What is So Interesting About Reinforcement Learning?

Andrew G. Barto (University of Massachusetts Amherst, Turing Laureate) 

Abstract: Reinforcement Learning (RL) is the common sense idea that learning is about adjusting behavior in order to gain rewards and avoid penalties.  This is a key principle of animal learning that has been around, in various forms, at least since Edward Thorndike proposed the “Law of Effect” in 1898. Why is this interesting now, and why is RL playing so many roles in today’s AI systems? This talk aims to answer these questions in four dimensions.  First is history. RL was the basis of AI long before the term AI was introduced in 1956. The first machine learning (ML) systems were based on RL before digital computers existed. The second is the clarification of some misunderstandings that have been prevalent in the ML community.  For example, underappreciated was the distinction between error-correction, the basis of supervised learning, and more general optimization, the basis of RL. The third is that new or rediscovered algorithms and connections to well developed mathematical and engineering methods have been worked out. For example, temporal difference algorithms help address a classic credit-assignment problem; connections to stochastic optimal control have helped put RL on a more rigorous basis, with stochastic dynamic programming serving as a key link to RL algorithms. The fourth is RL's influence on understanding animal reward systems, in particular, on the role that dopamine plays in motivation and learning. 

Biography: Andrew G. Barto is Professor Emeritus of Computer Science, University of Massachusetts Amherst, having retired in 2012. He served as Chair of the UMass Department of Computer Science from 2007 to 2011. He received a B.S. with distinction in mathematics in 1970, and a Ph.D. in Computer Science in 1975, both from the University of Michigan.  Before retiring he co-directed the Autonomous Learning Laboratory at UMass Amherst, which produced many notable machine learning researchers.  Professor Barto received the 2004 IEEE Neural Network Society Pioneer Award, the IJCAI-17 Award for Research Excellence, a University of Massachusetts Neurosciences Lifetime Achievement Award in 20019, and the 2024 ACM A. M. Turing Award for developing the conceptual and algorithmic foundations of reinforcement learning. He has published over one hundred papers or chapters in journals, books, and conference and workshop proceedings. He is co-author with Richard Sutton of the book "Reinforcement Learning: An Introduction," MIT Press, 1998. A much expanded second edition was published in 2018. 

Tickets $5.00 Non-Refundable 
We are not doing any refunds. 
Password: Nie2025Barto