Imperial College London > Talks@ee.imperial > Yiannis Demiris's list > Machine Learning for Sensorimotor Control

Machine Learning for Sensorimotor Control

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Yiannis Demiris.

Humans and other biological systems are very adept at performing fast, complicated control tasks in spite of large sensorimotor delays while being fairly robust to noise and perturbations. For example, one is able to react accurately and fast to catch a speeding ball while at the same time being flexible enough to ‘give-in’ when obstructed during the execution of a task.

There are various components involved in achieving such levels of robustness, accuracy and safety in anthropomorphic robotic systems. Broadly, speaking challenges lie in the domain of robust sensing, flexible planning, appropriate representation and learning dynamics under various contexts. Statistical Machine Learning provides ideal tools to deal with these challenges, especially in tackling issues like partial observability, noise, redundancy resolution, high dimensionality and the ability to perform and adapt in real time.

In my talk, I will talk about (a) novel techniques we have developed for real time acquisition of non-linear dynamics in a data driven manner, (b) techniques for automatic low-dimensional (latent space) representation of complex movement policies and trajectories and© planning methods capable of dealing with redundancy (e.g. variable impedance) and adaptation in the Optimal Feedback Control framework. Some of the techniques developed, in turn, provide novel insights into modeling human motor control behavior.

Videos of learning in high dimensional movement systems like anthropomorphic limbs (KUKA robot arm, SARCOS dexterous arm, iLIMB etc.) and humanoid robots (HONDA ASIMO , DB) will serve to validate the effectiveness of these machine learning techniques in real world applications.

Short Bio: Sethu Vijayakumar is the Director of the Institute for Perception, Action and Behavior (IPAB) in the School of Informatics at the University of Edinburgh. Since August 2007, he holds a Senior Research Fellowship of the Royal Academy of Engineering, co-funded by Microsoft Research in Learning Robotics. He also holds additional appointments as an Adjunct Faculty of the University of Southern California (USC), Los Angeles and as a Visiting Research Scientist at the RIKEN Brain Science Institute, Japan. His research interest spans a broad interdisciplinary curriculum involving basic research in the fields of statistical machine learning, robotics, human motor control, Bayesian inference techniques and computational neuroscience.

Personal Webpage: http://www.homepages.inf.ed.ac.uk/svijayak Research Group Webpage: http://www.ipab.inf.ed.ac.uk/slmc

This talk is part of the Yiannis Demiris's list series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

Changes to Talks@imperial | Privacy and Publicity