Reinforcement Learning and Control as Probabilistic Inference
Add to your list(s)
Download to your calendar using vCal
If you have a question about this talk, please contact jg801.
Reinforcement learning and inference offer powerful frameworks for solving sequential decision-making problems. While classically reinforcement learning and inference have been studied independently, it is possible to frame the decision-making problem itself as inference in a graphical model. This formalism allows us to utilize well-known approximate inference techniques and gives insights into how we can extend the model. The basic underlying framework has been proposed in the literature in several forms before, and remains an important source
of inspiration for novel algorithms.
This talk is part of the Machine Learning Reading Group @ CUED series.
This talk is included in these lists:
Note that ex-directory lists are not shown.
|