Andy Barto

Papers

Total Citations

H-Index

About

Andy Barto’s research lies at the intersection of reinforcement learning and intelligent control, where he has pioneered methods that blend supervised learning with actor-critic architectures. His most influential work, the 2012 paper “Supervised Actor-Critic Reinforcement Learning” (69 citations), introduced a novel framework that leverages a supervisor’s guidance to improve policy gradient learning, moving beyond purely stochastic search to achieve more efficient and stable policy updates. This contribution directly addresses a key limitation in reinforcement learning: the need for structure and domain knowledge. Earlier, in 2002, he laid the groundwork with “Supervised Learning Combined with an Actor-Critic Architecture” (19 citations), which formalized how to integrate external supervision into the actor-critic paradigm, enabling agents to learn faster and more reliably in complex environments. Barto’s work has been instrumental in shaping modern reinforcement learning, particularly in robotics and autonomous systems, where sample efficiency and safe exploration are critical. His research continues to inspire new generations of algorithms that bridge supervised and reinforcement learning, making him a respected figure in the AI community.