My name is Marwen Belkaid. I am a reseacher in Robotics and Cognitive Science. I received my PhD from the University of Cergy-Pontoise in December 2016. Priori to that, I received my degree in Computer Science engineering from the National School of Engineers of Tunis (Ecole Nationale d'Ingenieurs de Tunis), Tunisia in June 2012 and my MSc in Robotics and Artificial Intelligence from the University of Cergy-Pontoise and ENSEA (Ecole Nationale Superieure de l'Electronique et de ses Applications), France in September 2013.
I am currently lecturer/research associate at the University of Cergy-Pontoise within the Neurocybernetics team of ETIS lab (Equipes Traitement de l'Information et Systemes).
My research interests include questions about artificial and biological intelligence. Currently, my work consists in proposing computational models based on artificial neural networks and evaluating those models through robotic implementations. The goal is twofold: designing efficient artificial systems and understanding biological cognition.
My research in robotics and cognitive science encompasses two goals: designing efficient artificial systems and understanding biological cognition. In this process of understanding by design, I am interested in three topics: emotion, navigation and social interaction. In this context, the central question underlying my research interests is the study of how agents (e.g. humans, animals, robots) interact efficiently with their physical and social environment.
Emotion: In the last three decades, the research on emotion has witnessed significant expansion. Although often opposed to the rational intelligence, it is now becoming more and more admitted that emotional processes are tightly linked to cognition and constitute a foundation for the mind. Assessing events, focusing attention, enhancing communication, and motivating cognition and action are among the most consensual functions of emotion. Given this large variety of functions, a clear and comprehensive understanding of emotional phenomena is still lacking.
Navigation: Navigation provides a good framework for the study of cognition. Indeed, the act of moving, or finding a way, from one place to another is a fundamental cognitive function that is common to most species. On the one hand, it involves a variety of computational processes related to perception, attention, memory, decision and motricity. On the other hand, the resulting behavior can be expressed in relatively few dimensions. Studying navigation also allows for addressing broader questions about how individuals interact with their environment -- e.g. representing action potentials in the surrounding space, selecting goals/strategies, decision-making.
Social interaction: Social interactions are crucial for the survival and well-being of humans and animals. Non-verbal communication -- based on postures, gestures, facial expressions and rhythmical coordination -- forms the bedrock of these interactions. It is essential to understand the mechanisms that allow individuals to recognize and use such information. Moreover, the most evolved species have a remarkable competence that consists in projecting one's mental state on others (or putting oneself in the place of others). Humans for example can detect, interpret and even manipulate others mental states -- a competence often referred to as the theory of mind. Most of these aspects of social cognition can still be considered challenging puzzles.
To address these research topics, computational modeling is a powerful tool. It allows for testing hypotheses about how the involved processes organize and take place. Also, going from theoretical concepts to concrete implementations reveals ambiguities and forces more explicit definitions. To do so, artificial neural networks provide a compelling computational framework. In an analogy with biological systems, information processing is distributed over a collection of interconnected units that perform elementary operations in parallel. In this context, mental processes are dynamic patterns of activity that depend on internal constraints determining how connections are changed and how units (neurons) are activated by their inputs. The global coherence comes from these local properties. Additionally, implementations on robotic platforms can be very insightful. Indeed, as embodied and situated systems, robots have to deal with noise and uncertainty -- whether it is due to their sensors and actuators or to their environment. Thus, they are crucial when it comes to evaluating the model in real world situations.
|Belkaid M., Interactions between cognitive and emotional mechanisms: a study in neuromimetic mobile and social robotics, PhD thesis, 2016.||pdf   bibTeX|
|Belkaid M., Cuperlier N. and Gaussier P. Combining local and global visual information in context-based neurorobotic navigation. Proceedings of the International Joint Conference on Neural Networks (IJCNN), 2016.||pdf   bibTeX|
|Belkaid M., Lesueur-Grand C., Mostafaoui, G., Cuperlier N. and Gaussier P. Learning sensorimotor navigation using synchrony-based partner selection. Proceedings of the International Conference on Artificial Intelligence and Robotics (ICAIR), 2016.||pdf   bibTeX|
|Belkaid M., Cuperlier N. and Gaussier P. Emotional Modulation of peripersonal space as a way to represent reachable and comfort areas. Proceedings of the International Conference on Intelligent Robots and Systems (IROS), 2015.||pdf   bibTeX|
|Belkaid M., Cuperlier N. and Gaussier P. Emotional Modulation of peripersonal space impact the way robots interact. Proceedings of the European Conference on Artificial Life (ECAL), 2015.||pdf   bibTeX|
|Belkaid M., Sabouret N. A logical model of Theory of Mind for virtual agents in the context of job interview simulation. Proceedings of the International Workshop on Intelligent Digital Games for Empowerment and Inclusion (IDGEI) at the ACM International Conference on Intelligent User Interfaces (IUI), 2014.||pdf   bibTeX|
|Jauffret A., Belkaid M., Cuperlier N., Gaussier P. and Tarroux P. Frustration as a way toward autonomy and self-improvement in robotic navigation. Proceedings of the Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics (ICDL-EpiRob), 2013.||pdf   bibTeX|