Decision Making and Planning as Probabilistic Inference. Marc Toussaint

Thursday, 12 September, 2013 - 09:30 to 11:00

I'll first give a general introduction on formalizing planning (more precisely Stochastic Optimal Control problems) in terms of probabilistic inference (more precisely: KL divergence minimization), providing a unifying perspective of previous approaches. I'll then focus on novel algorithms that can be derived from this formulation, including an efficient model-free RL algorithm that I find very interesting. Besides providing novel efficient algorithms, the motivation of this work is to unify perspectives from control theory and machine learning, and perhaps to contribute to the discussion how neural systems might solve stochastic optimal control and RL problems. This is based on joint work with Konrad Rawlik and Sethu Vijayakumar from U Edinburgh.

Related papers are:


Prof Marc Toussaint
Machine Learning and Robotics Lab.,
Uni Stuttgart,
Stuttgart, Germany