Method IV-RL for Estimating Error in Reinforcement Learning

The seminar is based on the article “Sample Efficient Deep Reinforcement Learning Via Uncertainty Estimation” by Vincent Mai, Kaustubh Mani and Liam Paull, which discusses a new approach for error estimation in reinforcement learning .

In the first part of the seminar, the subject area and the problems associated with it will be considered.

In the second part of the seminar, the IV-RL error estimation method will be described.

In the third and final part, the results of experiments, use cases and further ways of research in this topic will be considered.

Speaker: Sorokumov Sergey Valerievich, student of the department. MOEM SPbGETU "LETI"

Date and time: February 15 19:00 (Moscow time)

Register via the link