 |
 |
|
|
|
Frédéric Kaplan - Detailed biography
Frédéric Kaplan graduated as an engineer of the Ecole Nationale Supérieure des Télécommunications in Paris and received a PhD degree in Artificial Intelligence from the University Paris VI. In his studies and research projects, he has always shown a double interest both in understanding principles governing living systems and in the construction of biologically inspired systems.
He conducted his PhD thesis under the supervision of Luc Steels and Alexis Drogoul on models of the collective dynamics involved in “cultural” self-organization of large populations of agents. In 2001, he published a first book out of this research [1] and continued working on formal aspects of such systems since then [2].
He started working at Sony Computer Science Laboratory in 1997 and spent each year several weeks in Tokyo in order to collaborate on the conception of the new four-legged robots and biped humanoids. Several technological innovations that resulted from this research were seen as crucial from an industrial perspective (Technology award [best innovation of year] at the Sony Research Forum 1999, best paper in 2000, Sony Implemented invention award in 2005).
Since 2000, he also worked closely with ethologists at Eotvos University in Budapest (HU) investigating how robot models could help make progress in this field and give insight for building trainable machines [3]. This collaboration resulted in the publication of one of the first papers using robots in an ethological experiment with mammals [4].
More recently, he initiated collaborations with developmental psychologists and have been active in creating an interdisciplinary domain entitled epigenetic robotics in which robots serve as tools to investigate the dynamics of cognitive development [5]. These interdisciplinary researches have been the subject of a second general audience book [6], published in 2005 and selected as one of the best French scientific books of the year (Selection Prix Roberval 2005).
In 2002, he started to develop with his colleague Pierre-Yves Oudeyer, a new family of theoretical architectures based on the notion of intrinsic motivation systems. In particular, he explored the principles that permit an agent to compute its own learning progress in various situations and act in order to maximize it [7,8]. The key idea was to introduce a meta-prediction system responsible for modelling how well the system is learning. As each technique is characterized by its own learning biases that make it appropriate for some contexts but not others, the idea was to let the system evaluate and decide by itself the most appropriate context to use. Such a system would go far beyond traditional learning systems because it can learn how to learn.
In 2005, he worked with designers of the Ecole Cantonale d’Art de Lausanne to create environments adapted where robots could learn in an open-ended manner. Together, they built the first robot’s playroom, a series of objects and furniture adapted to the morphology and learning capacities of robots.
In 2006, he was invited by Pierre Dillenbourg to supervise a new team on the design of interactive furnitures, inside the CRAFT group of the Ecole Polytechnique Fédérale de Lausanne. He devotes now most of his time to this new subject.
1. Kaplan, F. (2001) La naissance d’une langue chez les robots, Hermès Science, Paris
2. Kaplan, F. (2005) Simple models of distributed coordination, Connection Science, 17,(3-4) :249-270
3. Kaplan, F., Oudeyer, P-Y., Kubinyi, E. and Miklosi, A. (2002) Robotic clicker training, Robotics and Autonomous Systems, 38(3-4):197-206.
4. Kubinyi, E. , Miklosi, A. Kaplan, F. Gacsi, M. Topal, J. Csnayi, V. (2004) Social behaviour of dogs encountering AIBO, an animal-like robot in a neutral and in a feeding situation, Behavioural Processes, 65 (3 ) : 231-239
5. Kaplan, F. Oudeyer, P-Y, Revel, A. Gaussier, P., Nadel, J., Berthouze, L., Kozima, H., Prince, C. Balkenius, C. (2006), Proceedings of the Sixth International Workshop on Epigenetic Robotics: Modeling Cognitive Development in Robotic Systems, LUCS, 128, ISBN 91-974741-6-9
6. Kaplan, F. (2005) Les machines apprivoisees, Vuibert, Paris
7. Kaplan, F. and Oudeyer, P-Y. (2004) Maximizing learning progress: an internal reward system for development, LNAI 3139, Springer-Verlag ,259-270
8. Oudeyer, P-Y. and Kaplan, F. (2006) The discovery of communication, Connection Science, 18 (2) : 189- 206
|
|
|
|
|
|
|
