Theoretical Advances in Reinforcement Learning and Control

Description

Back to top

Reinforcement learning (RL) and control theory are concerned with training intelligent agents to make sequential decisions by interacting with an environment. In both formulations, an agent learns to navigate its surroundings through a process of trial and error, receiving feedback in the form of rewards or penalties based on the actions it takes. The agent’s goal is to learn a policy, a mapping from states to actions, that maximizes the cumulative reward over time.

In the past few years, there has been a notable increase in enthusiasm for RL and the interplay between learning and control. The surge of interest is driven by the compelling application of RL and control methods to diverse challenges in artificial intelligence, robotics, and the natural sciences. Numerous breakthroughs owe their success to large-scale computational resources, creative deployment of adaptable neural network structures and training approaches, as well as both modern and traditional decision-making algorithms. Nevertheless, there remains a significant gap in our understanding regarding the conditions, reasons, and the degree to which these algorithms effectively operate. Such a challenge has drawn significant attention from various communities including computer science, numerical analysis, artificial intelligence, control theory, operations research, and statistics. This program aims to advance the theoretical foundations of reinforcement learning (RL) and control, and foster new collaborations between these researchers.

Organizers

Back to top

X C

E H

C M

N S

A Z

Program Workshops

Back to top

Application

Back to top

Applications received by January 10, 2025 will have the first priority for consideration.

In order to apply for this program, you must first have an IMSI account and be logged in. Please login or create an account, and then return to this page to apply for the program.