critical conditions by clinicians, and improve the cardiovascular functioning of the patients. Personalisation is further beneficial when providing high-intensity training, eliciting and maintaining engagement (as measured through gaze and social interactions) and motivation throughout the programme. However, relying on full autonomy for personalisation in a real-world environment resulted in sensor and user recognition failures, which caused negative user perceptions and lowered the perceived utility of the robot. Nonetheless, personalisation was positively perceived, suggesting that potential drawbacks need to be weighed against various benefits of the personalised interaction.
Photonic reservoir computing for nonlinear equalization of 64-QAM signals with a Kramers-Kronig receiver

Masaad, Sarah, Gooskens, Emmanuel, Sackesyn, Stijn, Dambre, Joni, and Bienstman, Peter

NANOPHOTONICS 2023

Abs Biblio

Photonic reservoirs are machine learning based systems that boast energy efficiency and speediness. Thus they can be deployed as optical processors in fiber communication systems to aid or replace digital signal equalization. In this paper, we simulate the use of a passive photonic reservoir to target nonlinearity-induced errors originating from self-phase modulation in the fiber and from the nonlinear response of the modulator. A 64-level quadrature-amplitude modulated signal is directly detected using the recently proposed Kramers-Kronig (KK) receiver. We train the readout weights by backpropagating through the receiver pipeline, thereby providing extra nonlinearity. Statistically computed bit error rates for fiber lengths of up to 100 km fall below 1 x 10(-3) bit error rate, outperforming an optical feed-forward equalizer as a linear benchmark. This can find applications in inter-datacenter communications that benefit from the hardware simplicity of a KK receiver and the low power and low latency processing of a photonic reservoir.
Photonic reservoir computing for wavelength multiplexed nonlinear fiber distortion mitigation

Gooskens, Emmanuel, Sackesyn, Stijn, Masaad, Sarah, Dambre, Joni, and Bienstman, Peter

In 2023 IEEE SILICON PHOTONICS CONFERENCE, SIPHOTONICS 2023

Abs Biblio

We seek to improve nonlinear fiber distortion mitigation for wavelength multiplexed telecommunications in terms of both processing speed and energy efficiency. We propose a photonic reservoir computing hardware implementation maximizing the chip footprint to processing power ratio by employing a single readout for all wavelengths.
Point cloud classification with ModelNet40 : What is left?

Herrewegen, Jarne, Tourwé, Tom, and wyffels, Francis

In Data-centric Machine Learning Research Workshop at the 40th International Conference on Machine Learning, Proceedings 2023

Biblio
Quantifying forest structure, light, microclimate and carbon cycling in a temperate forest

Terryn, Louise, Calders, Kim, De Frenne, Pieter, Kuyken, Bart, Sanczuk, Pieter, Verbeeck, Hans, Verhelst, Tom, and wyffels, Francis

In SilviLaser Conference 2023, Abstracts 2023

Abs Biblio

European forests are undergoing large-scale changes in structure and species composition due to anthropogenic disturbances, climate change, and other canopy disturbances. Forest canopies across Europe are now opening up due to tree mortality associated with drought, pests, storms, and fire. Insights on how this will affect forest functioning are essential to understanding, predicting and managing our forests. Therefore, we want to better quantify the spatial and temporal relationships between forest structure, light and microclimate. For this purpose, we have set up a state-of-the-art edge-to-core transect in a temperate deciduous forest near Ghent (Aelmoeseneiebos, Gontrode, Belgium). The transect goes from the forest edge to 135 meters into the forest, covering both an oak-beach-dominated zone and an ash-dominated zone characterised by ash dieback. Here, we installed a spatially dense network of light and microclimate sensors every 15 meters, as well as a fiber optic sensing cable for distributed temperature sensing along the transect. The transect includes a 35 m high measuring tower, which also enables measuring light and microclimate along a vertical transect. Additionally, terrestrial laser scanning (TLS) data is collected from the transect monthly during leaf-off and leaf-on conditions. This setup will enable us to quantify the temporal and spatial variation of the microclimate and the forest structure in great detail. The continuous measurements will be complemented with campaign-based observations of (1) spectral components (reflectance, transmission & absorption) of leaves, bark and understorey using an ASD field Spectrometer to facilitate radiative transfer modelling (RTM) and (2) physiological measurements (photosynthesis, chlorophyll fluorescence, etc.) on the dominant tree species and understorey plants. In a further step of the project, we will reconstruct a 3D virtual forest (using the TLS data) that will be used as an input for radiative transfer modelling (RTM, which simulates the interaction between light and forest structure).
Querying a sign language dictionary with videos using dense vector search

De Coster, Mathieu, and Dambre, Joni

In 2023 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING WORKSHOPS, ICASSPW 2023

Abs Biblio

To search for an unknown sign in a sign language dictionary, users typically indicate parameters of the query, e.g., hand shape and signing location. Recent advances in sign language recognition enable video-based sign language dictionary search. In such a system, users can record an unknown sign and retrieve a list of signs that look similar, preferably including the queried sign as one of the top results. We have realized such a system by interpreting it as a dense vector search task. First, we learn a mapping (embedding) from sign videos to a vector space. The dictionary can then be searched by looking for the vectors in this space that are closest to the vector corresponding to the query. We present a proof of concept on a subset of the Flemish Sign Language dictionary. Further research is required to scale up our method to the large vocabularies of entire dictionaries.
Resource-efficient deep learning for computer vision

Bonnaerens, Maxim

2023

Abs Biblio

Deep-learning technieken hebben doorgaans veel rekenkracht en geheugen nodig. Dit proefschrift richt zich op het onderzoek naar efficiënte deep-learning technieken voor computervisie, met als doel het mogelijk te maken om geavanceerde AI-modellen in te zetten op mobiele en embedded apparaten met beperkte rekenkracht. Het bekijkt zowel het compact ontwerpen van neurale netwerkarchitecturen als het comprimeren van bestaande modellen. Specifieke onderzochte technieken zijn de selectie van efficiënte bouwstenen voor het ontwerpproces van neurale netwerken, het reduceren van de invoer pakketten in Vision Transformers, en het verminderen van het aantal ankers in objectdetectiemodellen. De voorgestelde methodes stellen ontwerpers in staat om een afweging te maken tussen nauwkeurigheid en efficiëntie, zodat de juiste balans kan worden gevonden voor de specifieke applicatie en het beschikbare hardwareplatform.
Revisiting proprioceptive sensing for articulated object manipulation

Lips, Thomas, and wyffels, Francis

In ICRA2023 : Embracing Contacts workshop : IEEE International Conference on Robotics and Automation, Proceedings 2023

Biblio
Seamless integration of tactile sensors for cobots

Proesmans, Remko, and wyffels, Francis

In RoboTac 2023 The 5th IEEE/RSJ International Workshop, Proceedings 2023

Abs Biblio

The development of tactile sensing is expected to enhance robotic systems in handling complex objects like deformables or reflective materials. However, readily available industrial grippers generally lack tactile feedback, which has led researchers to develop their own tactile sensors, resulting in a wide range of sensor hardware. Reading data from these sensors poses an integration challenge: either external wires must be routed along the robotic arm, or a wireless processing unit has to be fixed to the robot, increasing its size. We have developed a microcontroller-based sensor readout solution that seamlessly integrates with Robotiq grippers. Our Arduino compatible design takes away a major part of the integration complexity of tactile sensors and can serve as a valuable accelerator of research in the field. Design files and installation instructions can be found at https://github.com/RemkoPr/airo-halberd.
Self-supervised learning for robust object retrieval without human annotations

Herrewegen, Jarne, Tourwé, Tom, and wyffels, Francis

COMPUTERS & GRAPHICS-UK 2023

Abs Biblio

This paper explores the potential of self-supervised learning as an alternative to supervised learning in the context of geometry-based 3D object retrieval. With the ongoing digitalization of many industries, an exponentially increasing number of 3D objects are processed by retrieval systems. In order to support new shapes, modern deep learning-based retrieval systems require retraining. The dominant paradigm for optimizing neural networks in this field is supervised classification training. Supervised learning requires time-consuming and expensive data annotation. Moreover, training neural networks for classification introduces a bias towards the classes in the training data, which is undesirable for retrieval systems encountering unseen object types in the wild. Through extensive experiments, we make a direct comparison between supervised and self-supervised learning on four datasets from three different domains (household, manufacturing and medical). For object classes seen during training, self-supervised and supervised learning are competitive. For unseen classes, self-supervised learning outperforms supervised learning in many cases. We conclude that self-supervised learning provides a powerful tool for circumventing labeling costs and providing more robust retrieval systems.
SignON sign language translation : progress and challenges

Vandeghinste, Vincent, Shterionov, Dimitar, Sisto, Mirella De, Brady, Aoife, De Coster, Mathieu, Leeson, Lorraine, Blat, Josep, Picron, Frankie, Scipioni, Marcello Paolo, Parikh, Aditya, Bosch, Louis, O’Flaherty, John, Dambre, Joni, Rijckaert, Jorn, Vanroy, Bram, Nogales, Victor Ubieto, Gomez, Santiago Egea, Schuurman, Ineke, Labaka, Gorka, Núnez-Marcos, Adrián, Murtagh, Irene, McGill, Euan, and Saggion, Horacio

In Proceedings of the 24th Annual Conference of the European Association for Machine Translation 2023

Abs Biblio

SignON (\urlhttps://signon-project.eu/) is a Horizon 2020 project, running from 2021 until the end of 2023, which addresses the lack of technology and services for the automatic translation between sign languages (SLs) and spoken languages, through an inclusive, human-centric solution, hence contributing to the repertoire of communication media for deaf, hard of hearing (DHH) and hearing individuals. In this paper, we present an update of the status of the project, describing the approaches developed to address the challenges and peculiarities of SL machine translation (SLMT).
Sim2real flower detection towards automated Calendula harvesting

Vierbergen, Wout, Willekens, Axel, Dekeyser, Donald, Cool, Simon, and wyffels, Francis

BIOSYSTEMS ENGINEERING 2023

Biblio
Trends and challenges for sign language recognition with machine learning

Fink, J., De Coster, Mathieu, Dambre, Joni, and Frénay, B.

In ESANN 2023, 31st European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, Proceedings 2023

Abs Biblio

Research in natural language processing has led to the creation of powerful tools for individuals, companies... However, these successes for written languages have not yet affected signed languages (SLs) to the same extent. The creation of similar tools for signed languages would benefit deaf, hard of hearing, and hearing people by making SL content, learning, and communication more accessible for everyone. SL recognition and translation are related to AI, but require collaboration with linguists and stakeholders. This paper describes related challenges from an AI researcher’s point of view and summarizes the state of the art in these domains
UnfoldIR : tactile robotic unfolding of cloth

Proesmans, Remko, Verleysen, Andreas, and wyffels, Francis

IEEE ROBOTICS AND AUTOMATION LETTERS 2023

Abs Biblio

Robotic unfolding of cloth is challenging due to the wide range of textile materials and their ability to deform in unpredictable ways. Previous work has focused almost exclusively on visual feedback to solve this task. We present UnfoldIR ("unfolder"), a dual-arm robotic system relying on infrared (IR) tactile sensing and cloth manipulation heuristics to achieve in-air unfolding of randomly crumpled rectangular textiles by means of edge tracing. The system achieves > 85% coverage on multiple textiles of different sizes and textures. After unfolding, at least three corners are visible in 83.3 up to 94.7% of cases. Given these strong "tactile-only" results, we argue that the fusion of both tactile and visual sensing can bring cloth unfolding to a new level of performance.
Users’ perspectives on value awareness in social robots

Abbo, Giulio Antonio, and Belpaeme, Tony

In Proceedings of the 1st Workshop on Perspectives on Moral Agency in Human-Robot Interaction 2023

Biblio
Variational inference for SDEs driven by fractional noise

Daems, Rembert, Opper, Manfred, Crevecoeur, Guillaume, and Birdal, Tolga

In 37th Conference on Neural Information Processing Systems, NeurIPS 2023, Proceedings 2023

Biblio

2022

’Cool glasses, where did you get them?’ : generating visually grounded conversation starters for human-robot dialogue

Janssens, Ruben, Wolfert, Pieter, Demeester, Thomas, and Belpaeme, Tony

In PROCEEDINGS OF THE 2022 17TH ACM/IEEE INTERNATIONAL CONFERENCE ON HUMAN-ROBOT INTERACTION (HRI ’22) 2022

Abs Biblio

Visually situated language interaction is an important challenge in multi-modal Human-Robot Interaction (HRI). In this context we present a data-driven method to generate situated conversation starters based on visual context. We take visual data about the interactants and generate appropriate greetings for conversational agents in the context of HRI. For this, we constructed a novel open-source data set consisting of 4000 HRI-oriented images of people facing the camera, each augmented by three conversation-starting questions. We compared a baseline retrieval-based model and a generative model. Human evaluation of the models using crowdsourcing shows that the generative model scores best, specifically at correctly referencing visual features. We also investigated how automated metrics can be used as a proxy for human evaluation and found that common automated metrics are a poor substitute for human judgement. Finally, we provide a proof-of-concept demonstrator through an interaction with a Furhat social robot.
A comparative analysis on genome pleiotropy for evolved soft robots

Marzougui, Dries, Biondina, Matthijs, and wyffels, Francis

In Faculty of Engineering and Architecture Research Symposium 2022 (FEARS 2022), Abstracts 2022

Abs Biblio

Biological evolution shapes the body and brain of living creatures together over time. By contrast, in evolutionary robotics, brain-body co-optimization remains challenging. Conflicting mutations cause dissociation between morphology and control, which leads to premature convergence. Recent works have proposed algorithmic modifications to mitigate the impact of conflicting mutations. However, the importance of genetic design remains underexposed. Current approaches are divided between a single, pleiotropic genetic encoding and two isolated encodings representing morphology and control. This design choice is commonly made ad hoc, causing a lack of consistency for practitioners. To standardize this design, we performed a comparative analysis between these two configurations and two previously unexplored alternatives on a soft robot locomotion task.
A comparative analysis on genome pleiotropy for evolved soft robots

Marzougui, Dries, Biondina, Matthijs, and wyffels, Francis

In PROCEEDINGS OF THE 2022 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE COMPANION, GECCO 2022 2022

Abs Biblio

Biological evolution shapes the body and brain of living creatures together over time. By contrast, in evolutionary robotics, the co-optimization of these subsystems remains challenging. Conflicting mutations cause dissociation between morphology and control, which leads to premature convergence. Recent works have proposed algorithmic modifications to mitigate the impact of conflicting mutations. However, the importance of genetic design remains underexposed. Current approaches are divided between a single, pleiotropic genetic encoding and two isolated encodings representing morphology and control. This design choice is commonly made ad hoc, causing a lack of consistency for practitioners. To standardize this design, we performed a comparative analysis between these two configurations on a soft robot locomotion task. Additionally, we incorporated two currently unexplored alternatives that drive these configurations to their logical extremes. Our results demonstrate that pleiotropic representations yield superior performance in fitness and robustness towards premature convergence. Moreover, we showcase the importance of shared structure in the pleiotropic representation of robot morphology and control to achieve this performance gain. These findings provide valuable insights into genetic encoding design, which supply practitioners with a theoretical foundation to pursue efficient brain-body co-optimization.
A corpus for training sign language recognition and translation : the Belgian federal COVID-19 video corpus

Vandeghinste, Vincent, Van Dyck, Bob, De Coster, Mathieu, and Goddefroy, Maud

In 32nd Meeting of Computational Linguistics in The Netherlands, Abstracts 2022

Biblio
A review of evaluation practices of gesture generation in embodied conversational agents

Wolfert, Pieter, Robinson, Nicole, and Belpaeme, Tony

IEEE TRANSACTIONS ON HUMAN-MACHINE SYSTEMS 2022

Abs Biblio

Embodied conversational agents (ECAs) are often designed to produce nonverbal behavior to complement or enhance their verbal communication. One such form of the nonverbal behavior is co-speech gesturing, which involves movements that the agent makes with its arms and hands that are paired with verbal communication. Co-speech gestures for ECAs can be created using different generation methods, divided into rule-based and data-driven processes, with the latter, gaining traction because of the increasing interest from the applied machine learning community. However, reports on gesture generation methods use a variety of evaluation measures, which hinders comparison. To address this, we present a systematic review on co-speech gesture generation methods for iconic, metaphoric, deictic, and beat gestures, including reported evaluation methods. We review 22 studies that have an ECA with a human-like upper body that uses co-speech gesturing in social human-agent interaction. This includes studies that use human participants to evaluate performance. We found most studies use a within-subject design and rely on a form of subjective evaluation, but without a systematic approach. We argue that the field requires more rigorous and uniform tools for co-speech gesture evaluation, and formulate recommendations for empirical evaluation, including standardized phrases and example scenarios to help systematically test generative models across studies. Furthermore, we also propose a checklist that can be used to report relevant information for the evaluation of generative models, as well as to evaluate co-speech gesture use.
A social robot activity for novice programmers

Staey, Zimcke, Gesquière, Natacha, and wyffels, Francis

In RIE 2021, 12th International Conference on Robotics in Education, Proceedings 2022

Abs Biblio

We have developed an activity on social robots for learners of ages 11-14, as part of a broader AI and CT curriculum for secondary education. We thereby address some of the difficulties that teachers experience when teaching physical computing and challenges of CS education such as gender equality. During the project, learners will use a graphical programming language and simulator to design and implement their robot. Afterward, they can transfer their design to a physical robot.
Anchor pruning for object detection

Bonnaerens, Maxim, Freiberger, Matthias, and Dambre, Joni

COMPUTER VISION AND IMAGE UNDERSTANDING 2022

Abs Biblio

This paper proposes anchor pruning for object detection in one-stage anchor-based detectors. While pruning techniques are widely used to reduce the computational cost of convolutional neural networks, they tend to focus on optimizing the backbone networks where often most computations are. In this work we demonstrate an additional pruning technique, specifically for object detection: anchor pruning. With more efficient backbone networks and a growing trend of deploying object detectors on embedded systems where post-processing steps such as non-maximum suppression can be a bottleneck, the impact of the anchors used in the detection head is becoming increasingly more important. In this work, we show that many anchors in the object detection head can be removed without any loss in accuracy. With additional retraining, anchor pruning can even lead to improved accuracy. Extensive experiments on SSD and MS COCO show that the detection head can be made up to 44% more efficient while simultaneously increasing accuracy. Further experiments on RetinaNet and PASCAL VOC show the general effectiveness of our approach. We also introduce ’overanchorized’ models that can be used together with anchor pruning to eliminate hyperparameters related to the initial shape of anchors. Code and models are available at https://github.com/Mxbonn/anchor_pruning.
Assessment of code, which aspects do teachers consider and how are they valued?

Neutens, Tom, Coolsaet, Kris, and wyffels, Francis

ACM TRANSACTIONS ON COMPUTING EDUCATION 2022

Abs Biblio

In many countries, computer programming is becoming an integral part of the secondary school curriculum. However, many teachers, especially in the first years of Flemish secondary school, have limited experience with teaching programming. To improve their knowledge about programming, many different types of professional development programs have been proposed. Nevertheless, these programs mostly focus on technical skills and less on pedagogical skills. One aspect that is often overlooked in these programs is how teachers can assess code. To get insight into what teachers currently value when assessing code, we designed an experiment that analyzes the different aspects teachers consider during the assessment of code. During the experiment, the teachers (N=13) assess a set of programs from five different fictional learners. After the assessment, they participated in a structured interview giving us insight into the assessment process. We evaluated the transcripts of the interviews using deductive thematic analysis using a coding schema defining the different aspects of code that can be assessed. Additionally, we linked the assessment strategies of teachers to their teaching experience. Our results indicate that many teachers are unaware of the different concepts that can be part of the assessment of code which might lead to inaccurate or invalid feedback. Moreover, although our experimental group was too small to draw hard conclusions about the inter case results, our results indicate that the number of concepts considered by teachers seems to increase with experience. These results provide an initial insight into the code assessment practices of teachers and reveals interesting pathways for future research into the assessment of code.
Attention analysis of a sign language recognition task on the AUTSL dataset

Coppin, Jeanne, De Coster, Mathieu, and Dambre, Joni

In 32nd Meeting of Computational Linguistics in The Netherlands, Abstracts 2022

Biblio
BeCoS corpus : Belgian Covid-19 Sign language corpus : a corpus for training sign language recognition and translation

Vandeghinste, Vincent, Van Dyck, Bob, De Coster, Mathieu, Goddefroy, Maud, and Dambre, Joni

COMPUTATIONAL LINGUISTICS IN THE NETHERLANDS JOURNAL 2022

Abs Biblio

We are presenting the Belgian Federal COVID-19 corpus, nicknamed the BeCoS (Belgian Covid Sign language) corpus. It consists of the entire archive of official press conferences from the Belgian Federal Government concerning the COVID-19 pandemic. The speakers speak mostly in Dutch or French and occasionally in German, and nearly all speech is accompanied by a deaf signer who performs live interpreting from what is being said. We have preprocessed the corpus with speaker diarisation, applied Belgian Dutch ASR, and post-ASR language identification and punctuation prediction as well as signer diarisation, sign language identification and sign language keypoint recognition. The corpus is made publicly available.
Challenges with sign language datasets for sign language recognition and translation

De Sisto, Mirella, Vandeghinste, Vincent, Gomez, Santiago Egea, De Coster, Mathieu, Shterionov, Dimitar, and Saggion, Horacio

In PROCEEDINGS OF THE 13TH INTERNATIONAL CONFERENCE ON LANGUAGE RESOURCES AND EVALUATION (LREC 2022) 2022

Abs Biblio

Sign Languages (SLs) are the primary means of communication for at least half a million people in Europe alone. However, the development of SL recognition and translation tools is slowed down by a series of obstacles concerning resource scarcity and standardization issues in the available data. The former challenge relates to the volume of data available for machine learning as well as the time required to collect and process new data. The latter obstacle is linked to the variety of the data, i.e., annotation formats are not unified and vary amongst different resources. The available data formats are often not suitable for machine learning, obstructing the provision of automatic tools based on neural models. In the present paper, we give an overview of these challenges by comparing various SL corpora and SL machine learning datasets. Furthermore, we propose a framework to address the lack of standardization at format level, unify the available resources and facilitate SL research for different languages. Our framework takes ELAN files as inputs and returns textual and visual data ready to train SL recognition and translation models. We present a proof of concept, training neural translation models on the data produced by the proposed framework.
Computational thinking in Flanders’ compulsory education

Gesquière, Natacha, and wyffels, Francis

In Proceedings of Sixth APSCE International Conference on Computational Thinking and STEM Education 2022 (CTE-STEM) 2022

Abs Biblio

To modernise education, the Flemish government defined new learning goals that take account of 21st-century competences, in particular on â€˜digital competence and media literacyâ€™, of which â€˜computational thinking and actingâ€™ is one of the building blocks. Since September 2019, â€˜computational thinking and actingâ€™ has been compulsory in secondary schools in Flanders. The basic concepts decomposition, abstraction, pattern recognition and generalisation, and algorithm have been pushed forward. A closer look at the newly defined learning goals clarified that â€˜actingâ€™ is about basic knowledge in computer science and computational thinking practices. The learning objectives show that â€˜computational thinking and actingâ€™ is best addressed interdisciplinary in a socially relevant context. Based on the abundant scientific literature on the subject, we found these goals to fit into an international perspective. To support teachers, we are adjusting the teaching materials we already developed on physical computing, programming, and AI
Distance-based delays in echo state networks

Iacob, Stefan-Teodor, Freiberger, Matthias, and Dambre, Joni

In Intelligent Data Engineering and Automated Learning (IDEAL 2022) 2022

Abs Biblio

Physical reservoir computing, a paradigm bearing the promise of energy-efficient high-performance computing, has raised much attention in recent years. We argue though, that the effect of signal propagation delay on reservoir task performance, one of the most central aspects of physical reservoirs, is still insufficiently understood in a more general learning context. Such physically imposed delay has been found to play a crucial role in some specific physical realizations, such as integrated photonic reservoirs. While delays at the readout layer and input of Echo State Networks (ESNs) have been successfully exploited before to improve performance, to our knowledge this feature has not been studied in a more general setting. We introduce inter-node delays, based on physical distances, into ESNs as model systems for physical reservoir computing. We propose a novel ESN design that includes variable signal delays along the connections between neurons, comparable to varying axon lengths in biological neural networks or varying length delay lines in physical systems. We study the impact of the resulting variable inter-node delays in this setup in comparison with conventional ESNs and find that incorporating variable delays significantly improves reservoir performance on the NARMA-10 benchmark task.
Editorial : robots for learning

Johal, Wafa, Belpaeme, Tony, and Chetouani, Mohamed

2022

Biblio
Effective cloth folding trajectories in simulation with only two parameters

De Gusseme, Victor-Louis, and wyffels, Francis

FRONTIERS IN NEUROROBOTICS 2022

Abs Biblio

Robotic cloth folding remains challenging for robots due to its highly deformable nature. In order to deal with these deformations, several strategies with varying amounts of adaptability have been proposed. We study robotic cloth folding by simulating and evaluating a trajectory search space with only two parameters: one parameter for the trajectory’s height and one parameter to tilt it. We extensively analyzed folding a long-sleeved shirt in a high-fidelity simulator. To demonstrate that the trajectory is sufficiently adaptable and robust, we test several cloth shapes, cloth materials, an entire folding sequence and different folding velocities. We can deal with every folding scenario by tuning the two parameters correctly. The trajectories’ simplicity and their robustness in simulation make them ideal candidates for future transfer to real-world robotic setups.
Event-based vision for improved classification accuracy in label-free flow cytometry

Gouda, Muhammed Gouda Ahmed, Lugnan, Alessio, Dambre, Joni, Branden, Gerd, Posch, Christoph, and Bienstman, Peter

In IEEE Benelux Photonics Chapter - Annual Symposium 2022 2022

Biblio
Evolutionary co-optimisation of robot morphology and control : toward a seahorse-tail inspired robotic manipulator

Marzougui, Dries, Adriaens, Dominique, and wyffels, Francis

In SEB Annual Conference 2022, Abstracts 2022

Abs Biblio

The design of robotic manipulators is confronted with a seemingly unavoidable trade-off between the level of flexibility versus strength. Unsurprisingly, nature has already come up with a solution and encapsulated it in a particular group of organisms: the seahorses. A seahorse’s body is completely enclosed in highly articulated body armour, made of similar and modular bony plates. Yet, the combination of regional variation in this skeletal anatomy and the soft tissue interconnecting the skeletal units provides a rigid, yet controllable and flexible tail. A seahorse tail thereby intriguingly integrates these two seemingly mutually exclusive properties. Although the seahorse can serve as a bio-inspired foundation for designing a novel type of robotic manipulator, its actual design is far from intuitive and hard to do manually. This raises the need for automated design methodologies. One domain, named Evolutionary Robotics, applies evolution as an optimisation technique to provide a holistic perspective to automated robot design. Evolutionary Robotics has already shown its potential in conventional rigid robotics and provided the necessary backbone for the optimisation of more recent alternatives such as soft robotics. Our work lies at the intersection of biology and robotics, as we aim to create a generic evolutionary brain-body co-optimisation framework. Using this framework, we pursue the automated design of a seahorse-tail-inspired robotic manipulator. From there on, the framework maintains applicability to other biomimetic experiments, next to providing a testbed for in-silico evolutionary experiments in biology.
GENEA Workshop 2022 : the 3rd Workshop on Generation and Evaluation of Non-verbal Behaviour for Embodied Agents

Wolfert, Pieter, Kucherenko, Taras, Yoon, Youngwoo, Viegas, Carla, Yumak, Zerrin, and Henter, Gustav Eje

In Proceedings of the 2022 International Conference on Multimodal Interaction, ICMI 2022 2022

Abs Biblio

Embodied agents benefit from using non-verbal behavior when communicating with humans. Despite several decades of non-verbal behavior-generation research, there is currently no well-developed benchmarking culture in the field. For example, most researchers do not compare their outcomes with previous work, and if they do, they often do so in their own way which frequently is incompatible with others. With the GENEA Workshop 2022, we aim to bring the community together to discuss key challenges and solutions, and find the most appropriate ways to move the field forward.
Hardware-aware mobile building block evaluation for computer vision

Bonnaerens, Maxim, Freiberger, Matthias, Verhelst, Marian, and Dambre, Joni

APPLIED SCIENCES-BASEL 2022

Abs Biblio

In this paper, we propose a methodology to accurately evaluate and compare the performance of efficient neural network building blocks for computer vision in a hardware-aware manner. Our comparison uses pareto fronts based on randomly sampled networks from a design space to capture the underlying accuracy/complexity trade-offs. We show that our approach enables matching of information obtained by previous comparison paradigms, but provides more insights into the relationship between hardware cost and accuracy. We use our methodology to analyze different building blocks and evaluate their performance on a range of embedded hardware platforms. This highlights the importance of benchmarking building blocks as a preselection step in the design process of a neural network. We show that choosing the right building block can speed up inference by up to a factor of two on specific hardware ML accelerators.
Have I got the power? Analysing and reporting statistical power in HRI

Bartlett, Madeleine E., Edmunds, C. E. R., Belpaeme, Tony, and Thill, Serge

ACM TRANSACTIONS ON HUMAN-ROBOT INTERACTION 2022

Abs Biblio

This article presents a discussion of the importance of power analyses, providing an overview of when power analyses should be run in the context of the field of Human-Robot Interaction, as well as some examples of how to perform a power analysis. This work was motivated by the observation that the majority of papers published in the proceedings of recent HRI conferences did not report conducting a power analysis; an observation that has concerning implications for many conclusions drawn by these studies. This work is intended to raise awareness and encourage researchers to conduct power analyses when designing research studies using human participants.
Human-robot interaction

Belpaeme, Tony

2022

Biblio
Information processing capacity of a single-node reservoir computer : an experimental evaluation

Vettelschoss, Benedikt, Rohm, Andre, and Soriano, Miguel C.

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2022

Abs Biblio

Physical dynamical systems are able to process information in a nontrivial manner. The machine learning paradigm of reservoir computing (RC) provides a suitable framework for information processing in (analog) dynamical systems. The potential of dynamical systems for RC can be quantitatively characterized by the information processing capacity (IPC) measure. Here, we evaluate the IPC measure of a reservoir computer based on a single-analog nonlinear node coupled with delay. We link the extracted IPC measures to the dynamical regime of the reservoir, reporting an experimentally measured nonlinear memory of up to seventh order. In addition, we find a nonhomogeneous distribution of the linear and nonlinear contributions to the IPC as a function of the system operating conditions. Finally, we unveil the role of noise in the IPC of the analog implementation by performing ad hoc numerical simulations. In this manner, we identify the so-called edge of stability as being the most promising operating condition of the experimental implementation for RC purposes in terms of computational power and noise robustness. Similarly, a strong input drive is shown to have beneficial properties, albeit with a reduced memory depth.
Learning keypoints from synthetic data for robotic cloth folding

Lips, Thomas, De Gusseme, Victor-Louis, and wyffels, Francis

In Faculty of Engineering and Architecture Research Symposium 2022 (FEARS 2022), Abstracts 2022

Abs Biblio

Robotic cloth manipulation is challenging due to its deformability, which makes determining its full state infeasible. However, for cloth folding, it suffices to know the position of a few semantic keypoints. Convolutional neural networks (CNN) can be used to detect these keypoints, but require large amounts of annotated data, which is expensive to collect. To overcome this, we propose to learn these keypoint detectors purely from synthetic data, enabling low-cost data collection. In this paper, we procedurally generate images of towels and use them to train a CNN. We evaluate the performance of this detector for folding towels on a unimanual robot setup and find that the grasp and fold success rates are 77% and 53%, respectively. We conclude that learning keypoint detectors from synthetic data for cloth folding and related tasks is a promising research direction, discuss some failures and relate them to future work. A video of the system, as well as the codebase, more details on the CNN architecture and the training setup can be found at https://github.com/tlpss/workshop-icra-2022-cloth-keypoints
Learning keypoints from synthetic data for robotic cloth folding

Lips, Thomas, De Gusseme, Victor-Louis, and wyffels, Francis

In ICRA 2022 Workshop on Representing and Manipulating Deformable Objects 2022

Abs Biblio

Robotic cloth manipulation is challenging due to its deformability, which makes determining its full state infeasible. However, for cloth folding, it suffices to know the position of a few semantic keypoints. Convolutional neural networks (CNN) can be used to detect these keypoints, but require large amounts of annotated data, which is expensive to collect. To overcome this, we propose to learn these keypoint detectors purely from synthetic data, enabling low-cost data collection. In this paper, we procedurally generate images of towels and use them to train a CNN. We evaluate the performance of this detector for folding towels on a unimanual robot setup and find that the grasp and fold success rates are 77% and 53%, respectively. We conclude that learning keypoint detectors from synthetic data for cloth folding and related tasks is a promising research direction, discuss some failures and relate them to future work. A video of the system, as well as the codebase, more details on the CNN architecture and the training setup can be found at https://github.com/tlpss/workshop-icra-2022-cloth-keypoints.git.
Learning robotic cloth manipulation

Verleysen, Andreas

2022

Abs Biblio

Endowing robots with dexterous manipulation skills could spur economic welfare, create leisure time in society, reduce harmful manual labour, and provide care for an ageing population. However, while robots are producing our cars, we are still left to our own devices for doing the laundry at home. This shortcoming is due to the major difficulties in perceiving and handling the variability the real world presents. Robots in modern manufacturing require engineers to produce a safe and predictable environment: objects arrive at the same location, in the same orientation, and the robot is preprogrammed to perform a specific manipulation. Unfortunately, the need for a predictable environment is undesirable in dynamic environments that handle a wide range of objects, often in the presence of human activity. For example, should a human worker first unfold all clothes such that a robot can easily find the corner points and perform folding? Indeed, the high variability in modern production environments and households requires robots to handle objects that can take arbitrary shapes, weights, and configurations. This diversity renders traditional robotic control algorithms and grippers unsuitable for deployment in dynamic environments. To find methods that can handle the ever-changing nature of human environments, we study the perception and manipulation of objects that provide an infinite amount of variations: deformable objects. A deformable object changes shape on force interaction. Deformable objects are omnipresent in industry and society: food, paper, clothes, fruit, cables and sutures, among others. In particular, we study the task of automating the folding of clothes. Folding clothes is a common household task that will potentially be performed by service robots in the future. Handling cloth is also relevant in manufacturing, where technical textile is processed, and in the fashion industry. Dealing with the deformable nature of textiles requires fundamental improvements in both hardware and software. Mechanical engineering needs to incorporate actuators, links, joints and sensors into the limited space of a hand while using soft materials similar to the human skin. In addition to engineering more capable hands, control algorithms need to loosen assumptions about the environment in which robots operate. It is unattainable to expect highly deformable objects like cloth to always be in the same configuration before manipulating them with a robot. A solution for dealing with real-world variability can be found in the machine learning field. In particular, deep RLcombines the function approximation capabilities of deep neural networks with the learn by trial-and-error formalism provided by RL. Deep RL has shown to be capable of driving cars, flying helicopters and manipulating rigid objects. However, the data requirements for training highly parameterized functions, like neural networks, are considerable. This data-hungry property causes an incongruity between the representational learning features of deep neural networks and the high cost of generating real robotic trials. Our research focuses on reducing the required learning data for systems that perceive and manipulate clothing items. We implement a cloth simulation method to generate synthetic data, utilize smart textiles for state estimation of cloth, crowdsource a dataset of people folding clothing, and propose a method to estimate how well people are folding clothing without providing labels. Actuating a physical robot is slow, expensive and potentially dangerous. For this reason, roboticists resort to physics simulators that simulate the robot’s and environment’s dynamics. However, there exists no integrated robot and cloth simulator for use in learning experiments. Cloth simulators are built for offline render farms in the film industry, or for the game industry that sacrifices fidelity for real-time rendering. Unfortunately, cloth simulation for robotic learning requires performance characteristics similar to online rendering and accuracy aspects found in offline rendering. For this reason, we implement a custom cloth dynamics simulation on GPU and integrate it in the robotic simulation functionality of the Unity game engine. We found that we can utilize deep RL to train an agent in our simulation to fold a rectangular piece of cloth twice within 24 hours of wall time on standard computational hardware. The developed cloth simulation assumes full accessibility to the state of the cloth. However, state estimation of cloth in the real world relies on complex vision-based pipelines or high-cost sensing technology. We avoid this complexity and cost by integrating inexpensive tactile sensing technology into a cloth. The cloth becomes an active smart cloth by training a classifier that uses the tactile sensing data to estimate its state. We use this smart cloth to train a low-cost robotic platform to fold the cloth using RL. Our results demonstrate that it is possible to develop a smart cloth with off-the-shelf components and use it effectively for training on a real robotic platform. Our smart cloth bridges the gap between our cloth simulation on GPU and state estimation in the real world. However, it is still required to distil a scalar value that indicates task progress in order to acquire manipulation skills using RL. We believe that learning the reward function from demonstrations may overcome human bias in reward engineering. Unfortunately, when starting our research there existed no large dataset with people folding clothing. We fill this gap by crowdsourcing a dataset of people folding clothes. Our dataset consists of roughly 300000 multiperspective RGB-D frames, annotated with pose trajectories, quality labels and timestamps indicating substeps. This dataset can be used to benchmark research in action recognition and bootstrap learning by using example demonstrations. Learning from demonstrations is a prevalent domain in the robotics learning community. However, using our cloth folding dataset requires mapping the movements of demonstrators to the embodiment of a robot. Additionally, behavioural cloning is prone to blindly imitating trajectories instead of understanding how actions relate to solving the task. For this reason, estimating how well a process is being executed is preferred to learning the policy from demonstrations. Unfortunately, existing methods couple the learning of rewards with policy learning, thereby inheriting all problems associated with RL. To decouple reward and policy learning, we propose a method to learn the task progression from multiperspective videos of example demonstrations. We avoid incorporating human bias in the labelling process by using time as a self-supervised signal for learning. We demonstrate the first results on expressing task progression of people folding clothing, without labelling any data. General-purpose robots are not yet among us. Robots that are capable of working in dynamic environments will require a holistic view of software and hardware. We demonstrated the benefits of this approach by outsourcing the intelligence for state estimation to the cloth instead of the robot. By developing a smart cloth, we trained a robot to fold cloth in-vivo within a day. Extrapolating an integrated approach on hardware and software leads to embodied intelligence in which morphology closes the loop with control: co-optimizing the body and brain will allow evolving manipulators tailored to the tasks, and use them to build a representation of how the world works. Robots can use that feedback to understand how actions influence the environment and learn to solve tasks by using human examples, instrumented objects and their own experiences. This holistic process will enable future robots to understand human intent and solve a large repertoire of manipulation tasks.
Leveraging frozen pretrained written language models for neural sign language translation

De Coster, Mathieu, and Dambre, Joni

INFORMATION 2022

Abs Biblio

We consider neural sign language translation: machine translation from signed to written languages using encoder–decoder neural networks. Translating sign language videos to written language text is especially complex because of the difference in modality between source and target language and, consequently, the required video processing. At the same time, sign languages are low-resource languages, their datasets dwarfed by those available for written languages. Recent advances in written language processing and success stories of transfer learning raise the question of how pretrained written language models can be leveraged to improve sign language translation. We apply the Frozen Pretrained Transformer (FPT) technique to initialize the encoder, decoder, or both, of a sign language translation model with parts of a pretrained written language model. We observe that the attention patterns transfer in zero-shot to the different modality and, in some experiments, we obtain higher scores (from 18.85 to 21.39 BLEU-4). Especially when gloss annotations are unavailable, FPTs can increase performance on unseen data. However, current models appear to be limited primarily by data quality and only then by data quantity, limiting potential gains with FPTs. Therefore, in further research, we will focus on improving the representations used as inputs to translation models.
Leveraging plant physiological dynamics using physical reservoir computing

Pieters, Olivier, De Swaef, Tom, Stock, Michiel, and wyffels, Francis

SCIENTIFIC REPORTS 2022

Abs Biblio

Plants are complex organisms subject to variable environmental conditions, which influence their physiology and phenotype dynamically. We propose to interpret plants as reservoirs in physical reservoir computing. The physical reservoir computing paradigm originates from computer science; instead of relying on Boolean circuits to perform computations, any substrate that exhibits complex non-linear and temporal dynamics can serve as a computing element. Here, we present the first application of physical reservoir computing with plants. In addition to investigating classical benchmark tasks, we show that Fragaria x ananassa (strawberry) plants can solve environmental and eco-physiological tasks using only eight leaf thickness sensors. Although the results indicate that plants are not suitable for general-purpose computation but are well-suited for eco-physiological tasks such as photosynthetic rate and transpiration rate. Having the means to investigate the information processing by plants improves quantification and understanding of integrative plant responses to dynamic changes in their environment. This first demonstration of physical reservoir computing with plants is key for transitioning towards a holistic view of phenotyping and early stress detection in precision agriculture applications since physical reservoir computing enables us to analyse plant responses in a general way: environmental changes are processed by plants to optimise their phenotype.
Modeling electronic response properties with an explicit-electron machine learning potential

Cools-Ceuppens, Maarten, Dambre, Joni, and Verstraelen, Toon

JOURNAL OF CHEMICAL THEORY AND COMPUTATION 2022

Abs Biblio

Explicit-electron force fields introduce electrons or electron pairs as semiclassical particles in force fields or empirical potentials, which are suitable for molecular dynamics simulations. Even though semiclassical electrons are a drastic simplification compared to a quantum-mechanical electronic wave function, they still retain a relatively detailed electronic model compared to conventional polarizable and reactive force fields. The ability of explicit-electron models to describe chemical reactions and electronic response properties has already been demonstrated, yet the description of short-range interactions for a broad range of chemical systems remains challenging. In this work, we present the electron machine learning potential (eMLP), a new explicit electron force field in which the short-range interactions are modeled with machine learning. The electron pair particles will be located at well-defined positions, derived from localized molecular orbitals or Wannier centers, naturally imposing the correct dielectric and piezoelectric behavior of the system. The eMLP is benchmarked on two newly constructed data sets: eQM7, an extension of the QM7 data set for small molecules, and a data set for the crystalline beta-glycine. It is shown that the eMLP can predict dipole moments, polarizabilities, and IR-spectra of unseen molecules with high precision. Furthermore, a variety of response properties, for example, stiffness or piezoelectric constants, can be accurately reproduced.
Modular piezoresistive smart textile for state estimation of cloths

Proesmans, Remko, Verleysen, Andreas, Vleugels, Robbe, Veske-Lepp, Paula, De Gusseme, Victor-Louis, and wyffels, Francis

SENSORS 2022

Abs Biblio

Smart textiles have found numerous applications ranging from health monitoring to smart homes. Their main allure is their flexibility, which allows for seamless integration of sensing in everyday objects like clothing. The application domain also includes robotics; smart textiles have been used to improve human-robot interaction, to solve the problem of state estimation of soft robots, and for state estimation to enable learning of robotic manipulation of textiles. The latter application provides an alternative to computationally expensive vision-based pipelines and we believe it is the key to accelerate robotic learning of textile manipulation. Current smart textiles, however, maintain wired connections to external units, which impedes robotic manipulation, and lack modularity to facilitate state estimation of large cloths. In this work, we propose an open-source, fully wireless, highly flexible, light, and modular version of a piezoresistive smart textile. Its output stability was experimentally quantified and determined to be sufficient for classification tasks. Its functionality as a state sensor for larger cloths was also verified in a classification task where two of the smart textiles were sewn onto a piece of clothing of which three states are defined. The modular smart textile system was able to recognize these states with average per-class F1-scores ranging from 85.7 to 94.6% with a basic linear classifier.
Multi-modal open world user identification

Irfan, Bahar, Ortiz, Michael Garcia, Lyubova, Natalia, and Belpaeme, Tony

ACM TRANSACTIONS ON HUMAN-ROBOT INTERACTION 2022

Abs Biblio

User identification is an essential step in creating a personalised long-term interaction with robots. This requires learning the users continuously and incrementally, possibly starting from a state without any known user. In this article, we describe a multi-modal incremental Bayesian network with online learning, which is the first method that can be applied in such scenarios. Face recognition is used as the primary biometric, and it is combined with ancillary information, such as gender, age, height, and time of interaction to improve the recognition. The Multi-modal Long-term User Recognition Dataset is generated to simulate various human-robot interaction (HRI) scenarios and evaluate our approach in comparison to face recognition, soft biometrics, and a state-of-the-art open world recognition method (Extreme Value Machine). The results show that the proposed methods significantly outperform the baselines, with an increase in the identification rate up to 47.9% in open-set and closed-set scenarios, and a significant decrease in long-term recognition performance loss. The proposed models generalise well to new users, provide stability, improve over time, and decrease the bias of face recognition. The models were applied in HRI studies for user recognition, personalised rehabilitation, and customer-oriented service, which showed that they are suitable for long-term HRI in the real world.
On the pivotal role of water potential to model plant physiological processes

De Swaef, Tom, Pieters, Olivier, Appeltans, Simon, Borra-Serrano, Irene, Coudron, Willem, Couvreur, Valentin, Garré, Sarah, Lootens, Peter, Nicolaï, Bart, Pols, Leroi, Saint Cast, Clément, Šalagovič, Jakub, Van Haeverbeke, Maxime, Stock, Michiel, and wyffels, Francis

IN SILICO PLANTS 2022

Abs Biblio

Water potential explains water transport in the Soil-Plant-Atmosphere Continuum (SPAC), and is gaining interest as connecting variable between ‘pedo-, bio- and atmosphere’. It is primarily used to simulate hydraulics in the SPAC, and is thus essential for studying drought effects. Recent implementations of hydraulics in large-scale Terrestrial Biosphere Models (TBMs) improved their performance under water-limited conditions, while hydraulic features of recent detailed Functional-Structural Plant Models (FSPMs) open new possibilities for dissecting complex traits for drought tolerance. These developments in models across scales deserve a critical appraisal to evaluate its potential for wider use in FSPMs, but also in crop systems models (CSMs), where hydraulics are currently still absent. After refreshing the physical basis, we first address models where water potential is primarily used for describing water transport along the transpiration pathway from the soil to the leaves, through the roots, the xylem and the leaf mesophyll. Then, we highlight models for three ecophysiological processes, which have well-recognised links to water potential: phloem transport, stomatal conductance and organ growth. We identify water potential as the bridge between soil, root and shoot models, as the physiological variable integrating below- and above-ground abiotic drivers, but also as the link between water status and growth. Models making these connections enable identifying crucial traits for ecosystem resilience to drought and for breeding towards improved drought tolerance in crops. Including hydraulics often increases model complexity, and thus requires experimental data on soil and plant hydraulics. Nevertheless, modelling hydraulics is insightful at different scales (FSPMs, CSMs and TBMs).
Photonic reservoir computing for nonlinear equalization of 64-QAM signals with a Kramers-Kronig receiver

Masaad, Sarah, Gooskens, Emmanuel, Sackesyn, Stijn, Dambre, Joni, and Bienstman, Peter

In 2022 EUROPEAN CONFERENCE ON OPTICAL COMMUNICATION (ECOC) 2022

Abs Biblio

Photonic reservoir computing is a promising processing solution for the equalization of fiber optic communication signals. We simulate the nonlinear equalization of 64 Quadrature-Amplitude Modulated signals using a fully passive space multiplexed reservoir. The system deploys direct detection using the recently proposed Kramers-Kronig receiver. (C) 2022 The Author(s)
Physical computing in primary and secondary education : techniques and tools for teaching and assessment

Neutens, Tom

2022

Abs Biblio

Computer science is becoming an increasingly important part of primary and secondary education. This change is backed by a societal shift towards a digital economy. The European Commis- sion wants 100% of key public services to be online and 90% of small- and medium-sized businesses to reach at least a basic level of digital intensity. It is clear that the integration of computer science in education is important in order to reach these goals. Nevertheless, this is not a trivial task. There are many unknowns about how to do this effectively, moreover, these unknowns are distributed across multiple areas of policy and practice. In this dissertation, we identify some of these unknowns and formulate some possible solutions. We mainly focus on supporting teach- ing practice but also formulate insights which can be helpful for making policy decisions. Since the field of computing education is very broad, we focus on the integration of programming in a physical computing context for learners ages 10 to 14. To conduct research in this field, we first created a visual programming environment for physical systems. This environment is based on existing literature and is used as a tool throughout this work. It allows learners to build and program robots both in the real world and in simulation. This strong link between the physical and virtual word allows the environment to be used in many different learning contexts, allowing us to conduct a variety of in classroom experiments. Using the environment we created, we have shown that not all ways of integrating programming into a physical computing class are equally effective. Specifically, our results show that having learners create a program from scratch is preferable to having them find errors in existing programs. Our metrics indicate that finding errors in an existing program has more risk of causing ex- traneous cognitive load in learners. Cognitive load theory states that the human mind has a limited working memory. In order for effective learning to take place, the capacity of the working memory should not be exceeded. Even though fixing a program only required learners to change one element in the code, the act of understanding what the code does, applying error detection strategies, and deducing what they can do to fix the issue, causes more cognitive load than constructing a fully functional program from scratch. As a result, learners in the fix group get stuck more often, lose motivation more quickly, tend to use trial-and-error solution strategies more often, and score lower on programming tests. Since effective teaching not only depends on the teaching method but also on the teacher, we have developed methods for support- ing teachers in their efforts to integrate programming into their teaching. To identify the most pressing needs teachers have, we set up a qualitative experiment during one of the professional de- velopment programs we created. By observing and interviewing teachers, we identified two main challenges teachers encounter. 1) The ability and willingness to create, reuse and share edu- cational materials. 2) How to assess the coding ability of their learners especially in an open learning context. We aimed to address these issues by designing and developing different tools and techniques for supporting teachers. Our ef- forts include the design and implementation of a learning object repository, a tool for sharing learning content on the web, as well as a framework for the assessment of code. The learning object repository defines a unified way of creating and sharing educational content. It solves the issues of granularity, heterogene- ity, reuse, personalization, and sharing. We argue that having a uniform system for defining learning content is important to facilitate personalized learning. The assessment framework we developed can be used by teach- ers to get insight into the aspects of code that can be assessed, giving them a grip on how to give quality feedback. It focusses on feedback of visual programs to learners ages 12 to 14. The teachers for these age groups need the most support since they usually have to teach multiple subjects and are less of an expert in each of the topics they teach. To develop this framework, we identified the gap between the existing literature about the assessment of code and how teachers perform the assessment in practice. Consequently, its strength is the link between theory and practice. The framework consists of eight main categories: 1) functionality 2) readability 3) programming concepts 4) algo- rithmic thinking 5) testing and debugging 6) mathematical con- cepts 7) creativity 8) personal aspects. Each of these categories is then split up into concepts relevant in the context of visual programming. Comparing this framework to what teachers as- sess in practice has shown that teachers are mostly unaware of these categories. This often results in incomplete, inaccurate, or invalid feedback. Additionally, the range of concepts being considered by teachers seems to grow with teaching experience. Nevertheless, even more experienced teachers are often not fully aware of which exact concepts they are assessing. Finally, we describe a method to assist teachers with their assess- ment of code by visualizing the solution paths of learners. This requires techniques to extract educationally relevant informa- tion from code. Consequently, we developed different machine learning based techniques for extracting didactical information from code and for visualizing solution paths in our programming environment. Our system extracts structural and functional in- formation from programs and embeds it into a vector space. The structural information is extracted using a Gaussian vertex edge histogram kernel. This kernel defines the structural similarity between two abstract syntax trees which can be used to cluster similar programs together. To extract functional information, we executed the programs in a simulated environment and captured the internal program state at each time step. By transforming these state variations in time into both frequency and amplitude components, we are able to create a vector representing the func- tion of a program. These vectors can be used as input for several machine learning algorithms. This allows us to cluster similar programs together and create visualizations based on the similarity of solution paths. This proof of concept application lays the groundwork for automated feedback systems.
Progress, challenges and innovations of the SignON project

Vandeghinste, Vincent, De Sisto, Mirella, Shterionov, Dimitar, Brady, Aoife, De Coster, Mathieu, Leeson, Lorraine, Blat, Josep, Picron, Frankie, Scipioni, Marcello, Parikh, Aditya, Bosch, Louis, O’Flaherty, John, Dambre, Joni, and Rijckaert, Jorn

In 32nd Meeting of Computational Linguistics in The Netherlands, Abstracts 2022

Biblio
Reservoir computing with plants

Pieters, Olivier

2022

Abs Biblio

Planten zijn complexe organismen, onderhevig aan een groot aantal omgevingsvariabelen, dewelke op hun beurt de fysiologie en het fenotype van de plant beïnvloeden. In ons werk wordt dit complex ingangsgedreven systeem beschouwd als een reservoir in physical reservoir computing (PRC). PRC is een paradigma uit de computerwetenschappen dat een fysisch substraat gebruikt als rekenmedium. In een eerste luik van het doctoraat werd de toepasbaarheid van hyperspectrale camera’s onderzocht. Dergelijke camera’s blijken niet geschikt te zijn om subtiele spectrale veranderingen, nodig voor PRC, te detecteren. Daarop werd een nieuw sensor platform ontwikkeld, specifiek ontworpen om contactsensoren uit te lezen. Aan de hand van bladdikte sensoren toonden we aan dat planten inderdaad als reservoir gebruikt kunnen worden voor PRC. Planten waren performanter dan een controle substraat in het oplossen van eco-fysiologische- en omgevingstaken. We onderzochten ook controle taken zoals NARMA en een vertragingslijn. Deze resultaten geven aan dat planten niet geschikt zijn om niet- plantspecifieke taken op te lossen, maar wel goed presteren in plant- gerelateerde taken. Deze eerste experimentele demonstratie van PRC met planten is een belangrijke mijlpaal naar een meer holistische kijk op fenotypering en informatieverwerking door planten.
Sign language translation : ongoing development, challenges and innovations in the SignON project

Shterionov, Dimitar, De Sisto, Mirella, Vandeghinste, Vincent, Brady, Aoife, De Coster, Mathieu, Leeson, Lorraine, Blat, Josep, Picron, Frankie, Scipioni, Marcello, Parikh, Aditya, Bosch, Louis, O’Flaherty, John, Dambre, Joni, and Rijckaert, Jorn

In Proceedings of the 23rd Annual Conference of the European Association for Machine Translation 2022

Abs Biblio

The SignON project (www.signon-project.eu) focuses on the research and development of a Sign Language (SL) translation mobile application and an open communications framework. SignON rectifies the lack of technology and services for the automatic translation between signed and spoken languages, through an inclusive, humancentric solution which facilitates communication between deaf, hard of hearing (DHH) and hearing individuals. We present an overview of the current status of the project, describing the milestones reached to date and the approaches that are being developed to address the challenges and peculiarities of Sign Language Machine Translation (SLMT).
Tactile interaction with a robot leads to increased risk-taking

Ren, Qiaoqiao, and Belpaeme, Tony

In SOCIAL ROBOTICS, ICSR 2022, PT1 2022

Abs Biblio

Tactile interaction plays a crucial role in interactions between people. Touch can, for example, help people calm down and lower physiological stress responses. Consequently, it is believed that tactile and haptic interaction matter also in human-robot interaction. We study if the intensity of the tactile interaction has an impact on people, and do so by studying whether different intensities of tactile interaction modulate physiological measures and task performance. We use a paradigm in which a small humanoid robot is used to encourage risk-taking behaviour, relying on peer encouragement to take more risks which might lead to a higher pay-off, but potentially also to higher losses. For this, the Balloon Analogue Risk Task (BART) is used as a proxy for the propensity to take risks. We study four conditions, one control condition in which the task is completed without a robot, and three experimental conditions in which a robot is present that encourages risk-taking behaviour with different degrees of tactile interaction. The results show that both low-intensity and high-intensity tactile interaction increase people’s risk-taking behaviour. However, low-intensity tactile interaction increases comfort and lowers stress, whereas high-intensity touch does not.
The GENEA Challenge 2022 : a large evaluation of data-driven co-speech gesture generation

Yoon, Youngwoo, Wolfert, Pieter, Kucherenko, Taras, Viegas, Carla, Nikolov, Teodor, Tsakov, Mihail, and Henter, Gustav Eje

In PROCEEDINGS OF THE 2022 INTERNATIONAL CONFERENCE ON MULTIMODAL INTERACTION, ICMI 2022 2022

Abs Biblio

This paper reports on the second GENEA Challenge to benchmark data-driven automatic co-speech gesture generation. Participating teams used the same speech and motion dataset to build gesture generation systems. Motion generated by all these systems was rendered to video using a standardised visualisation pipeline and evaluated in several large, crowdsourced user studies. Unlike when comparing diferent research papers, differences in results are here only due to differences between methods, enabling direct comparison between systems. This year’s dataset was based on 18 hours of full-body motion capture, including fingers, of diferent persons engaging in dyadic conversation. Ten teams participated in the challenge across two tiers: full-body and upper-body gesticulation. For each tier we evaluated both the human-likeness of the gesture motion and its appropriateness for the specific speech signal. Our evaluations decouple human-likeness from gesture appropriateness, which previously was a major challenge in the field. The evaluation results are a revolution, and a revelation. Some synthetic conditions are rated as significantly more human-like than human motion capture. To the best of our knowledge, this has never been shown before on a high-fidelity avatar. On the other hand, all synthetic motion is found to be vastly less appropriate for the speech than the original motion-capture recordings.
The effectiveness of dynamically processed incremental descriptions in human robot interaction

Wallbridge, Christopher D., Smith, Alex, Giuliani, Manuel, Melhuish, Chris, Belpaeme, Tony, and Lemaignan, Séverin

ACM TRANSACTIONS ON HUMAN-ROBOT INTERACTION 2022

Abs Biblio

We explore the effectiveness of a dynamically processed incremental referring description system using under-specified ambiguous descriptions that are then built upon using linguistic repair statements, which we refer to as a dynamic system. We build a dynamically processed incremental referring description generation system that is able to provide contextual navigational statements to describe an object in a potential real-world situation of nuclear waste sorting and maintenance. In a study of 31 participants, we test the dynamic system in a case where a user is remote operating a robot to sort nuclear waste, with the robot assisting them in identifying the correct barrels to be removed. We compare these against a static non-ambiguous description given in the same scenario. As well as looking at efficiency with time and distance measurements, we also look at user preference. Results show that our dynamic system was a much more efficient method—taking only 62% of the time on average—for finding the correct barrel. Participants also favoured our dynamic system.
Update logic synthesis objectives for better placement and routing

Louage, Marieke, and Stroobandt, Dirk

In Faculty of Engineering and Architecture Research Symposium 2022 (FEARS 2022), Abstracts 2022

Abs Biblio

This research is situated in the design of integrated circuits (ICs). ICs are virtually everywhere. They are physical implementations of digital circuits that perform some logical functionality. The steps involved in IC design are twofold. First in the Logic Synthesis step, an abstract representation of the logic functionality is designed in the form of a graph called a netlist. Second, this netlist needs to be further developed into a physically realizable model in a step called Physical Design (Placement and Routing). ICs consist of two main components: logic cells and interconnects. Logic cells are responsible for the data manipulation and interconnects transport the data between the logic cells. Both logic cells and interconnects amount to some delay that ultimately add up to determine the speed of the IC. Back in 1980, most of the delay was accounted for by the logic cells and interconnect delay was negligible. A consequence of this early imbalance is that the netlist is optimized to have as little logic cells as possible. However, as IC technology progressed, the delay in logic cells decreased significantly and the delay in interconnects became more important. Today, the speed of an IC is considerably impacted by lengthy interconnects. Despite this shift in delay causes, the general objective of minimizing the amount of logic cells has not changed. In this poster, we argue that the netlist should be optimized for different, yet unknown, objectives. This is done by giving a simple example that shows a discrepancy between minimizing the amount of logic cells and optimizing speed.
Visual conversation starters for human-robot interaction

Janssens, Ruben, Demeester, Thomas, and Belpaeme, Tony

In BNAIC/BeNeLearn 2022 : joint International Scientific Conferences on AI and Machine Learning, Abstracts 2022

Biblio
Wavelength dimension in waveguide-based photonic reservoir computing

Gooskens, Emmanuel, Laporte, Floris, Ma, Chonghuai, Sackesyn, Stijn, Dambre, Joni, and Bienstman, Peter

OPTICS EXPRESS 2022

Abs Biblio

Existing work on coherent photonic reservoir computing (PRC) mostly concentrates on single-wavelength solutions. In this paper, we discuss the opportunities and challenges related to exploiting the wavelength dimension in integrated photonic reservoir computing systems. Different strategies are presented to be able to process several wavelengths in parallel using the same readout. Additionally, we present multiwavelength training techniques that allow to increase the stable operating wavelength range by at least a factor of two. It is shown that a single-readout photonic reservoir system can perform with approximate to 0% BER on several WDM channels in parallel for bit-level tasks and nonlinear signal equalization. This even when taking manufacturing deviations and laser wavelength drift into account. (C) 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement

2021

A large, crowdsourced evaluation of gesture generation systems on common data : the GENEA Challenge 2020

Kucherenko, T., Jonell, P., Yoon, Y., Wolfert, Pieter, and Henter, G. E.

In IUI ’21: 26th International Conference on Intelligent User Interfaces 2021

Abs Biblio

Co-speech gestures, gestures that accompany speech, play an important role in human communication. Automatic co-speech gesture generation is thus a key enabling technology for embodied conversational agents (ECAs), since humans expect ECAs to be capable of multi-modal communication. Research into gesture generation is rapidly gravitating towards data-driven methods. Unfortunately, individual research efforts in the field are difficult to compare: there are no established benchmarks, and each study tends to use its own dataset, motion visualisation, and evaluation methodology. To address this situation, we launched the GENEA Challenge, a gesture-generation challenge wherein participating teams built automatic gesture-generation systems on a common dataset, and the resulting systems were evaluated in parallel in a large, crowdsourced user study using the same motion-rendering pipeline. Since differences in evaluation outcomes between systems now are solely attributable to differences between the motion-generation methods, this enables benchmarking recent approaches against one another in order to get a better impression of the state of the art in the field. This paper reports on the purpose, design, results, and implications of our challenge.
A proposal of accessibility guidelines for human-robot interaction

Qbilat, Malak, Iglesias, Ana, and Belpaeme, Tony

ELECTRONICS 2021

Abs Biblio

We will increasingly become dependent on automation to support our manufacturing and daily living, and robots are likely to take an important place in this. Unfortunately, currently not all the robots are accessible for all users. This is due to the different characteristics of users, as users with visual, hearing, motor or cognitive disabilities were not considered during the design, implementation or interaction phase, causing accessibility barriers to users who have limitations. This research presents a proposal for accessibility guidelines for human-robot interaction (HRI). The guidelines have been evaluated by seventeen HRI designers and/or developers. A questionnaire of nine five-point Likert Scale questions and 6 open-ended questions was developed to evaluate the proposed guidelines for developers and designers, in terms of four main factors: usability, social acceptance, user experience and social impact. The questions act as indicators for each factor. The majority (15 of 17 participants) agreed that the guidelines are helpful for them to design and implement accessible robot interfaces and applications. Some of them had considered some ad hoc guidelines in their design practice, but none of them showed awareness of or had applied all the proposed guidelines in their design practice, 72% of the proposed guidelines have been applied by less than or equal to 8 participants for each guideline. Moreover, 16 of 17 participants would use the proposed guidelines in their future robot designs or evaluation. The participants recommended the importance of aligning the proposed guidelines with safety requirements, environment of interaction (indoor or outdoor), cost and users’ expectations.
A socially assistive robot for long-term cardiac rehabilitation in the real world

Cespedes, Nathalia, Irfan, Bahar, Senft, Emmanuel, Cifuentes, Carlos A., Gutierrez, Luisa F., Rincon-Roncancio, Monica, Belpaeme, Tony, and Munera, Marcela

FRONTIERS IN NEUROROBOTICS 2021

Abs Biblio

What are the benefits of using a socially assistive robot for long-term cardiac rehabilitation? To answer this question we designed and conducted a real-world long-term study, in collaboration with medical specialists, at the Fundacion Cardioinfantil-Instituto de Cardiologia clinic (Bogota, Colombia) lasting 2.5 years. The study took place within the context of the outpatient phase of patients’ cardiac rehabilitation programme and aimed to compare the patients’ progress and adherence in the conventional cardiac rehabilitation programme (control condition) against rehabilitation supported by a fully autonomous socially assistive robot which continuously monitored the patients during exercise to provide immediate feedback and motivation based on sensory measures (robot condition). The explicit aim of the social robot is to improve patient motivation and increase adherence to the programme to ensure a complete recovery. We recruited 15 patients per condition. The cardiac rehabilitation programme was designed to last 36 sessions (18 weeks) per patient. The findings suggest that robot increases adherence (by 13.3%) and leads to faster completion of the programme. In addition, the patients assisted by the robot had more rapid improvement in their recovery heart rate, better physical activity performance and a higher improvement in cardiovascular functioning, which indicate a successful cardiac rehabilitation programme performance. Moreover, the medical staff and the patients acknowledged that the robot improved the patient motivation and adherence to the programme, supporting its potential in addressing the major challenges in rehabilitation programmes.
Addressing limited weight resolution in a fully optical neuromorphic reservoir computing readout

Ma, Chonghuai, Laporte, Floris, Dambre, Joni, and Bienstman, Peter

SCIENTIFIC REPORTS 2021

Abs Biblio

Using optical hardware for neuromorphic computing has become more and more popular recently, due to its efficient high-speed data processing capabilities and low power consumption. However, there are still some remaining obstacles to realizing the vision of a completely optical neuromorphic computer. One of them is that, depending on the technology used, optical weighting elements may not share the same resolution as in the electrical domain. Moreover, noise of the weighting elements are important considerations as well. In this article, we investigate a new method for improving the performance of optical weighting components, even in the presence of noise and in the case of very low resolution. Our method utilizes an iterative training procedure and is able to select weight connections that are more robust to quantization and noise. As a result, even with only 8 to 32 levels of resolution, in noisy weighting environments, the method can outperform both nearest rounding low-resolution weighting and random rounding weighting by up to several orders of magnitude in terms of bit error rate and can deliver performance very close to full-resolution weighting elements.
Analysis of attention in child-robot interaction among children diagnosed with cognitive impairment

Ismail, Luthffi, Hanapiah, Fazah Akhtar, Belpaeme, Tony, Dambre, Joni, and wyffels, Francis

INTERNATIONAL JOURNAL OF SOCIAL ROBOTICS 2021

Abs Biblio

Interacting with social robots has been reported as potentially beneficial for children with social communication difficulties, with one of the promising applications being the practising of social skills, such as joint attention. We present the analysis of attention skills in children with cognitive impairments over a series of child-robot interaction sessions. Here, an interaction consists of five different modules. The first module introduces the child to the robot. The next three modules are the task modules during which children are expected to improve their attention skills during the completion of a series of social tasks. The final module is a free style interaction, where the duration of interaction between the child and robot was used as a proxy to indicate the attention of the child towards a robot. Our analysis showed that the majority of the children reduced their task completion time in modules two to four, indicating an improvement in attention. Moreover, most of the children showed positive engagement towards the robot and spent an average of 120 s during the free style interaction in module five. The positive response suggests that the robot, via child-robot interaction could be a useful and engaging tool to improve attention skills of the children with cognitive impairment.
Analyzing coding behaviour of novice programmers in different instructional settings : creating vs. debugging

Neutens, Tom, and wyffels, Francis

In 2020 International Conference on Computational Science and Computational Intelligence (CSCI) 2021

Biblio
Biologically inspired locomotion of compliant robots

Urbain, Gabriel

2021

Abs Biblio

A better understanding of locomotion, and the processes that it involves, has potential benefits in our society. On one hand, it could help to design efficient legged robots, with better accessibil- ity to the different environments on the globe, whereas wheeled platforms remain generally limited on even terrains without ob- stacles. This could be used to build social robots or to deal with exploration and rescue operations in hazardous environments. On the other hand, improving the understanding of the locomotion mechanisms can also contribute to biological sciences and, in par- ticular, neurosciences. In this regard, studies on locomotion are a typical illustration of an embodied problem, which is often cited as a key concept to bridge the different scales in brain research, from the chemical and physical processes to the behavioral and psychological aspects. To improve the agility and adaptability of robotic locomotion plat- forms, an appealing path is to use compliant structures and actua- tors rather than stiff elements. However, compliant and soft robots are not well suited for control with traditional computational ar- chitectures, generally directed towards centralized commands and exactness. New models, driven by data, or inspired by the broad locomotion abilities encountered within biological systems, create an opportunity to improve the state of the art in robotics. But they also raise questions on several fundamental aspects, limiting their maturity and their diffusion in the society. This dissertation tries to better investigate three of these research questions: how can we transfer knowledge from simulation to real robots, how does the mechanical compliance correlate with the locomotion performance and the controller complexity, and, thirdly, how can reflex-based control on compliant structures benefit from a stance correction mechanism taking its inspiration in the biological Cerebellum. In the introduction and the state-of-the-art chapters, I identify these questions more precisely and provide an overview of the existing literature on the subject. The next chapter gives more de- tails on the control methodology used in this dissertation. Chapter 4 presents three different robot platforms to target the research questions: a simulated network of masses and springs (MSD struc- ture), a cheap passive compliant quadruped robot (Tigrillo), and a state-of-the-art quadruped robot with active compliant actuators (HyQ). The next four chapters expose the approach and discuss the results of the experimental trials conducted on these platforms, to formulate contributions in the domain. Finally, a conclusion is provided in the last chapter. The contribution of this dissertation is manifold and five main topics are investigated throughout the manuscript. First, the re- cent progress in machine learning has led to impressive results for the locomotion of simulated creatures. However, mechanical compliance is not often considered in the work from this field and the difficulty to transfer trained parametric models from simula- tion to the real world has been highlighted in different research tracks. In this dissertation, I suggest an optimization procedure of the physics simulation model to reduce the difference between observations in simulation and the real world. Secondly, an analysis of the state of the art in the field of robotics shows that compliance is not a straightforward defined concept, although it is generally linked to two physical parameters, damping and stiffness. I further analyze this relation through empirical analysis of non-linear robotic systems. They seem to indicate that the concept of resonance, only strictly defined in systems with second-order ordinary differential equations, could serve as a first approximation of the link between stiffness, damping, and optimal locomotion frequency. An investigation of this dependence is conducted on the MSD structures and the HyQ robot. Thirdly, this dissertation presents a training method in two steps: a parameter optimization of open-loop biologically inspired models, followed by a supervised training of a feed-forward neural network, linking robot’s sensors and actuators, to reproduce the targets obtained in the first step in closed-loop. This architecture demon- strates the ability to learn a reflex-based locomotion model without the need for centralized control. The properties of this closed-loop dynamical system are investigated on the three different robotic platforms and the robustness against external disturbance is also discussed. Fourthly, to connect this reflex-based model with biological obser- vations, I also compare two architectural hypotheses in a simple stability controller for a quadruped robot: one using an internal spatiotemporal representation of the body system, and the other based on afferent sensor signals from the lower limbs. I show that the first model performs better in maintaining a target locomotion frequency and resisting external disturbances, which corroborates biological observations conducted on the Cerebellum’s functioning. All these items put together establish an ideal framework to fur- ther investigate the potential exchange of computation capacity between physical body and controller during locomotion, as ex- pected from the theory of morphological computation. A reflection on the matter constitutes the fifth contribution of this disserta- tion. The experiments on HyQ and the MSD networks promote another formulation of this phenomenon: although it is not pos- sible to confirm a transfer of computation per se in my trials, we observe that increased structural complexity and larger mechan- ical compliance contribute to the simplification of computational requirements in the controller and promote a more stable locomo- tion process against external disturbance. In conclusion, this work contributes to better understanding the impact of mechanical compliance on the design and the tuning of a locomotion controller. The experimental results advocate in favor of the use of compliance in robotics, not only to improve performance but also to simplify the control, in association with generic and data-driven controllers. Furthermore, anchoring the architecture of these controllers into biological observations proves to be a source of inspiration for enhanced robots but also a way to test hypotheses and better understand the phenomena in action during human and animal locomotion.
CASIE : computing affect and social intelligence for healthcare in an ethical and trustworthy manner

Vasiliu, Laurentiu, Cortis, Keith, McDermott, Ross, Kerr, Aphra, Peters, Arne, Hesse, Marc, Hagemeyer, Jens, Belpaeme, Tony, McDonald, John, Villing, Rudi, Mileo, Alessandra, Capulto, Annalina, Scriney, Michael, Griffiths, Sascha, Koumpis, Adamantios, and Davis, Brian

PALADYN 2021

Abs Biblio

This article explores the rapidly advancing innovation to endow robots with social intelligence capabilities in the form of multilingual and multimodal emotion recognition, and emotion-aware decision-making capabilities, for contextually appropriate robot behaviours and cooperative social human–robot interaction for the healthcare domain. The objective is to enable robots to become trustworthy and versatile social robots capable of having human-friendly and human assistive interactions, utilised to better assist human users’ needs by enabling the robot to sense, adapt, and respond appropriately to their requirements while taking into consideration their wider affective, motivational states, and behaviour. We propose an innovative approach to the difficult research challenge of endowing robots with social intelligence capabilities for human assistive interactions, going beyond the conventional robotic sense-think-act loop. We propose an architecture that addresses a wide range of social cooperation skills and features required for real human–robot social interaction, which includes language and vision analysis, dynamic emotional analysis (long-term affect and mood), semantic mapping to improve the robot’s knowledge of the local context, situational knowledge representation, and emotion-aware decision-making. Fundamental to this architecture is a normative ethical and social framework adapted to the specific challenges of robots engaging with caregivers and care-receivers.
Chatbot : leerkrachtenhandleiding

Gesquière, Natacha, Hoste, Veronique, Van Hee, Cynthia, and wyffels, Francis

2021

Abs Biblio

Waar taal en technologie samenkomen, ontstaat het domein van Natural Language Processing (NLP). Met technieken uit machine learning, zoals het trainen van modellen om tekst, afbeeldingen of geluidsfrequenties te herkennen, kunnen we computers op een intelligente manier laten werken met taal. Bekende NLP-toepassingen van machine learning zijn bijvoorbeeld vertaalprogramma’s, spraakbesturingssystemen of virtual assistants zoals Siri, en chatrobots. We buigen ons over vragen als "Hoe kunnen we automatisch teksten begrijpen, vertalen en zelfs schrijven?", "Kan een computer emoties of ironie herkennen in een tekst?", en vooral "Hoe belangrijk zijn data?”. Leerlingen van de tweede en de derde graad secundair onderwijs krijgen inzicht in het domein NLP en zetten hun eerste stappen in de wereld van het programmeren met Python.
Coffee with a hint of data : towards using data-driven approaches in personalised long-term interactions

Irfan, Bahar, Hellou, Mehdi, and Belpaeme, Tony

FRONTIERS IN ROBOTICS AND AI 2021

Abs Biblio

While earlier research in human-robot interaction pre-dominantly uses rule-based architectures for natural language interaction, these approaches are not flexible enough for long-term interactions in the real world due to the large variation in user utterances. In contrast, data-driven approaches map the user input to the agent output directly, hence, provide more flexibility with these variations without requiring any set of rules. However, data-driven approaches are generally applied to single dialogue exchanges with a user and do not build up a memory over long-term conversation with different users, whereas long-term interactions require remembering users and their preferences incrementally and continuously and recalling previous interactions with users to adapt and personalise the interactions, known as the lifelong learning problem. In addition, it is desirable to learn user preferences from a few samples of interactions (i.e., few-shot learning). These are known to be challenging problems in machine learning, while they are trivial for rule-based approaches, creating a trade-off between flexibility and robustness. Correspondingly, in this work, we present the text-based Barista Datasets generated to evaluate the potential of data-driven approaches in generic and personalised long-term human-robot interactions with simulated real-world problems, such as recognition errors, incorrect recalls and changes to the user preferences. Based on these datasets, we explore the performance and the underlying inaccuracies of the state-of-the-art data-driven dialogue models that are strong baselines in other domains of personalisation in single interactions, namely Supervised Embeddings, Sequence-to-Sequence, End-to-End Memory Network, Key-Value Memory Network, and Generative Profile Memory Network. The experiments show that while data-driven approaches are suitable for generic task-oriented dialogue and real-time interactions, no model performs sufficiently well to be deployed in personalised long-term interactions in the real world, because of their inability to learn and use new identities, and their poor performance in recalling user-related data.
</p>
Collaborative and inclusive process with the autism community : a case study in Colombia about social robot design

Ramirez-Duque, Andres A., Aycardi, Luis F., Villa, Adriana, Munera, Marcela, Bastos, Teodiano, Belpaeme, Tony, Frizera-Neto, Anselmo, and Cifuentes, Carlos A.

INTERNATIONAL JOURNAL OF SOCIAL ROBOTICS 2021

Abs Biblio

One of the most promising areas in which social assistive robotics has been introduced is therapeutic intervention for children with autism spectrum disorders (CwASD). Even though there are promising results in therapeutic contexts, there is a lack of guidelines on how to select the appropriate robot and how to design and implement the child-robot interaction. The use of participatory design (PD) methods in the design of technology-based processes for CwASD is a recognition of the stakeholders as "experts" in their fields. This work explores the benefits brought by the use of PD methods in the design of a social robot, with a specific focus on their use in autism spectrum disorders therapies on the Colombian autism community. Based on what proved to be effective in our previous research, we implemented participatory methods for both the CwASD and the stakeholders. The process leverages the active role of participants using a focus group approach with parents and specialists, and scene cards, narrative and handmade generative methods with the children. To overcome some challenges of traditional PD processes, where not all community actors are considered, we included a Colombian community consisting of therapists, nurses, caregivers and parents. The proposed PD process provides an opportunity to learn from several community actors (and thus different cultural and social aspects of developing countries), improving traditional robot design methods. In this way, the findings are summarized through a set of guidelines regarding the design of a social robot-device suitable to be implemented for robot-assisted therapy for CwASD.
Communicative function of eye blinks of virtual avatars may not translate onto physical platforms

Oetringer, Djamari, Wolfert, Pieter, Deschuyteneer, Jorre, Thill, Serge, and Belpaeme, Tony

In HRI ’21 Companion: Companion of the 2021 ACM/IEEE International Conference on Human-Robot Interaction 2021

Abs Biblio

Eye behaviour is one of the main modalities used to regulate face-to-face conversation. Gaze aversion and mutual gaze, for example, serve to signal cognitive load, interest or turns during a conversation. While eye blinking is mainly thought to have a physiological function, the rate of blinking is known to increase during conversation suggesting a communicative function for the eye. Recently, it has been shown that a virtual avatar, acting as the receiver in a conversation, could use blinking as a kind of conversational marker influencing the speaker’s communicative behaviour. In particular, it has been demonstrated that long eye blinks resulted in shorter answers by the speaker compared with short ones. Here, we set out to investigate this effect when using a humanoid robot as interaction partner, given that robots have both a physical and social presence. Interestingly, however, we could not replicate the result: short or long blinks did not modulate the length of the responses by the human interactant.
Comparing different nonlinearities in readout systems for optical neuromorphic computing networks

Ma, Chonghuai, Lambrecht, Joris, Laporte, Floris, Yin, Xin, Dambre, Joni, and Bienstman, Peter

SCIENTIFIC REPORTS 2021

Abs Biblio

Nonlinear activation is a crucial building block of most machine-learning systems. However, unlike in the digital electrical domain, applying a saturating nonlinear function in a neural network in the analog optical domain is not as easy, especially in integrated systems. In this paper, we first investigate in detail the photodetector nonlinearity in two main readout schemes: electrical readout and optical readout. On a 3-bit-delayed XOR task, we show that optical readout trained with backpropagation gives the best performance. Furthermore, we propose an additional saturating nonlinearity coming from a deliberately non-ideal voltage amplifier after the detector. Compared to an all-optical nonlinearity, these two kinds of nonlinearities are extremely easy to obtain at no additional cost, since photodiodes and voltage amplifiers are present in any system. Moreover, not having to design ideal linear amplifiers could relax their design requirements. We show through simulation that for long-distance nonlinear fiber distortion compensation, using only the photodiode nonlinearity in an optical readout delivers BER improvements over three orders of magnitude. Combined with the amplifier saturation nonlinearity, we obtain another three orders of magnitude improvement of the BER.
Comparing learning ecologies of primary graphical programming : create or fix?

Neutens, Tom, Barbion, Evelien, Coolsaet, Kris, and wyffels, Francis

JOURNAL OF COMPUTER ASSISTED LEARNING 2021

Abs Biblio

In the last few years, programming, computational thinking, and robotics are more frequently integrated into elementary education. This integration can be done in many different ways. However, it is still unclear which teaching methods work in which situations. To provide some clarity in this area, we compared two methods of integrating programming into a primary robotics workshop for learners aged ten to twelve. In one method, students create programs from scratch; in the other, they start with a faulty program they have to fix. These teaching methods were evaluated using the framework of learning ecology, which provides a holistic framework for assessing complex learning environments. We identified different indicators of learning ecology and assessed our workshops using a mixed-methods approach. Our results showed no difference between the groups on the intrinsic dimension of a learning ecology. However, on the experiential dimension, the learners in the create group scored better on all tests. Our results show the value of a multidimensional assessment of learning ecology to understand different teaching techniques. Additionally, the results provide us with important insights on how to integrate programming into a primary robotics curriculum enabling teachers to select better methods for teaching computing in their classroom.
Computing with integrated photonic reservoirs

Dambre, Joni, Katumba, Andrew, Ma, Chonghuai, Sackesyn, Stijn, Laporte, Floris, Freiberger, Matthias, and Bienstman, Peter

2021

Abs Biblio

The idea of using photonic systems as reservoirs to perform general-purpose computing was first introduced in 2008. Since then, a wide range of systems using either discrete or integrated optical components has been explored. In this chapter, we summarise a decade of research into integrated coherent photonic reservoirs. In these systems, information is carried by the intensity and the phase of light waves. Computations emerge from the way the light propagates inside the system, and the ways in which light that travels along different paths is mixed and transformed. We discuss the computational capabilities of these reservoirs and the trade-offs that can be made to optimise them. We also discuss the technological constraints that are encountered in building such systems and the ways these reflect on their design and training. Finally, we give an overview of recent approaches to combining multiple such reservoirs into larger and computationally more powerful systems.
Effect of compliance on morphological control of dynamic locomotion with HyQ

Urbain, Gabriel, Barasuol, Victor, Semini, Claudio, Dambre, Joni, and wyffels, Francis

AUTONOMOUS ROBOTS 2021

Abs Biblio

Classic control theory applied to compliant and soft robots generally involves an increment of computation that has no equivalent in biology. To tackle this, morphological computation describes a theoretical framework that takes advantage of the computational capabilities of physical bodies. However, concrete applications in robotic locomotion control are still rare. Also, the trade-off between compliance and the capacity of a physical body to facilitate its own control has not been thoroughly studied in a real locomotion task. In this paper, we address these two problems on the state-of-the-art hydraulic robot HyQ. An end-to-end neural network is trained to control HyQ’s joints positions and velocities using only Ground Reaction Forces (GRF). Our simulations and experiments demonstrate better controllability using less memory and computational resources when increasing compliance. However, we show empirically that this effect cannot be attributed to the ability of the body to perform intrinsic computation. It invites to give an increased emphasis on compliance and co-design of the controller and the robot to facilitate attempts in machine learning locomotion.
Experimental demonstration of nonlinear fibre distortion compensation with integrated photonic reservoir computing

Sackesyn, Stijn, Ma, Chonghuai, Dambre, Joni, and Bienstman, Peter

In 2021 European Conference on Optical Communication (ECOC) 2021

Abs Biblio

Optical reservoir computing is a machine learning technique in which a photonic chip can be trained on classification tasks of time signals. This paper presents experimental results where linear and nonlinear fibre distortions are mitigated to below the 0.2×10 −3 FEC limit using a photonic reservoir.
Experimental realization of integrated photonic reservoir computing for nonlinear fiber distortion compensation

Sackesyn, Stijn, Ma, Chonghuai, Dambre, Joni, and Bienstman, Peter

OPTICS EXPRESS 2021

Abs Biblio

Nonlinearity mitigation in optical fiber networks is typically handled by electronic Digital Signal Processing (DSP) chips. Such DSP chips are costly, power-hungry and can introduce high latencies. Therefore, optical techniques are investigated which are more efficient in both power consumption and processing cost. One such a machine learning technique is optical reservoir computing, in which a photonic chip can be trained on certain tasks, with the potential advantages of higher speed, reduced power consumption and lower latency compared to its electronic counterparts. In this paper, experimental results are presented where nonlinear distortions in a 32 GBPS OOK signal are mitigated to below the 0.2 x 10(-3) FEC limit using a photonic reservoir. Furthermore, the results of the reservoir chip are compared to a tapped delay line filter to clearly show that the system performs nonlinear equalisation. (C) 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Frozen pretrained transformers for neural sign language translation

De Coster, Mathieu, D’Oosterlinck, Karel, Pizurica, Marija, Rabaey, Paloma, Verlinden, Severine, Van Herreweghe, Mieke, and Dambre, Joni

In Proceedings of the 1st International Workshop on Automatic Translation for Signed and Spoken Languages (AT4SSL) 2021

Abs Biblio

One of the major challenges in sign language translation from a sign language to a spoken language is the lack of parallel corpora. Recent works have achieved promising results on the RWTH-PHOENIX-Weather 2014T dataset, which consists of over eight thousand parallel sentences between German sign language and German. However, from the perspective of neural machine translation, this is still a tiny dataset. To improve the performance of models trained on small datasets, transfer learning can be used. While this has been previously applied in sign language translation for feature extraction, to the best of our knowledge, pretrained language models have not yet been investigated. We use pretrained BERT-base and mBART-50 models to initialize our sign language video to spoken language text translation model. To mitigate overfitting, we apply the frozen pretrained transformer technique: we freeze the majority of parameters during training. Using a pretrained BERT model, we outperform a baseline trained from scratch by 1 to 2 BLEU-4. Our results show that pretrained language models can be used to improve sign language translation performance and that the self-attention patterns in BERT transfer in zero-shot to the encoder and decoder of sign language translation models.
GENEA Workshop 2021 : the 2nd Workshop on Generation and Evaluation of Non-verbal Behaviour for Embodied Agents

Kucherenko, Taras, Jonell, Patrik, Yoon, Youngwoo, Wolfert, Pieter, Yumak, Zerrin, and Henter, Gustav

In ICMI ’21 : proceedings of the 2021 International Conference on Multimodal Interaction 2021

Abs Biblio

Embodied agents benefit from using non-verbal behavior when communicating with humans. Despite several decades of non-verbal behavior-generation research, there is currently no well-developed benchmarking culture in the field. For example, most researchers do not compare their outcomes with previous work, and if they do, they often do so in their own way which frequently is incompatible with others. With the GENEA Workshop 2021, we aim to bring the community together to discuss key challenges and solutions, and find the most appropriate ways to move the field forward.
HEMVIP : human evaluation of multiple videos in parallel

Jonell, Patrik, Yoon, Youngwoo, Wolfert, Pieter, Kucherenko, Taras, and Henter, Gustav Eje

In ICMI ’21 : Proceedings of the 2021 International Conference on Multimodal Interaction 2021

Abs Biblio

In many research areas, for example motion and gesture generation, objective measures alone do not provide an accurate impression of key stimulus traits such as perceived quality or appropriateness. The gold standard is instead to evaluate these aspects through user studies, especially subjective evaluations of video stimuli. Common evaluation paradigms either present individual stimuli to be scored on Likert-type scales, or ask users to compare and rate videos in a pairwise fashion. However, the time and resources required for such evaluations scale poorly as the number of conditions to be compared increases. Building on standards used for evaluating the quality of multimedia codecs, this paper instead introduces a framework for granular rating of multiple comparable videos in parallel. This methodology essentially analyses all condition pairs at once. Our contributions are 1) a proposed framework, called HEMVIP, for parallel and granular evaluation of multiple video stimuli and 2) a validation study confirming that results obtained using the tool are in close agreement with results of prior studies using conventional multiple pairwise comparisons.
Isolated sign recognition from RGB video using pose flow and self-attention

De Coster, Mathieu, Van Herreweghe, Mieke, and Dambre, Joni

In 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) 2021

Abs Biblio

Automatic sign language recognition lies at the intersection of natural language processing (NLP) and computer vision. The highly successful transformer architectures, based on multi-head attention, originate from the field of NLP. The Video Transformer Network (VTN) is an adaptation of this concept for tasks that require video understanding, e.g., action recognition. However, due to the limited amount of labeled data that is commonly available for training automatic sign (language) recognition, the VTN cannot reach its full potential in this domain. In this work, we reduce the impact of this data limitation by automatically pre-extracting useful information from the sign language videos. In our approach, different types of information are offered to a VTN in a multi-modal setup. It includes per-frame human pose keypoints (extracted by OpenPose) to capture the body movement and hand crops to capture the (evolution of) hand shapes. We evaluate our method on the recently released AUTSL dataset for isolated sign recognition and obtain 92.92% accuracy on the test set using only RGB data. For comparison: the VTN architecture without hand crops and pose flow achieved 82% accuracy. A qualitative inspection of our model hints at further potential of multi-modal multi-head attention in a sign language recognition context.
MIRRA : a modular and cost-effective microclimate monitoring system for real-time remote applications

Pieters, Olivier, Deprost, Emiel, Van Der Donckt, Jonas, Brosens, Lore, Sanczuk, Pieter, Vangansbeke, Pieter, De Swaef, Tom, De Frenne, Pieter, and wyffels, Francis

SENSORS 2021

Abs Biblio

Monitoring climate change, and its impacts on ecological, agricultural, and other societal systems, is often based on temperature data derived from official weather stations. Yet, these data do not capture most microclimates, influenced by soil, vegetation and topography, operating at spatial scales relevant to the majority of organisms on Earth. Detecting and attributing climate change impacts with confidence and certainty will only be possible by a better quantification of temperature changes in forests, croplands, mountains, shrublands, and other remote habitats. There is an urgent need for a novel, miniature and simple device filling the gap between low-cost devices with manual data download (no instantaneous data) and high-end, expensive weather stations with real-time data access. Here, we develop an integrative real-time monitoring system for microclimate measurements: MIRRA (Microclimate Instrument for Real-time Remote Applications) to tackle this problem. The goal of this platform is the design of a miniature and simple instrument for near instantaneous, long-term and remote measurements of microclimates. To that end, we optimised power consumption and transfer data using a cellular uplink. MIRRA is modular, enabling the use of different sensors (e.g., air and soil temperature, soil moisture and radiation) depending upon the application, and uses an innovative node system highly suitable for remote locations. Data from separate sensor modules are wirelessly sent to a gateway, thus avoiding the drawbacks of cables. With this sensor technology for the long-term, low-cost, real-time and remote sensing of microclimates, we lay the foundation and open a wide range of possibilities to map microclimates in different ecosystems, feeding a next generation of models. MIRRA is, however, not limited to microclimate monitoring thanks to its modular and wireless design. Within limits, it is suitable or any application requiring real-time data logging of power-efficient sensors over long periods of time. We compare the performance of this system to a reference system in real-world conditions in the field, indicating excellent correlation with data collected by established data loggers. This proof-of-concept forms an important foundation to creating the next version of MIRRA, fit for large scale deployment and possible commercialisation. In conclusion, we developed a novel wireless cost-effective sensor system for microclimates.
Photonic neuromorphic computing using silicon chips

Bienstman, Peter, Dambre, Joni, Lugnan, Alessio, Sackesyn, Stijn, Ma, Chonghuai, Gooskens, Emmanuel, Gouda, Muhammed Gouda Ahmed, and Masaad, Sarah

In Huawei STW, Abstracts 2021

Biblio
Photonic neuromorphic computing using silicon chips

Bienstman, Peter, Dambre, Joni, Lugnan, Alessio, Sackesyn, Stijn, Ma, Chonghuai, Gooskens, Emmanuel, Gouda, Muhammed Gouda Ahmed, and Masaad, Sarah

In bePOM - Belgian Photonics Online Meetup, Abstracts 2021

Biblio
Photonic reservoir computing for high-speed neuromorphic computing applications

Lugnan, Alessio, Sackesyn, Stijn, Ma, Chonghuai, Gooskens, Emmanuel, Gouda, Muhammed Gouda Ahmed, Masaad, Sarah, Dambre, Joni, and Bienstman, Peter

In 2021 Summer Topicals Meetings Series, Abstracts 2021

Biblio
Reservoir computing for high-speed photonic information processing (invited)

Bienstman, Peter, Dambre, Joni, Lugnan, Alessio, Sackesyn, Stijn, Ma, Chonghuai, Gooskens, Emmanuel, Gouda, Muhammed Gouda Ahmed, and Masaad, Sarah

In Photonics in Switching and Computing, Abstracts 2021

Biblio
Roboter in der Bildung : wie Robotik das Lernen im digitalen Zeitalter bereichern kann

Alnajjar, Fady, Bartneck, Christoph, Baxter, Paul, Belpaeme, Tony, Cappuccio, Massimiliano L., Dio, Cinzia, Eyssel, Friederike, Handke, Jürgen, Mubin, Omar, Obaid, Mohammad, and Reich-Stiebert, Natalia

2021

Biblio
Robots in education : an introduction to high-tech social agents, intelligent tutors, and curricular tools

Alnajjar, Fady, Bartneck, Christoph, Baxter, Paul, Belpaeme, Tony, Cappuccio, Massimiliano, Di Dio, Cinzia, Eyssel, Friederike, Handke, Jürgen, Mubin, Omar, Obaid, Mohammad, and Reich-Stiebert, Natalia

2021

Biblio
Silicon photonics for brain-inspired neuromorphic information processing

Bienstman, Peter, Dambre, Joni, Lugnan, Alessio, Sackesyn, Stijn, Ma, Chonghuai, Gooskens, Emmanuel, Gouda, Muhammed Gouda Ahmed, and Masaad, Sarah

In 1st workshop on neuromorphic photonics, Abstracts 2021

Biblio
Simulating self-learning in photorefractive optical reservoir computers

Laporte, Floris, Dambre, Joni, and Bienstman, Peter

SCIENTIFIC REPORTS 2021

Abs Biblio

Photorefractive materials exhibit an interesting plasticity under the influence of an optical field. By extending the finite-difference time-domain method to include the photorefractive effect, we explore how this property can be exploited in the context of neuromorphic computing for telecom applications. By first priming the photorefractive material with a random bit stream, the material reorganizes itself to better recognize simple patterns in the stream. We demonstrate this by simulating a typical reservoir computing setup, which gets a significant performance boost on performing the XOR on two consecutive bits in the stream after this initial priming step.
Social assistive robots : assessing the impact of a training assistant robot in cardiac rehabilitation

Casas, Jonathan, Senft, Emmanuel, Gutierrez, Luisa F., Rincon-Rocancio, Monica, Munera, Marcela, Belpaeme, Tony, and Cifuentes, Carlos A.

INTERNATIONAL JOURNAL OF SOCIAL ROBOTICS 2021

Abs Biblio

According to the world health organization, cardiovascular diseases are a major cause of death worldwide. Cardiac rehabilitation programmes are dedicated to approach this problem and reduce mortality rates due to the presence of a second event. However, the adherence and motivation of patients to assist to these programmes is not the expected. Therefore, this paper presents the incorporation of a SAR system into a cardiac rehabilitation scenario, where a social robot had the role of a training assistant during the therapy, aiming to increase motivation and encourage people to continue with the therapy. This study carried out a longitudinal experimental setup with a total of 209 sessions observed for a group of 6 patients in a period between 3 and 6 months. Results show that patients felt more encouraged to perform physical activity and continue with the rehabilitation when they perceived that monitored and supervised by the system, demonstrating that it can be implemented as a reliable tool that would potentially leverage tasks carried out by health professionals.
The SignON project : a sign language translation framework

Shterionov, Dimitar, Vandeghinste, Vincent, Saggion, Horacio, Blat, Josep, De Coster, Mathieu, Dambre, Joni, Heuvel, Henk, Murtagh, Irene, Leeson, Lorraine, and Schuurman, Ineke

In 31st Meeting of Computational Linguistics in The Netherlands, Abstracts 2021

Biblio
The robot made me do it : human-robot interaction and risk-taking behavior

Hanoch, Yaniv, Arvizzigno, Francesco, Hernandez García, Daniel, Denham, Sue, Belpaeme, Tony, and Gummerum, Michaela

CYBERPSYCHOLOGY BEHAVIOR AND SOCIAL NETWORKING 2021

Abs Biblio

Empirical evidence has shown that peer pressure can impact human risk-taking behavior. With robots becoming ever more present in a range of human settings, it is crucial to examine whether robots can have a similar impact. Using the balloon analogue risk task (BART), participants’ risk-taking behavior was measured when alone, in the presence of a silent robot, or in the presence of a robot that actively encouraged risk-taking behavior. In the BART, shown to be a proxy for real risk-taking behavior, participants must weigh risk against potential payout. Our results reveal that participants who were encouraged by the robot did take more risks, while the mere presence of the robot in the robot control condition did not entice participants to show more risk-taking behavior. Our results point to both possible benefits and perils that robots might pose to human decision-making. Although increasing risk-taking behavior in some cases has obvious advantages, it could also have detrimental consequences that are only now starting to emerge.
To rate or not to rate : investigating evaluation methods for generated co-speech gestures

Wolfert, Pieter, Girard, Jeffrey M., Kucherenko, Taras, and Belpaeme, Tony

In ICMI ’21 : Proceedings of the 2021 International Conference on Multimodal Interaction 2021

Abs Biblio

While automatic performance metrics are crucial for machine learning of artificial human-like behaviour, the gold standard for evaluation remains human judgement. The subjective evaluation of artificial human-like behaviour in embodied conversational agents is however expensive and little is known about the quality of the data it returns. Two approaches to subjective evaluation can be largely distinguished, one relying on ratings, the other on pairwise comparisons. In this study we use co-speech gestures to compare the two against each other and answer questions about their appropriateness for evaluation of artificial behaviour. We consider their ability to rate quality, but also aspects pertaining to the effort of use and the time required to collect subjective data. We use crowd sourcing to rate the quality of co-speech gestures in avatars, assessing which method picks up more detail in subjective assessments. We compared gestures generated by three different machine learning models with various level of behavioural quality. We found that both approaches were able to rank the videos according to quality and that the ranking significantly correlated, showing that in terms of quality there is no preference of one method over the other. We also found that pairwise comparisons were slightly faster and came with improved inter-rater reliability, suggesting that for small-scale studies pairwise comparisons are to be favoured over ratings.
Unsupervised functional analysis of graphical programs for physical computing

Neutens, Tom, and wyffels, Francis

In 2020 International Conference on Computational Science and Computational Intelligence (CSCI) 2021

Biblio
cREAtIve : reconfigurable embedded artificial intelligence

Bahrebar, Poona, Denis, Leon, Bonnaerens, Maxim, Coddens, Kristof, Dambre, Joni, Favoreel, Wouter, Khvastunov, Illia, Munteanu, Adrian, Nguyen-Duc, Hung, Schulte, Stefan, Stroobandt, Dirk, Valvekens, Ramses, Broeck, Nick, and Verbruggen, Geert

In PROCEEDINGS OF THE 18TH ACM INTERNATIONAL CONFERENCE ON COMPUTING FRONTIERS 2021 (CF 2021) 2021

Abs Biblio

cREAtIve targets the development of novel highly-adaptable embedded deep learning solutions for automotive and traffic monitoring applications, including position sensor processing, scene interpretation based on LiDAR, and object detection and classification in thermal images for traffic camera systems. These applications share the need for deep learning solutions tailored for deployment on embedded devices with limited resources and featuring high adaptability and robustness to changing environmental conditions. cREAtIve develops knowledge, tools and methods that enable hardware-efficient, adaptable, and robust deep learning.

2020

A training algorithm for networks of high-variability reservoirs

Freiberger, Matthias, Bienstman, Peter, and Dambre, Joni

SCIENTIFIC REPORTS 2020

Abs Biblio

Physical reservoir computing approaches have gained increased attention in recent years due to their potential for low-energy high-performance computing. Despite recent successes, there are bounds to what one can achieve simply by making physical reservoirs larger. Therefore, we argue that a switch from single-reservoir computing to multi-reservoir and even deep physical reservoir computing is desirable. Given that error backpropagation cannot be used directly to train a large class of multi-reservoir systems, we propose an alternative framework that combines the power of backpropagation with the speed and simplicity of classic training algorithms. In this work we report our findings on a conducted experiment to evaluate the general feasibility of our approach. We train a network of 3 Echo State Networks to perform the well-known NARMA-10 task, where we use intermediate targets derived through backpropagation. Our results indicate that our proposed method is well-suited to train multi-reservoir systems in an efficient way.
Advice to new human-robot interaction researchers

Belpaeme, Tony

2020

Biblio
An open-source social robot based on CompliAnt SofT Robotics for therapy with children with ASD

Casas-Bocanegra, Diego, Gomez-Vargas, Daniel, Pinto Bernal, Maria Jose, Maldonado, Juan, Munera, Marcela, Villa-Moreno, Adriana, Stoelen, Martin F., Belpaeme, Tony, and Cifuentes, Carlos A.

ACTUATORS 2020

Abs Biblio

Therapy with robotic tools is a promising way to help improve verbal and nonverbal communication in children. The robotic tools are able to increase aspects such as eye contact and the ability to follow instructions and to empathize with others. This work presents the design methodology, development, and experimental validation of a novel social robot based on CompliAnt SofT Robotics called the CASTOR robot, which intends to be used as an open-source platform for the long-term therapy of children with autism spectrum disorder (CwASD). CASTOR integrates the concepts of soft actuators and compliant mechanisms to create a replicable robotic platform aimed at real therapy scenarios involving physical interaction between the children and the robot. The validation shows promising results in terms of robustness and the safety of the user and robot. Likewise, mechanical tests assess the robot’s response to blocking conditions for two critical modules (i.e., neck and arm) in interaction scenarios. Future works should focus on the validation of the robot’s effectiveness in the therapy of CwASD.
Automatic detection of the aortic annular plane and coronary ostia from multidetector computed tomography

Astudillo, Patricio, Mortier, Peter, Bosmans, Johan, De Backer, Ole, Jaegere, Peter, Iannaccone, Francesco, De Beule, Matthieu, and Dambre, Joni

JOURNAL OF INTERVENTIONAL CARDIOLOGY 2020

Abs Biblio

Anatomic landmark detection is crucial during preoperative planning of transcatheter aortic valve implantation (TAVI) to select the proper device size and assess the risk of complications. The detection is currently a time-consuming manual process influenced by the image quality and subject to operator variability. In this work, we propose a novel automatic method to detect the relevant aortic landmarks from MDCT images using deep learning techniques. We trained three convolutional neural networks (CNNs) with 344 multidetector computed tomography (MDCT) acquisitions to detect five anatomical landmarks relevant for TAVI planning: the three basal attachment points of the aortic valve leaflets and the left and right coronary ostia. The detection strategy used these three CNN models to analyse a single MDCT image and yield three segmentation volumes as output. These segmentation volumes were averaged into one final segmentation volume, and the final predicted landmarks were obtained during a postprocessing step. Finally, we constructed the aortic annular plane, defined by the three predicted hinge points, and measured the distances from this plane to the predicted coronary ostia (i.e., coronary height). The methodology was validated on 100 patients. The automatic landmark detection was able to detect all the landmarks and showed high accuracy as the median distance between the ground truth and predictions is lower than the interobserver variations (1.5 mm [1.1-2.1], 2.0 mm [1.3-2.8] with a paired difference -0.5 +/- 1.3 mm and p value <0.001). Furthermore, a high correlation is observed between predicted and manually measured coronary heights (for both R-2 = 0.8). The image analysis time per patient was below one second. The proposed method is accurate, fast, and reproducible. Embedding this tool based on deep learning in the preoperative planning routine may have an impact in the TAVI environments by reducing the time and cost and improving accuracy.
Century‐long apparent decrease in iWUE with no evidence of progressive nutrient limitation in African tropical forests

Bauters, Marijn, Meeus, Sofie, Barthel, Matti, Stoffelen, Piet, De Deurwaerder, Hannes, Meunier, Félicien, Drake, Travis W., Ponette, Quentin, Ebuy, Jerôme, Vermeir, Pieter, Beeckman, Hans, wyffels, Francis, Bodé, Samuel, Verbeeck, Hans, Vandelook, Filip, and Boeckx, Pascal

GLOBAL CHANGE BIOLOGY 2020

Abs Biblio

Forests exhibit leaf- and ecosystem-level responses to environmental changes. Specifically, rising carbon dioxide (CO2) levels over the past century are expected to have increased the intrinsic water-use efficiency (iWUE) of tropical trees while the ecosystem is gradually pushed into progressive nutrient limitation. Due to the long-term character of these changes, however, observational datasets to validate both paradigms are limited in space and time. In this study, we used a unique herbarium record to go back nearly a century and show that despite the rise in CO2 concentrations, iWUE has decreased in central African tropical trees in the Congo Basin. Although we find evidence that points to leaf-level adaptation to increasing CO2-that is, increasing photosynthesis-related nutrients and decreasing maximum stomatal conductance, a decrease in leaf delta C-13 clearly indicates a decreasing iWUE over time. Additionally, the stoichiometric carbon to nitrogen and nitrogen to phosphorus ratios in the leaves show no sign of progressive nutrient limitation as they have remained constant since 1938, which suggests that nutrients have not increasingly limited productivity in this biome. Altogether, the data suggest that other environmental factors, such as increasing temperature, might have negatively affected net photosynthesis and consequently downregulated the iWUE. Results from this study reveal that the second largest tropical forest on Earth has responded differently to recent environmental changes than expected, highlighting the need for further on-ground monitoring in the Congo Basin.
Challenges of a real-world HRI study with non-native English speakers : can personalisation save the day?

Irfan, Bahar, Hellou, Mehdi, Mazel, Alexandre, and Belpaeme, Tony

In HRI’20: COMPANION OF THE 2020 ACM/IEEE INTERNATIONAL CONFERENCE ON HUMAN-ROBOT INTERACTION 2020

Abs Biblio

Real-world studies allow for testing the limits of HRI systems and observing how people react to failures. We developed a fully autonomous personalised barista robot and deployed the robot on an international student campus for five days. We experienced several challenges, the most important one being speech recognition failures due to foreign accents. Nonetheless, these failures showed a different perspective on HRI, and we demonstrate how personalisation can overcome a negative user experience.
Close range hyperspectral camera dataset with high temporal resolution of strawberry with eco-physiological data of one leaf

Pieters, Olivier, De Swaef, Tom, Lootens, Peter, Stock, Michiel, Roldán-Ruiz, Isabel, and wyffels, Francis

2020

Abs Biblio

This high temporal resolution dataset of a strawberry plant was captured in two experiments, each lasting 100h. On leaf was inserted into a leaf chamber of the LI-6400XT gas exchange system, capturing information on transpiration, photosynthesis and stomatal conductance. Environmental characteristics are also captured at canopy height. These experiments were conducted in a growth chamber at ILVO (Melle, Belgium) and only covered conditions that did not result in stress in the plant. As such, this dataset attempts to capture subtle dynamic variation in the plant.
Conditional BRUNO : a neural process for exchangeable labelled data

Korshunova, Iryna, Gal, Yarin, Gretton, Arthur, and Dambre, Joni

NEUROCOMPUTING 2020

Biblio
Curriculum deep reinforcement learning with different exploration strategies : a feasibility study on cardiac landmark detection

Astudillo, Patricio, Mortier, Peter, De Beule, Matthieu, and wyffels, Francis

In PROCEEDINGS OF THE 13TH INTERNATIONAL JOINT CONFERENCE ON BIOMEDICAL ENGINEERING SYSTEMS AND TECHNOLOGIES, VOL 2: BIOIMAGING 2020

Abs Biblio

Transcatheter aortic valve implantation (TAVI) is associated with conduction abnormalities and the mechanical interaction between the prosthesis and the atrioventricular (AV) conduction path cause these life-threatening arrhythmias. Pre-operative assessment of the location of the AV conduction path can help to understand the risk of post-TAVI conduction abnormalities. As the AV conduction path is not visible on cardiac CT, the inferior border of the membranous septum can be used as an anatomical landmark. Detecting this border automatically, accurately and efficiently would save operator time and thus benefit pre-operative planning. This preliminary study was performed to identify the feasibility of 3D landmark detection in cardiac CT images with curriculum deep Q-learning. In this study, curriculum learning was used to gradually teach an artificial agent to detect this anatomical landmark from cardiac CT. This agent was equipped with a small field of view and burdened with a large ac tion-space. Moreover, we introduced two novel action-selection strategies: α-decay and action-dropout. We compared these two strategies to the already established ε-decay strategy and observed that α-decay yielded the most accurate results. Limited computational resources were used to ensure reproducibility. In order to maximize the amount of patient data, the method was cross-validated with k-folding for all three action-selection strategies. An inter-operator variability study was conducted to assess the accuracy of the method
Dendritic computation in a point neuron model

Vandesompele, Alexander, wyffels, Francis, and Dambre, Joni

In ARTIFICIAL NEURAL NETWORKS AND MACHINE LEARNING, ICANN 2020, PT II 2020

Abs Biblio

Biological neurons possess elaborate dendrites that perform elaborate computations. They are however ignored in the widely used point neuron models. Here, we present a simple addition to the commonly used leaky integrate-and-fire model that introduces the concept of a dendrite. All synapses on the dendrite have a mutual relationship. The result is a form of short term plasticity in which synapse strengths are influenced by recent activity in other synapses. This improves the ability of the neuron to recognize temporal sequences.
Development of a quantitative comparison tool for plant models

Pieters, Olivier, De Swaef, Tom, and wyffels, Francis

In FSPM2020: Towards Computable Plants; 9th International Conference on Functional-structural Plant Models 2020

Biblio
From leaf to label : a robust automated workflow for stomata detection

Meeus, Sofie, Bulcke, Jan, and wyffels, Francis

ECOLOGY AND EVOLUTION 2020

Abs Biblio

1. Plant leaf stomata are the gatekeepers of the atmosphere-plant interface and are essential building blocks of land surface models as they control transpiration and photosynthesis. Although more stomatal trait data is needed to significantly reduce the error in these model predictions, recording these traits is time-consuming and no standardized protocol is currently available. Some attempts were made to automate stomatal detection from photomicrographs, however, these approaches have the disadvantage of using classic image processing or targeting a narrow taxonomic entity which makes these technologies less robust and generalizable to other plant species. We propose an easy-to-use and adaptable workflow from leaf to label. A methodology for automatic stomata detection was developed using deep neural networks according to the state-of-the-art and its applicability demonstrated across the phylogeny of the angiosperms. 2. We used a patch-based approach for training/tuning three different deep learning architectures. For training, we used 431 micrographs taken from leaf prints made according the nail polish method from herbarium specimens of 19 species. The best performing architecture was tested on 595 images of 16 additional species spread across the angiosperm phylogeny. 3. The nail polish method was successfully applied in 78% of the species sampled here. The VGG19 architecture slightly outperformed the basic shallow and deep architectures, with a confidence threshold equal to 0.7 resulting in an optimal trade-off between precision and recall. Applying this threshold the VGG19 architecture obtained an average F-score of 0.87, 0.89 and 0.67 on the training, validation and unseen test set, respectively. The average accuracy was very high (94%) for computed stomatal counts on unseen images of species used for training. 4. The leaf-to-label pipeline is an easy-to-use workflow for researchers of different areas of expertise interested in detecting stomata more efficiently. The described methodology was based on multiple species and well-established methods so that it can serve as a reference for future work.
Gloxinia—an open-source sensing platform to monitor the dynamic responses of plants

Pieters, Olivier, De Swaef, Tom, Lootens, Peter, Stock, Michiel, Roldán-Ruiz, Isabel, and wyffels, Francis

SENSORS 2020

Abs Biblio

The study of the dynamic responses of plants to short-term environmental changes is becoming increasingly important in basic plant science, phenotyping, breeding, crop management, and modelling. These short-term variations are crucial in plant adaptation to new environments and, consequently, in plant fitness and productivity. Scalable, versatile, accurate, and low-cost data-logging solutions are necessary to advance these fields and complement existing sensing platforms such as high-throughput phenotyping. However, current data logging and sensing platforms do not meet the requirements to monitor these responses. Therefore, a new modular data logging platform was designed, named Gloxinia. Different sensor boards are interconnected depending upon the needs, with the potential to scale to hundreds of sensors in a distributed sensor system. To demonstrate the architecture, two sensor boards were designed—one for single-ended measurements and one for lock-in amplifier based measurements, named Sylvatica and Planalta, respectively. To evaluate the performance of the system in small setups, a small-scale trial was conducted in a growth chamber. Expected plant dynamics were successfully captured, indicating proper operation of the system. Though a large scale trial was not performed, we expect the system to scale very well to larger setups. Additionally, the platform is open-source, enabling other users to easily build upon our work and perform application-specific optimisations.
HRI’20 : Proceedings of the 2020 ACM/IEEE International Conference on Human-Robot Interaction

2020

Biblio
High-resolution time series of leaf length and leaf thickness of strawberry and tomato measured in a growth chamber

Pieters, Olivier, De Swaef, Tom, Lootens, Peter, Stock, Michiel, Roldàn-Ruiz, Isabel, and wyffels, Francis

2020

Abs Biblio

This dataset captures leaf thickness and leaf length of two strawberry plants and one tomato seeding. Leaf thickness measurements are only available of the strawberry plants. Additional data included light intensity data, soil water content (of one plant), relative humidity and temperature data. All sensors were read every second over a period of two weeks. Original data and calibrated data is provided, along with the necessary scripts to process the data.
Human-robot interaction : an introduction

Bartneck, Christoph, Belpaeme, Tony, Eyssel, Friederike, Kanda, Takayuki, Keijsers, Merel, and Šabanović, Selma

2020

Abs Biblio

The role of robots in society keeps expanding and diversifying, bringing with it a host of issues surrounding the relationship between robots and humans. This introduction to human-robot interaction (HRI), written by leading researchers in this developing field, is the first to provide a broad overview of the multidisciplinary topics central to modern HRI research. Students and researchers from robotics, artificial intelligence, psychology, sociology, and design will find it a concise and accessible guide to the current state of the field. Written for students from diverse backgrounds, it presents relevant background concepts, describing how robots work, how to design them, and how to evaluate their performance. Self-contained chapters discuss a wide range of topics, including the different communication modalities such as speech and language, non-verbal communication and the processing of emotions, as well as ethical issues around the application of robots today and in the context of our future society.
Human-robot interaction : evaluation methods and their standardization

2020

Biblio
Improvement of child-robot interaction for children diagnosed with cognitive impairments

Ismail, Luthffi

2020

Biblio
Improving time series recognition and prediction with networks and ensembles of passive photonic reservoirs

Freiberger, Matthias, Sackesyn, Stijn, Ma, Chonghuai, Katumba, Andrew, Bienstman, Peter, and Dambre, Joni

IEEE JOURNAL OF SELECTED TOPICS IN QUANTUM ELECTRONICS 2020

Abs Biblio

As the performance increase of traditional Von-Neumann computing attenuates, new approaches to computing need to be found. A promising approach for low-power computing at high bitrates is integrated photonic reservoir computing. In the past though, the feasible reservoir size and computational power of integrated photonic reservoirs have been limited by hardware constraints. An alternative solution to building larger reservoirs is the combination of several small reservoirs to match or exceed the performance of a single bigger one. This paper summarizes our efforts to increase the available computational power by combining multiple reservoirs into a single computing architecture. We investigate several possible combination techniques and evaluate their performance using the classic XOR and header recognition tasks as well as the well-known Santa Fe chaotic laser prediction task. Our findings suggest that a new paradigm of feeding a reservoir’s output into the readout structure of the next one shows consistently good results for various tasks as well as for both electrical and optical readouts and coupling schemes.
Kunstmatige intelligentie, klimaatverandering, stomata : KIKS : lerarenhandleiding

Gesquière, Natacha, Meeus, Sofie, Bulcke, Jan, and wyffels, Francis

2020

Biblio
Limitations of snapshot hyperspectral cameras to monitor plant response dynamics in stress-free conditions

Pieters, Olivier, De Swaef, Tom, Lootens, Peter, Stock, Michiel, Roldán-Ruiz, Isabel, and wyffels, Francis

COMPUTERS AND ELECTRONICS IN AGRICULTURE 2020

Abs Biblio

Plants’ dynamic eco-physiological responses are vital to their productivity in continuously fluctuating conditions, such as those in agricultural fields. However, it is currently still very difficult to capture these responses at the field scale for phenotyping purposes. Advanced hyperspectral imaging tools are increasingly used in phenotyping, and have been applied to detect changes in plants in response to a specific treatment, phenological state or monitor its growth and development. Phenotyping has to evolve towards capturing dynamic behaviour under more subtle fluctuations in environmental conditions, without the presence of clear treatments or stresses. Therefore, we investigated the potential of hyperspectral imaging to capture dynamic behaviour of plants in stress-free conditions at a temporal resolution of seconds. Two growth chamber experiments were set up, in which strawberry plants and four different background materials, serving as controls, were monitored by a snapshot hyperspectral camera in variable conditions of light, temperature and relative humidity. The sampling period was set to three seconds, triggering image acquisition and gas exchange measurements. Different background materials were used to assess the influence of the environment and the camera in both experiments. To separate the plant and background data, static masks were determined. Two datasets were created, which encompass both experiments. One dataset was constructed after averaging over the entire mask to acquire one value per spectral band. These values were then used to calculate a set of vegetation indices. The other dataset used spatial subsampling to retain spatial information. From both datasets, linear models were constructed using ridge regression, which estimated the measured eco-physiological and environmental data. Leaf temperature and vapour pressure deficit based on leaf temperature are the two main eco-physiological characteristics that could be predicted successfully. Stomatal conductance, photosynthesis and transpiration rate show less promising results. We suspect that limited variation, and low spectral resolution and range are the main causes of the inability of the models to extract meaningful predictions. Furthermore, the models that were only trained on background data also showed good predictive performance. This is probably because the main drivers for good performing eco-physiological variables are temperature and incident light intensity. Environmental characteristics that have good performance are photosynthetically active radiation and air temperature. Current hyperspectral sensing technologies are not yet able to uncover most plant dynamic eco-physiological responses when plants are cultivated in stress-free conditions.
Machine learning issues and opportunities in ultrafast particle classification for label-free microflow cytometry

Lugnan, Alessio, Gooskens, Emmanuel, Vatin, Jeremy, Dambre, Joni, and Bienstman, Peter

SCIENTIFIC REPORTS 2020

Abs Biblio

Machine learning offers promising solutions for high-throughput single-particle analysis in label-free imaging microflow cytomtery. However, the throughput of online operations such as cell sorting is often limited by the large computational cost of the image analysis while offline operations may require the storage of an exceedingly large amount of data. Moreover, the training of machine learning systems can be easily biased by slight drifts of the measurement conditions, giving rise to a significant but difficult to detect degradation of the learned operations. We propose a simple and versatile machine learning approach to perform microparticle classification at an extremely low computational cost, showing good generalization over large variations in particle position. We present proof-of-principle classification of interference patterns projected by flowing transparent PMMA microbeads with diameters of 15.2 mu m and 18.6 mu m. To this end, a simple, cheap and compact label-free microflow cytometer is employed. We also discuss in detail the detection and prevention of machine learning bias in training and testing due to slight drifts of the measurement conditions. Moreover, we investigate the implications of modifying the projected particle pattern by means of a diffraction grating, in the context of optical extreme learning machine implementations.
Photonic neuromorphic information processing and reservoir computing

Lugnan, Alessio, Katumba, Andrew, Laporte, Floris, Freiberger, Matthias, Sackesyn, Stijn, Ma, C., Gooskens, Emmanuel, Dambre, Joni, and Bienstman, Peter

APL PHOTONICS 2020

Abs Biblio

Photonic neuromorphic computing is attracting tremendous research interest now, catalyzed in no small part by the rise of deep learning in many applications. In this paper, we will review some of the exciting work that has been going in this area and then focus on one particular technology, namely, photonic reservoir computing.
Photonic reservoir computing

Laporte, Floris, Katumba, Andrew, Freiberger, Matthias, Ma, Chonghuai, Sackesyn, Stijn, Gooskens, Emmanuel, Lugnan, Alessio, Dambre, Joni, and Bienstman, Peter

In Photonic Integration Week, Abstracts 2020

Biblio
PyTorch-Hebbian : facilitating local learning in a deep learning framework

Talloen, Jules, Vandesompele, Alexander, and Dambre, Joni

In NeurIPS 2020 Beyond backpropagation Workshop 2020

Biblio
Pyglmnet : Python implementation of elastic-net regularized generalized linear models

Jas, Mainak, Achakulvisut, Titipat, Idrizović, Aid, Acuna, Daniel, Antalek, Matthew, Marques, Vinicius, Odland, Tommy, Garg, Ravi Prakash, Agrawal, Mayank, Umegaki, Yu, Foley, Peter, Fernandes, Hugo, Harris, Drew, Li, Beibin, Pieters, Olivier, Otterson, Scott, De Toni, Giovanni, Rodgers, Chris, Dyer, Eva, Hamalainen, Matti, Kording, Konrad, and Ramkumar, Pavan

JOURNAL OF OPEN SOURCE SOFTWARE 2020

Biblio
Scaling up integrated photonic reservoirs towards low-power high-bandwidth computing

Freiberger, Matthias

2020

Biblio
Security risks of social robots used to persuade and manipulate : a proof of concept study

Wolfert, Pieter, Deschuyteneer, Jorre, Oetringer, Djamari, Robinson, Nicole, and Belpaeme, Tony

In HRI’20: COMPANION OF THE 2020 ACM/IEEE INTERNATIONAL CONFERENCE ON HUMAN-ROBOT INTERACTION 2020

Abs Biblio

Earlier research has shown that robots can provoke social responses in people, and that robots often elicit compliance. In this paper we discuss three proof of concept studies in which we explore the possibility of robots being hacked and taken over by others with the explicit purpose of using the robot’s social capabilities. Three scenarios are explored: gaining access to secured areas, extracting sensitive and personal information, and convincing people to take unsafe action. We find that people are willing to do these tasks, and that social robots tend to be trusted, even in situations that would normally cause suspicion.
Self-organized dynamic attractors in recurrent neural networks

Vettelschoss, Benedikt, Freiberger, Matthias, and Dambre, Joni

In ESANN 2020, European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, Proceediings 2020

Biblio
Sign language recognition with transformer networks

De Coster, Mathieu, Van Herreweghe, Mieke, and Dambre, Joni

In PROCEEDINGS OF THE 12TH INTERNATIONAL CONFERENCE ON LANGUAGE RESOURCES AND EVALUATION (LREC 2020) 2020

Abs Biblio

Sign languages are complex languages. Research into them is ongoing, supported by large video corpora of which only small parts are annotated. Sign language recognition can be used to speed up the annotation process of these corpora, in order to aid research into sign languages and sign language recognition. Previous research has approached sign language recognition in various ways, using feature extraction techniques or end-to-end deep learning. In this work, we apply a combination of feature extraction using OpenPose for human keypoint estimation and end-to-end feature learning with Convolutional Neural Networks. The proven multi-head attention mechanism used in transformers is applied to recognize isolated signs in the Flemish Sign Language corpus. Our proposed method significantly outperforms the previous state of the art of sign language recognition on the Flemish Sign Language corpus: we obtain an accuracy of 74.7% on a vocabulary of 100 classes. Our results will be implemented as a suggestion system for sign language corpus annotation.
Simpler learning of robotic manipulation of clothing by utilizing DIY smart textile technology

Verleysen, Andreas, Holvoet, Thomas, Proesmans, Remko, Den Haese, Cedric, and wyffels, Francis

APPLIED SCIENCES-BASEL 2020

Abs Biblio

Deformable objects such as ropes, wires, and clothing are omnipresent in society and industry but are little researched in robotics research. This is due to the infinite amount of possible state configurations caused by the deformations of the deformable object. Engineered approaches try to cope with this by implementing highly complex operations in order to estimate the state of the deformable object. This complexity can be circumvented by utilizing learning-based approaches, such as reinforcement learning, which can deal with the intrinsic high-dimensional state space of deformable objects. However, the reward function in reinforcement learning needs to measure the state configuration of the highly deformable object. Vision-based reward functions are difficult to implement, given the high dimensionality of the state and complex dynamic behavior. In this work, we propose the consideration of concepts beyond vision and incorporate other modalities which can be extracted from deformable objects. By integrating tactile sensor cells into a textile piece, proprioceptive capabilities are gained that are valuable as they provide a reward function to a reinforcement learning agent. We demonstrate on a low-cost dual robotic arm setup that a physical agent can learn on a single CPU core to fold a rectangular patch of textile in the real world based on a learned reward function from tactile information.
Stance control inspired by cerebellum stabilizes reflex-based locomotion on HyQ robot

Urbain, Gabriel, Barasuol, Victor, Semini, Claudio, Dambre, Joni, and wyffels, Francis

In 2020 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA) 2020

Abs Biblio

Advances in legged robotics are strongly rooted in animal observations. A clear illustration of this claim is the generalization of Central Pattern Generators (CPG), first identified in the cat spinal cord, to generate cyclic motion in robotic locomotion. Despite a global endorsement of this model, physiological and functional experiments in mammals have also indicated the presence of descending signals from the cerebellum, and reflex feedback from the lower limb sensory cells, that closely interact with CPGs. To this day, these interactions are not fully understood. In some studies, it was demonstrated that pure reflex-based locomotion in the absence of oscillatory signals could be achieved in realistic musculoskeletal simulation models or small compliant quadruped robots. At the same time, biological evidence has attested the functional role of the cerebellum for predictive control of balance and stance within mammals. In this paper, we promote both approaches and successfully apply reflex-based dynamic locomotion, coupled with a balance and gravity compensation mechanism, on the state-of-art HyQ robot. We discuss the importance of this stability module to ensure a correct foot lift-off and maintain a reliable gait. The robotic platform is further used to test two different architectural hypotheses inspired by the cerebellum. An analysis of experimental results demonstrates that the most biologically plausible alternative also leads to better results for robust locomotion.
The DREAM dataset : supporting a data-driven study of autism spectrum disorder and robot enhanced therapy

Billing, Erik, Belpaeme, Tony, Cai, Haibin, Cao, Hoang-Long, Ciocan, Anamaria, Costescu, Cristina, David, Daniel, Homewood, Robert, Garcia, Daniel Hernandez, Esteban, Pablo Gomez, Liu, Honghai, Nair, Vipul, Matu, Silviu, Mazel, Alexandre, Selescu, Mihaela, Senft, Emmanuel, Thill, Serge, Vanderborght, Bram, Vernon, David, and Ziemke, Tom

PLOS ONE 2020

Abs Biblio

We present a dataset of behavioral data recorded from 61 children diagnosed with Autism Spectrum Disorder (ASD). The data was collected during a large-scale evaluation of Robot Enhanced Therapy (RET). The dataset covers over 3000 therapy sessions and more than 300 hours of therapy. Half of the children interacted with the social robot NAO supervised by a therapist. The other half, constituting a control group, interacted directly with a therapist. Both groups followed the Applied Behavior Analysis (ABA) protocol. Each session was recorded with three RGB cameras and two RGBD (Kinect) cameras, providing detailed information of children’s behavior during therapy. This public release of the dataset comprises body motion, head position and orientation, and eye gaze variables, all specified as 3D data in a joint frame of reference. In addition, metadata including participant age, gender, and autism diagnosis (ADOS) variables are included. We release this data with the hope of supporting further data-driven studies towards improved therapy methods as well as a better understanding of ASD in general.
The GENEA Challenge 2020 : benchmarking gesture-generation systems on common data

Kucherenko, Taras, Jonell, Patrik, Yoon, Youngwoo, Wolfert, Pieter, and Henter, Gustav Eje

In GENEA workshop 2020, part of IVA’20, Proceedings 2020

Abs Biblio

Automatic gesture generation is a field of growing interest, and a key technology for enabling embodied conversational agents. Research into gesture generation is rapidly gravitating towards data-driven methods. Unfortunately, individual research efforts in the field are difficult to compare: there are no established benchmarks, and each study tends to use its own dataset, motion visualisation, and evaluation methodology. To address this situation, we launched the GENEA gesture-generation challenge, wherein participating teams built automatic gesture-generation systems on a common dataset, and the resulting systems were evaluated in parallel in a large, crowdsourced user study. Since differences in evaluation outcomes between systems now are solely attributable to differences between the motion-generation methods, this enables benchmarking recent approaches against one another and investigating the state of the art in the field. This paper provides a first report on the purpose, design, and results of our challenge, with each individual team’s entry described in a separate paper also presented at the GENEA Workshop. Additional information about the workshop can be found at genea-workshop.github.io/202
Unsupervised orientation learning using autoencoders

Daems, Rembert, and wyffels, Francis

In NeurIPS 2020 DiffGeo4DL Workshop 2020

Abs Biblio

We present a method to learn the orientation of symmetric objects in real-world images in an unsupervised way. Our method explicitly maps in-plane relative rotations to the latent space of an autoencoder, by rotating both in the image domain and latent domain. This is achieved by adding a proposed \textitcrossing loss to a standard autoencoder training framework which enforces consistency between the image domain and latent domain rotations. This relative representation of rotation is made absolute, by using the symmetry of the observed object, resulting in an unsupervised method to learn the orientation. Furthermore, orientation is disentangled in latent space from other descriptive factors. We apply this method on two real-world datasets: aerial images of planes in the DOTA dataset and images of densely packed honeybees. We empirically show this method can learn orientation using no annotations with high accuracy compared to the same models trained with annotations.
User-driven design of robot costume for child-robot interactions among children with cognitive impairment

Ismail, Luthffi, Hanapiah, Fazah Akhtar, and wyffels, Francis

In ROBOTICS IN EDUCATION: CURRENT RESEARCH AND INNOVATIONS 2020

Abs Biblio

The involvement of arts and psychology elements in robotics research for children with cognitive impairment is still limited. However, the combination of robots, arts, psychology and education in the development of robots could significantly contribute to the improvement of social interaction skills among children with cognitive impairment. In this article, we would like to share our work on building and innovating the costume of LUCA’s robot, which incorporating the positive psychological perspectives and arts values for children with cognitive impairment. Our goals are (1) to educate arts students in secondary arts school on the importance of social robot appearance for children with cognitive impairment, and (2) to select the best costume for future child-robot interaction study with children with cognitive impairments.
Using a personalised socially assistive robot for cardiac rehabilitation : a long-term case study

Irfan, Bahar, Gomez, Nathalia Cespedes, Casas, Jonathan, Senft, Emmanuel, Gutierrez, Luisa F., Rincon-Roncancio, Monica, Munera, Marcela, Belpaeme, Tony, and Cifuentes, Carlos A.

In 2020 29th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN) 2020

Abs Biblio

This paper presents a longitudinal case study of Robot Assisted Therapy for cardiac rehabilitation. The patient, who is a 60-year old male that suffered a myocardial infarction and received angioplasty surgery, successfully recovered after 35 sessions of rehabilitation with a social robot, lasting 18 weeks. The sessions took place directly at the clinic and relied on an exercise regime which was designed by the clinicians and delivered with the support of a social robot and a sensor suite. The robot monitored the patient’s progress, and provided personalised encouragement and feedback. We discuss the recovery of the patient and illustrate how the use of a social robot, its sensory systems and its personalised interaction was instrumental to maintain engagement with the programme and to the patient’s recovery. Of note is a critical event that was promptly detected by the robot, which allowed fast intervention measures to be taken by the medical staff for the referral of the patient for further surgery.
Van motorvoertuig tot autonome robot

wyffels, Francis

2020

Biblio
Video dataset of human demonstrations of folding clothing for robotic folding

Verleysen, Andreas, Biondina, Matthijs, and wyffels, Francis

INTERNATIONAL JOURNAL OF ROBOTICS RESEARCH 2020

Abs Biblio

General-purpose clothes-folding robots do not yet exist owing to the deformable nature of textiles, making it hard to engineer manipulation pipelines or learn this task. In order to accelerate research for the learning of the robotic clothes folding task, we introduce a video dataset of human folding demonstrations. In total, we provide 8.5 hours of demonstrations from multiple perspectives leading to 1,000 folding samples of different types of textiles. The demonstrations are recorded in multiple public places, in different conditions with a diverse set of people. Our dataset consists of anonymized RGB images, depth frames, skeleton keypoint trajectories, and object labels. In this article, we describe our recording setup, the data format, and utility scripts, which can be accessed at https://adverley.github.io/folding-demonstrations.

2019

A CycleGAN for style transfer between drum and bass subgenres

Vande Veire, Len, De Bie, Tijl, and Dambre, Joni

In ML4MD at ICML2019, Proceedings 2019

Abs Biblio

In this work, we apply the CycleGAN image-to-image translation framework to Mel-scaled log-amplitude spectrograms, successfully realizing audio texture transfer between excerpts from two musically related genres. Such automatic musical transfer could provide music producers and DJs with new creative tools, e.g. to quickly prototype a remix of an existing song in another genre, or to use as an advanced effect during a live performance. We show that meaningful style transfer can be realized using only a limited amount of data and computational resources. A high-quality audio reconstruction is obtained from the generated amplitude spectrogram by simply using the phase of the original audio as an approximation for the phase of the generated spectrogram. This results in a significant quality improvement over traditional phase reconstruction methods.
A differentiable physics engine for deep learning in robotics

Degrave, Jonas, Hermans, Michiel, Dambre, Joni, and wyffels, Francis

FRONTIERS IN NEUROROBOTICS 2019

Abs Biblio

An important field in robotics is the optimization of controllers. Currently, robots are often treated as a black box in this optimization process, which is the reason why derivative-free optimization methods such as evolutionary algorithms or reinforcement learning are omnipresent. When gradient-based methods are used, models are kept small or rely on finite difference approximations for the Jacobian. This method quickly grows expensive with increasing numbers of parameters, such as found in deep learning. We propose an implementation of a modern physics engine, which can differentiate control parameters. This engine is implemented for both CPU and GPU. Firstly, this paper shows how such an engine speeds up the optimization process, even for small problems. Furthermore, it explains why this is an alternative approach to deep Q-learning, for using deep learning in robotics. Finally, we argue that this is a big step for deep learning in robotics, as it opens up new possibilities to optimize robots, both in hardware and software.
A neuromorphic silicon photonics nonlinear equalizer for optical communications with intensity modulation and direct detection

Katumba, Andrew, Yin, Xin, Dambre, Joni, and Bienstman, Peter

JOURNAL OF LIGHTWAVE TECHNOLOGY 2019

Abs Biblio

We present the design and numerical study of a nonlinear equalizer for optical communications based on silicon photonics and reservoir computing. The proposed equalizer leverages the optical information processing capabilities of integrated photonic reservoirs to combat distortions both in metro links of a few hundred kilometers and in high-speed short-reach intensity-modulation-direct-detection links. We show nonlinear compensation in unrepeated metro links of up to 200 km that outperform electrical feedforward equalizers based equalizers, and ultimately any linear compensation device. For a high-speed short-reach 40Gb/s link based on a distributed feedback laser and an electroabsorptive modulator, and considering a hard decision forward error correction limit of 0.2 x 10(-2), we can increase the reach by almost 10 km. Our equalizer is compact (only 16 nodes) and operates in the optical domain without the need for complex electronic DSP, meaning its performance is not bandwidth constrained. The approach is, therefore, a viable candidate even for equalization techniques far beyond 100G optical communication links.
A personalized and platform-independent behavior control system for social robots in therapy : development and applications

Cao, Hoang-Long, Perre, Greet, Kennedy, James, Senft, Emmanuel, Esteban, Pablo Gomez, De Beir, Albert, Simut, Ramona, Belpaeme, Tony, Lefeber, Dirk, and Vanderborght, Bram

IEEE TRANSACTIONS ON COGNITIVE AND DEVELOPMENTAL SYSTEMS 2019

Abs Biblio

Social robots have been proven beneficial in different types of healthcare interventions. An ongoing trend is to develop (semi-)autonomous socially assistive robotic systems in healthcare context to improve the level of autonomy and reduce human workload. This paper presents a behavior control system for social robots in therapies with a focus on personalization and platform-independence. This system architecture provides the robot an ability to behave as a personable character, which behaviors are adapted to user profiles and responses during the human-robot interaction. Robot behaviors are designed at abstract levels and can be transferred to different social robot platforms. We adopt the component-based software engineering approach to implement our proposed architecture to allow for the replaceability and reusability of the developed components. We introduce three different experimental scenarios to validate the usability of our system. Results show that the system is potentially applicable to different therapies and social robots. With the component-based approach, the system can serve as a basic framework for researchers to customize and expand the system for their targeted healthcare applications.
A power-efficient architecture for on-chip reservoir computing

Sackesyn, Stijn, Ma, Chonghuai, Katumba, Andrew, Dambre, Joni, and Bienstman, Peter

In Artificial neural networks and machine learning – ICANN 2019 : workshop and special sessions 2019

Abs Biblio

Reservoir computing is a neuromorphic computing paradigm which is well suited for hardware implementations. In this work, an enhanced reservoir architecture is introduced as to lower the losses and improve mixing behaviour in silicon photonic reservoir computing designs.
Advanced X-ray CT scanning can boost tree-ring research for earth-system sciences

Bulcke, Jan, Boone, Marijn A, Dhaene, Jelle, Van Loo, Denis, Van Hoorebeke, Luc, Boone, Matthieu, wyffels, Francis, Beeckman, Hans, Van Acker, Joris, and De Mil, Tom

ANNALS OF BOTANY 2019

Abs Biblio

Background and Aims: Tree rings, as archives of the past and biosensors of the present, offer unique opportunities to study influences of the fluctuating environment over decades to centuries. As such, tree-ring-based wood traits are capital input for global vegetation models. To contribute to earth system sciences, however, sufficient spatial coverage is required of detailed individual-based measurements, necessitating large amounts of data. X-ray computed tomography (CT) scanning is one of the few techniques that can deliver such data sets. Methods: Increment cores of four different temperate tree species were scanned with a state-of-the-art X-ray CT system at resolutions ranging from 60 mu m down to 4.5 mu m, with an additional scan at a resolution of 0.8 mu m of a splinter-sized sample using a second X-ray CT system to highlight the potential of cell-level scanning. Calibration-free densitometry, based on full scanner simulation of a third X-ray CT system, is illustrated on increment cores of a tropical tree species. Key Results: We show how multiscale scanning offers unprecedented potential for mapping tree rings and wood traits without sample manipulation and with limited operator intervention. Custom-designed sample holders enable simultaneous scanning of multiple increment cores at resolutions sufficient for tree ring analysis and densitometry as well as single core scanning enabling quantitative wood anatomy, thereby approaching the conventional thin section approach. Standardized X-ray CT volumes are, furthermore, ideal input imagery for automated pipelines with neural-based learning for tree ring detection and measurements of wood traits. Conclusions: Advanced X-ray CT scanning for high-throughput processing of increment cores is within reach, generating pith-to-bark ring width series, density profiles and wood trait data. This would allow contribution to large-scale monitoring and modelling efforts with sufficient global coverage.
All-optical readout for integrated photonic reservoir computing

Ma, Chonghuai, Sackesyn, Stijn, Dambre, Joni, and Bienstman, Peter

In 2019 21st International Conference on Transparent Optical Networks (ICTON) 2019

Abs Biblio

Photonic neuromorphic computing has gained a lot of attention for its strong potential to deliver machine learning computation capability at high bitrates (> 32 Gbps) with very low energy consumption. Reservoir computing is one of the strong candidates that delivers a huge advantage on a real hardware implementation. However, one of the challenges is that current readout systems are the bottleneck of the high-speed link involving heavy power consumption from opto-electrical conversions. In this paper, we present our design of an integrated all-optical readout system that overcomes the challenges with optical weighting elements, which works at 32 Gbps and can deliver computation capability on-chip with one final readout signal channel. We especially compare the memory capability difference between the optical readout scheme with conventional electrical readout and show that optical readout is superior. Furthermore, this paper discusses the problem of some non-volatile optical weighting elements having limited weighting resolution. As a result, our readout system can still perform very well at an low resolution (4 bit) compared to using full resolution weighting elements.
Body randomization reduces the sim-to-real gap for compliant quadruped locomotion

Vandesompele, Alexander, Urbain, Gabriel, Mahmud, Hossain, wyffels, Francis, and Dambre, Joni

FRONTIERS IN NEUROROBOTICS 2019

Abs Biblio

Designing controllers for compliant, underactuated robots is challenging and usually requires a learning procedure. Learning robotic control in simulated environments can speed up the process whilst lowering risk of physical damage. Since perfect simulations are unfeasible, several techniques are used to improve transfer to the real world. Here, we investigate the impact of randomizing body parameters during learning of CPG controllers in simulation. The controllers are evaluated on our physical quadruped robot. We find that body randomization in simulation increases chances of finding gaits that function well on the real robot.
Classification of aortic valve morphology using a three-dimensional convolutional neural network

Dowling, Cameron, Astudillo, Patricio, Brecker, Stephen, De Beule, Matthieu, Mortier, Peter, and Dambre, Joni

In PCR London Valves, Abstracts 2019

Biblio
Closed loop control of a compliant quadruped with spiking neural networks

Vandesompele, Alexander, Urbain, Gabriel, wyffels, Francis, and Dambre, Joni

In Biologically inspired cognitive architectures 2019 2019

Biblio
Combining evolutionary and adaptive control strategies for quadruped robotic locomotion

Massi, Elisa, Vannucci, Lorenzo, Albanese, Ugo, Capolei, Marie Claire, Vandesompele, Alexander, Urbain, Gabriel, Sabatini, Angelo Maria, Dambre, Joni, Laschi, Cecilia, Tolu, Silvia, and Falotico, Egidio

FRONTIERS IN NEUROROBOTICS 2019

Abs Biblio

In traditional robotics, model-based controllers are usually needed in order to bring a robotic plant to the next desired state, but they present critical issues when the dimensionality of the control problem increases and disturbances from the external environment affect the system behavior, in particular during locomotion tasks. It is generally accepted that the motion control of quadruped animals is performed by neural circuits located in the spinal cord that act as a Central Pattern Generator and can generate appropriate locomotion patterns. This is thought to be the result of evolutionary processes that have optimized this network. On top of this, fine motor control is learned during the lifetime of the animal thanks to the plastic connections of the cerebellum that provide descending corrective inputs. This research aims at understanding and identifying the possible advantages of using learning during an evolution-inspired optimization for finding the best locomotion patterns in a robotic locomotion task. Accordingly, we propose a comparative study between two bio-inspired control architectures for quadruped legged robots where learning takes place either during the evolutionary search or only after that. The evolutionary process is carried out in a simulated environment, on a quadruped legged robot. To verify the possibility of overcoming the reality gap, the performance of both systems has been analyzed by changing the robot dynamics and its interaction with the external environment. Results show better performance metrics for the robotic agent whose locomotion method has been discovered by applying the adaptive module during the evolutionary exploration for the locomotion trajectories. Even when the motion dynamics and the interaction with the environment is altered, the locomotion patterns found on the learning robotic system are more stable, both in the joint and in the task space.
Conditional BRUNO : a neural process for exchangeable labelled data

Korshunova, Iryna, Gal, Yarin, Gretton, Arthur, and Dambre, Joni

In Proceedings of the 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2019) 2019

Abs Biblio

We present a neural process that models exchangeable sequences of high-dimensional complex observations conditionally on a set of labels or tags. Our model combines the expressiveness of deep neural networks with the data-efficiency of Gaussian processes, resulting in a probabilistic model for which the posterior distribution is easy to evaluate and sample from, and the computational complexity scales linearly with the number of observations. The advantages of the proposed architecture are demonstrated on a challenging few-shot view reconstruction task which requires generalisation from short sequences of viewpoints.
Design of a bio-inspired compliant quadruped robot for research on closed-loop locomotion control

Urbain, Gabriel, Vandesompele, Alexander, wyffels, Francis, and Dambre, Joni

In 2nd HBP student conference : transdisciplinary research linking neuroscience, brain medicine and computer science : book of abstracts 2019

Biblio
Development of a smart cycling support device for people with an intellectual disability

Deprez, Kenneth, Holvoet, Emiel, Reynders, Esther, Vandepoele, Michelle, Verstringe, Maud, wyffels, Francis, Verstockt, Steven, and Saldien, Jelle

In Project and design literacy as cornerstones of smart education : proceedings of the 4th International Conference on Smart Learning Ecosystems and Regional Development 2019

Abs Biblio

While cycling in traffic, people encounter many different situations with external triggers, that need special attention, such as vehicles, other bicycles, pedestrians, road surface and intersections. Furthermore, people need to remember the route to their destination and take into account the traffic regulations. For people with an intellectual disability, this can be a challenging task.
Enabling automated device size selection for transcatheter aortic valve implantation

Astudillo, Patricio, Mortier, Peter, Bosmans, Johan, De Backer, Ole, De Jaegere, Peter, De Beule, Matthieu, and Dambre, Joni

JOURNAL OF INTERVENTIONAL CARDIOLOGY 2019

Abs Biblio

The number of transcatheter aortic valve implantation (TAVI) procedures is expected to increase significantly in the coming years. Improving efficiency will become essential for experienced operators performing large TAVI volumes, while new operators will require training and may benefit from accurate support. In this work, we present a fast deep learning method that can predict aortic annulus perimeter and area automatically from aortic annular plane images. We propose a method combining two deep convolutional neural networks followed by a postprocessing step. The models were trained with 355 patients using modern deep learning techniques, and the method was evaluated on another 118 patients. The method was validated against an interoperator variability study of the same 118 patients. The differences between the manually obtained aortic annulus measurements and the automatic predictions were similar to the differences between two independent observers (paired diff. of 3.3 +/- 16.8 mm(2) vs. 1.3 +/- 21.1 mm(2) for the area and a paired diff. of 0.6 +/- 1.7 mm vs. 0.2 +/- 2.5 mm for the perimeter). The area and perimeter were used to retrieve the suggested prosthesis sizes for the Edwards Sapien 3 and the Medtronic Evolut device retrospectively. The automatically obtained device size selections accorded well with the device sizes selected by operator 1. The total analysis time from aortic annular plane to prosthesis size was below one second. This study showed that automated TAVI device size selection using the proposed method is fast, accurate, and reproducible. Comparison with the interobserver variability has shown the reliability of the strategy, and embedding this tool based on deep learning in the preoperative planning routine has the potential to increase the efficiency while ensuring accuracy.
From leaf to label : a robust automated workflow for stomata detection

Meeus, Sofie, wyffels, Francis, and Bulcke, Jan

In BIODIVERSITY INFORMATION SCIENCE AND STANDARDS 2019

Biblio
Generating spatial referring expressions in a social robot : dynamic vs. non-ambiguous

Wallbridge, Christopher D., Lemaignan, Severin, Senft, Emmanuel, and Belpaeme, Tony

FRONTIERS IN ROBOTICS AND AI 2019

Abs Biblio

Generating spatial referring expressions is key to allowing robots to communicate with people in an environment. The focus of most algorithms for generation is to create a non-ambiguous description, and how best to deal with the combination explosion this can create in a complex environment. However, this is not how people naturally communicate. Humans tend to give an under-specified description and then rely on a strategy of repair to reduce the number of possible locations or objects until the correct one is identified, what we refer to here as a dynamic description. We present here a method for generating these dynamic descriptions for Human Robot Interaction, using machine learning to generate repair statements. We also present a study with 61 participants in a task on object placement. This task was presented in a 2D environment that favored a non-ambiguous description. In this study we demonstrate that our dynamic method of communication can be more efficient for people to identify a location compared to one that is non-ambiguous.
Highly parallel simulation and optimization of photonic circuits in time and frequency domain based on the deep-learning framework PyTorch

Laporte, Floris, Dambre, Joni, and Bienstman, Peter

SCIENTIFIC REPORTS 2019

Abs Biblio

We propose a new method for performing photonic circuit simulations based on the scatter matrix formalism. We leverage the popular deep-learning framework PyTorch to reimagine photonic circuits as sparsely connected complex-valued neural networks. This allows for highly parallel simulation of large photonic circuits on graphical processing units in time and frequency domain while all parameters of each individual component can easily be optimized with well-established machine learning algorithms such as backpropagation.
Information processing and computation with photonic reservoir systems

Dambre, Joni

2019

Biblio
Leveraging robotics research for children with autism : a review

Ismail, Luthffi, Verhoeven, Thibault, Dambre, Joni, and wyffels, Francis

INTERNATIONAL JOURNAL OF SOCIAL ROBOTICS 2019

Biblio
Leveraging deep learning for eye contact analysis in child-robot interaction

Ismail, Luthffi, Hanapiah, Fazah Akhtar, Dambre, Joni, Belpaeme, Tony, and wyffels, Francis

2019

Biblio
Neuromorphic information processing using silicon photonics

Bienstman, Peter, Dambre, Joni, Katumba, Andrew, Freiberger, Matthias, Laporte, Floris, Lugnan, Alessio, Sackesyn, Stijn, and Ma, Chonghuai

In Active Photonic Platforms XI 2019

Biblio
Non-linear signal equalisation using silicon photonic reservoir computing

Bienstman, Peter, Dambre, Joni, Katumba, Andrew, Freiberger, Matthias, Laporte, Floris, Lugnan, Alessio, Sackesyn, Stijn, Ma, Chonghuai, and Gooskens, Emmanuel

In ECOC Machine Learning Workshop 2019, Abstracts 2019

Biblio
Photontorch : simulation and optimization of large photonic circuits using the deep learning framework PyTorch

Laporte, Floris, Dambre, Joni, and Bienstman, Peter

In 2019 IEEE Photonics Society Summer Topical Meeting Series (SUM) 2019

Abs Biblio

We propose a new framework for simulating and optimizing large photonic circuits. The framework utilizes gpu-acceleration to efficiently simulate the circuits in a highly parallel manner, while optimization is achieved through backpropagation by essentially viewing the network as a sparsely connected recurrent neural network.
Populations of spiking neurons for reservoir computing : closed loop control of a compliant quadruped

Vandesompele, Alexander, Urbain, Gabriel, wyffels, Francis, and Dambre, Joni

COGNITIVE SYSTEMS RESEARCH 2019

Abs Biblio

Compliant robots can be more versatile than traditional robots, but their control is more complex. The dynamics of compliant bodies can however be turned into an advantage using the physical reservoir computing frame- work. By feeding sensor signals to the reservoir and extracting motor signals from the reservoir, closed loop robot control is possible. Here, we present a novel framework for implementing central pattern generators with spik- ing neural networks to obtain closed loop robot control. Using the FORCE learning paradigm, we train a reservoir of spiking neuron populations to act as a central pattern generator. We demonstrate the learning of predefined gait patterns, speed control and gait transition on a simulated model of a compliant quadrupedal robot.
Reinforcement learning and insight in the artificial pigeon

Colin, Thomas R., and Belpaeme, Tony

In Creativity + cognition + computation : 41st Annual Meeting of the Cognitive Science Society (CogSci 2019) 2019

Abs Biblio

The phenomenon of insight (also called "Aha!" or "Eureka!" moments) is considered a core component of creative cognition. It is also a puzzle and a challenge for statistics-based approaches to behavior such as associative learning and reinforcement learning. We simulate a classic experiment on insight in pigeons using deep Reinforcement Learning. We show that prior experience may produce large and rapid performance improvements reminiscent of insights, and we suggest theoretical connections between concepts from machine learning (such as the value function or overfitting) and concepts from psychology (such as feelings-of-warmth and the einstellung effect). However, the simulated pigeons were slower than the real pigeons at solving the test problem, requiring a greater amount of trial and error: their "insightful" behavior was sudden by comparison with learning from scratch, but slow by comparison with real pigeons. This leaves open the question of whether incremental improvements to reinforcement learning algorithms will be sufficient to produce insightful behavior.
Response pattern of child-robot interaction among children with cognitive impairment

Ismail, Luthffi, Hanapiah, Fazah Akhtar, Belpaeme, Tony, and wyffels, Francis

In Proceedings of HUSO 2019, the fifth international conference on human and social analytics 2019

Abs Biblio

Children with cognitive impairments reported having an engaging initial response in child-robot interaction. However, the interaction modules for child-robot interaction vary and depend on the aim of the studies in the literature. In this study, we designed child-robot interaction modules in order to help children with cognitive impairment in improving their social interaction skills. Hence, we would like to report the overall pattern of their response for each module in 3 sessions of child-robot interaction. Their response pattern for each module is important in future behaviour analysis, especially with regards to their attention skills analysis and their eye contact engagement analysis during child-robot interaction.
Robot-enhanced therapy developing and validating a supervised autonomous robotic system for autism spectrum disorders therapy

Cao, Hoang-Long, Esteban, Pablo G., Bartlett, Madeleine, Baxter, Paul, Belpaeme, Tony, Billing, Erik, Cai, Haibin, Coeckelbergh, Mark, Costescu, Cristina, David, Daniel, De Beir, Albert, Hernandez, Daniel, Kennedy, James, Liu, Honghai, Matu, Silviu, Mazel, Alexandre, Pandey, Amit, Richardson, Kathleen, Senft, Emmanuel, Thill, Serge, Perre, Greet, Vanderborght, Bram, Vernon, David, Wakunuma, Kutoma, Yu, Hui, Zhou, Xiaolong, and Ziemke, Tom

IEEE ROBOTICS & AUTOMATION MAGAZINE 2019

Biblio
Second language tutoring using social robots : L2TOR-the movie

Vogt, Paul, Berghe, Rianne, Haas, Mirjam, Hoffman, Laura, Kanero, Junko, Mamus, Ezgi, Montanier, Jean-Marc, Oranc, Cansu, Oudgenoeg-Paz, Ora, Garcia, Daniel Hernandez, Papadopoulos, Fotios, Schodde, Thorsten, Verhagen, Josje, Wallbridge, Christopher D., Willemsen, Bram, Wit, Jan, Belpaeme, Tony, Goksun, Tilbe, Kopp, Stefan, Krahmer, Emiel, Kuntay, Aylin C., Leseman, Paul, and Pandey, Amit Kumar

In HRI ’19: 2019 14TH ACM/IEEE INTERNATIONAL CONFERENCE ON HUMAN-ROBOT INTERACTION 2019

Biblio
Second language tutoring using social robots : a large-scale study

Vogt, Paul, Berghe, Rianne, Haas, Mirjam, Hoffman, Laura, Kanero, Junko, Mamus, Ezgi, Montanier, Jean-Marc, Oranc, Cansu, Oudgenoeg-Paz, Ora, Garcia, Daniel Hernandez, Papadopoulos, Fotios, Schodde, Thorsten, Verhagen, Josje, Wallbridge, Christopher D., Willemsen, Bram, Wit, Jan, Belpaeme, Tony, Goksun, Tilbe, Kopp, Stefan, Krahmer, Emiel, Kuntay, Aylin C., Leseman, Paul, and Pandey, Amit Kumar

In HRI ’19: 2019 14TH ACM/IEEE INTERNATIONAL CONFERENCE ON HUMAN-ROBOT INTERACTION 2019

Abs Biblio

We present a large-scale study of a series of seven lessons designed to help young children learn English vocabulary as a foreign language using a social robot. The experiment was designed to investigate 1) the effectiveness of a social robot teaching children new words over the course of multiple interactions (supported by a tablet), 2) the added benefit of a robot’s iconic gestures on word learning and retention, and 3) the effect of learning from a robot tutor accompanied by a tablet versus learning from a tablet application alone. For reasons of transparency, the research questions, hypotheses and methods were preregistered. With a sample size of 194 children, our study was statistically well-powered. Our findings demonstrate that children are able to acquire and retain English vocabulary words taught by a robot tutor to a similar extent as when they are taught by a tablet application. In addition, we found no beneficial effect of a robot’s iconic gestures on learning gains.
Sensing-enhanced therapy system for assessing children with autism spectrum disorders : a feasibility study

Cai, Haibin, Fang, Yinfeng, Costescu, Cristina, David, Daniel, Billing, Erik, Ziemke, Tom, Thill, Serge, Belpaeme, Tony, Vanderborght, Bram, Vernon, David, Richardson, Kathleen, and Ju, Zhaojie

IEEE SENSORS JOURNAL 2019

Biblio
Should beat gestures be learned or designed? A benchmarking user study

Wolfert, Pieter, Kucherenko, T., Kjelström, H., and Belpaeme, Tony

In ICDL-EPIROB 2019 : workshop on naturalistic non-verbal and affective human-robot interactions, Proceedings 2019

Biblio
Silicon photonics reservoir computing at 32 Gbit/s

Bienstman, Peter, Dambre, Joni, Katumba, Andrew, Freiberger, Matthias, Laporte, Floris, Lugnan, Alessio, Sackesyn, Stijn, Ma, Chonghuai, and Gooskens, Emmanuel

In 2019 Workshop on Dynamical Systems and Brain-Inspired Information Processing, Abstracts 2019

Biblio
Teaching robots social autonomy from in situ human guidance

Senft, Emmanuel, Lemaignan, Séverin, Baxter, Paul E., Bartlett, Madeleine, and Belpaeme, Tony

SCIENCE ROBOTICS 2019

Abs Biblio

Striking the right balance between robot autonomy and human control is a core challenge in social robotics, in both technical and ethical terms. On the one hand, extended robot autonomy offers the potential for increased human productivity and for the off-loading of physical and cognitive tasks. On the other hand, making the most of human technical and social expertise, as well as maintaining accountability, is highly desirable. This is particularly relevant in domains such as medical therapy and education, where social robots hold substantial promise, but where there is a high cost to poorly performing autonomous systems, compounded by ethical concerns. We present a field study in which we evaluate SPARC (supervised progressively autonomous robot competencies), an innovative approach addressing this challenge whereby a robot progressively learns appropriate autonomous behavior from in situ human demonstrations and guidance. Using online machine learning techniques, we demonstrate that the robot could effectively acquire legible and congruent social policies in a high-dimensional child-tutoring situation needing only a limited number of demonstrations while preserving human supervision whenever desirable. By exploiting human expertise, our technique enables rapid learning of autonomous social and domain-specific policies in complex and nondeterministic environments. Last, we underline the generic properties of SPARC and discuss how this paradigm is relevant to a broad range of difficult human-robot interaction scenarios.
Towards automatic sign language corpus annotation using deep learning

De Coster, Mathieu, Van Herreweghe, Mieke, and Dambre, Joni

In 6th Workshop on Sign Language Translation and Avatar Technology (SLTAT) 2019

Abs Biblio

Sign classification in sign language corpora is a challenging problem that requires large datasets. Unfortunately, only a small portion of those corpora is labeled. To expedite the annotation process, we propose a gloss suggestion system based on deep learning. We improve upon previous research in three ways. Firstly, we use a proven feature extraction method called OpenPose, rather than learning end-to-end. Secondly, we propose a more suitable and powerful network architecture, based on GRU layers. Finally, we exploit domain and task knowledge to further increase the accuracy. We show that we greatly outperform the previous state of the art on the used dataset. Our method can be used for suggesting a top 5 of annotations given a video fragment that is selected by the corpus annotator. We expect that it will expedite the annotation process to the benefit of sign language translation research.
Towards generating spatial referring expressions in a social robot : dynamic vs non-ambiguous

Wallbridge, Christopher D., Lemaignan, Severin, Senft, Emmanuel, and Belpaeme, Tony

In HRI ’19: 2019 14TH ACM/IEEE INTERNATIONAL CONFERENCE ON HUMAN-ROBOT INTERACTION 2019

Abs Biblio

We present in this paper our work towards a new dynamic method of generating spatial referring expressions. While people are generally ambiguous in their description of locations, previous methods of artificial generation mostly considered non-ambiguous descriptions. However, to increase the naturalness of interaction and share workload in the communication, robots should be able to generate language in a more dynamic way. Our method initially produces ambiguous spatial referring expressions followed by dynamically generating repair statements. We built a classifier using data from 18 participants as they described locations to each other. We perform a preliminary analysis on this method using two further pilot studies.
Towards safe and efficient pre-operative planning of transcatheter mitral valve interventions

Astudillo, Patricio, De Beule, Matthieu, Dambre, Joni, and Mortier, Peter

MORPHOLOGIE 2019

Biblio
Training passive photonic reservoirs with integrated optical readout

Freiberger, Matthias, Katumba, Andrew, Bienstman, Peter, and Dambre, Joni

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2019

Abs Biblio

As Moore’s law comes to an end, neuromorphic approaches to computing are on the rise. One of these, passive photonic reservoir computing, is a strong candidate for computing at high bitrates (>10 Gb/s) and with low energy consumption. Currently though, both benefits are limited by the necessity to perform training and readout operations in the electrical domain. Thus, efforts are currently underway in the photonic community to design an integrated optical readout, which allows to perform all operations in the optical domain. In addition to the technological challenge of designing such a readout, new algorithms have to be designed in order to train it. Foremost, suitable algorithms need to be able to deal with the fact that the actual on-chip reservoir states are not directly observable. In this paper, we investigate several options for such a training algorithm and propose a solution in which the complex states of the reservoir can be observed by appropriately setting the readout weights, while iterating over a predefined input sequence. We perform numerical simulations in order to compare our method with an ideal baseline requiring full observability as well as with an established black-box optimization approach (CMA-ES).
What can you see? Identifying cues on internal states from the movements of natural social interactions

Bartlett, Madeleine E., Edmunds, Charlotte E. R., Belpaeme, Tony, Thill, Serge, and Lemaignan, Severin

FRONTIERS IN ROBOTICS AND AI 2019

Abs Biblio

In recent years, the field of Human-Robot Interaction (HRI) has seen an increasing demand for technologies that can recognize and adapt to human behaviors and internal states (e.g., emotions and intentions). Psychological research suggests that human movements are important for inferring internal states. There is, however, a need to better understand what kind of information can be extracted from movement data, particularly in unconstrained, natural interactions. The present study examines which internal states and social constructs humans identify from movement in naturalistic social interactions. Participants either viewed clips of the full scene or processed versions of it displaying 2D positional data. Then, they were asked to fill out questionnaires assessing their social perception of the viewed material. We analyzed whether the full scene clips were more informative than the 2D positional data clips. First, we calculated the inter-rater agreement between participants in both conditions. Then, we employed machine learning classifiers to predict the internal states of the individuals in the videos based on the ratings obtained. Although we found a higher inter-rater agreement for full scenes compared to positional data, the level of agreement in the latter case was still above chance, thus demonstrating that the internal states and social constructs under study were identifiable in both conditions. A factor analysis run on participants’ responses showed that participants identified the constructs interaction imbalance, interaction valence and engagement regardless of video condition. The machine learning classifiers achieved a similar performance in both conditions, again supporting the idea that movement alone carries relevant information. Overall, our results suggest it is reasonable to expect a machine learning algorithm, and consequently a robot, to successfully decode and classify a range of internal states and social constructs using low-dimensional data (such as the movements and poses of observed individuals) as input.