THE NATIONAL-SCIENCE-FOUNDATION WORKSHOP ON REINFORCEMENT LEARNING

Reinforcement learning has become one of the most actively studied lea rning frameworks in the area of intelligent autonomous agents. This ar ticle describes the results of a three-day meeting of leading research ers in this area that was sponsored by the National Science Foundation . Because reinforcement learning is an interdisciplinary topic, the wo rkshop brought together researchers from a variety of fields, includin g machine learning, neural networks, Al, robotics, and operations rese arch. Thirty leading researchers from the United States, Canada, Europ e, and Japan, representing from many different universities, governmen t, and industrial research laboratories participated in the workshop. The goals of the meeting were to (1) understand limitations of current reinforcement-learning systems and define promising directions for fu rther research; (2) clarify the relationships between reinforcement le arning and existing work in engineering fields, such as operations res earch; and (3) identify potential industrial applications of reinforce ment learning.