Bayesian reinforcement learning: Algorithms and models

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

• Dr. Christos Dimitrakakis, Computer Science and Engineering, Chalmers University of Technology, Göteborg
• Monday 17 March 2014, 16:00-17:00
• UG05, Learning Centre.

If you have a question about this talk, please contact Leandro Minku.

Host: Prof. Jeremy Wyatt

We will cover a couple of topics in decision making under uncertainty in the context of reinforcement learning. This is the problem of learning how to act optimally in an unknown environmnet only by interaction. The focus will be in Bayesian methods, where uncertainty about the environment is modelled as a probabilistic belief. The problems we face are two-fold: Firstly, how can we plan optimally given that our current belief will change as soon as we obtain more information? While this problem is formally solved, approximate algorithms are required in practice. Secondly, what kind of models should we use for representing our current belief that are simultaneously widely applicable and easy to integrate with planning? The main challenges here are avoiding an explosion in complexity as we accumulate more data and incoroprating diverse prior knowledge into the problem.

Speaker’s home page: http://www.cse.chalmers.se/%7Echrdimi/

This talk is part of the Artificial Intelligence and Natural Computation seminars series.

This talk is included in these lists:

• Artificial Intelligence and Natural Computation seminars
• Computer Science Departmental Series
• Computer Science Distinguished Seminars
• UG05, Learning Centre
• computer sience

Note that ex-directory lists are not shown.