Conferences/Symposia
Self-Calibration of Sensorimotor Loops with Active Efficient Coding
Title: Self-Calibration of Sensorimotor Loops with Active Efficient Coding
Speaker: Jochen Triesch, Ph.D.
Date, Time, and Venue: March 10, 2015; 15:00-16:30; IS Lecture Hall
Abstract:
The goal of biological and artificial perceptual systems is to provide
useful knowledge about the environment and to encode this information
efficiently. Importantly, perception is an active process that often
involves the movement of sense organs such as the eyes. This active nature
of perception has typically been neglected in popular theories describing
how biological nervous systems learn sensory representations. Here we
present an approach for the self-calibration of sensorimotor loops for
active perception that treats the learning of sensory representations and
the learning of movements of the sense organs in an integrated manner. In
this „active efficient coding" approach, a generative model learns to encode
the sensory data while a reinforcement learner directs the sense organs so
as to make the generative model work as efficiently as possible. To this
end, the reinforcement learner receives an intrinsically generated reward
signal that measures the encoding quality currently obtained by the
generative model. We show how this approach allows a robot vision system to
self-calibrate vergence and pursuit eye movements in a completely autonomous
fashion. Somewhat surprisingly, the approach also offers a new perspective
on the development of imitation abilities.
Biosktech: ( <http://fias.uni-frankfurt.de/~triesch/resume.html>
http://fias.uni-frankfurt.de/~triesch/resume.html)
Jochen Triesch received the Diploma and Ph.D. degrees in physics from the
Ruhr-University Bochum, Germany, in 1994 and 1999, respectively. He spent
two years as a postdoctoral Fellow in the Computer Science Department of the
University of Rochester, NY. In 2001, he joined the faculty of the Cognitive
Science Department of the University of California San Diego, La Jolla, CA,
as an Assistant Professor. In 2005, he became a Fellow of the Frankfurt
Institute for Advanced Studies (FIAS), Frankfurt am Main, Germany. Since
2007 he has been the Johanna Quandt Research Professor for Theoretical Life
Sciences at FIAS, where he also serves as the Vice Chairman of the Board of
Directors. He also holds Professorships in the Department of Physics and the
Department of Computer Science and Mathematics at the Goethe University in
Frankfurt am Main, Germany. His research interests span computational
neuroscience, cognitive science, developmental robotics, and biologically
inspired computer vision.
Primary Contact Person: Assistant Professor Sungmoon Jeong
(jeongsm@jaist.ac.jp, ext. 1362)
