Biography

Jun Tani received a doctor of engineering degree in electrical engineering from Sophia Univ. in Tokyo in 1995. He worked for Sony Corp. and later for Sony Computer Science Lab as a researcher from 1990 to 2001. Then, he worked at Riken Brain Science Institute from 2001 to 2012 where he has been a PI of Lab. for Behavior and Dynamic Cognition. He became a full professor of Electrical Engineering in Korean Advanced Inst. of Science and Technology in Korea in 2012. Then, he moved to Okinawa Institute of Science and Technology for starting Cognitive Neurorobotics Research Unit in 2017. His research interests include neurorobotics, deep learning, complex systems, brain science, developmental psychology, and philosophy of mind. He is an author of “Exploring Robotic Minds: Actions, Symbols, and Consciousness as Self-Organizing Dynamic Phenomena.” published from Oxford Univ. Press in 2016.

Abstract

How can compositionality develop through the self-exploration as well as the supervised tutoring?

This talk introduces possible mechanisms to develop compositionality using multiple spatio-temporal scale constraints imposed on modelled internal neurodynamics in terms of downward causation. I will explain each case in this development using supervised training, reward-based reinforcement learning, and self-exploration with intrinsic motivation. I will discuss how these different learning schemes can be integrated to achieve efficient development of compositionality in agents.