Reinforcement Learning Architects Honored with 2024 Turing Award
Andrew Barto, professor emeritus at the University of Massachusetts Amherst, and Richard Sutton, professor of computer science at the University of Alberta, have been awarded the 2024 Association for Computing Machinery A.M. Turing Award for “developing the conceptual and algorithmic foundations of reinforcement learning.” The accolade, often referred to as the “Nobel Prize of computing,” recognizes their decades-long collaboration and transformative impact on artificial intelligence.
Their work in reinforcement learning (RL) has defined how computational systems can learn to act based on evaluative feedback, creating algorithms that adapt through experience rather than explicit instruction. RL’s reach spans computer science, engineering, mathematics, neuroscience, psychology, and economics, influencing disciplines from optimal control theory to behavioral conditioning models. The methods they developed are now embedded in technologies that shape daily life and advanced research.
Barto’s contributions were sustained by long-term support from the U.S. National Science Foundation through programs such as the National Robotics Initiative, Robust Intelligence, and Collaborative Research in Computational Neuroscience. Greg Hager, NSF assistant director for Computer and Information Science and Engineering, stated, “Barto’s research exemplifies the power of foundational computational research that has not only advanced state-of-the-art decision-making machines and intelligent systems but has also provided critical insights into understanding intelligence itself.” Michael Littman, director for the NSF Division of Information and Intelligent Systems, added, “Andy Barto’s work laid the foundation for modern reinforcement learning, influencing generations of researchers, including myself. His insights with Rich Sutton into how agents can learn and adapt in complex environments form the backbone of how automated behavior is generated in the field of artificial intelligence. Without his pioneering research, many of today’s — and tomorrow’s — AI breakthroughs wouldn’t be possible.”
The partnership between Barto and Sutton began when Sutton was Barto’s first doctoral student. Their collaboration continued through Sutton’s tenure as a senior research scientist at UMass Amherst from 1995 to 1998, producing foundational RL approaches still in use today. These methods underpin diverse applications: conversational AI agents like ChatGPT use reinforcement learning from human feedback to refine responses; RL-driven game-playing systems have achieved superhuman performance in domains from Jeopardy to Go; robots acquire complex motor skills through trial-and-error training; RL optimizes microprocessor layouts and circuit designs; streaming platforms tailor recommendations using RL models; autonomous vehicles navigate traffic through learned decision-making; supply chains are fine-tuned for efficiency; and researchers employ RL to design novel algorithms.
The commercial and research impact is substantial. RL is a core technology for companies such as DeepMind and OpenAI, and many major technology firms maintain dedicated RL research groups. Its importance has also been recognized in education, with RL added to the Computer Science Standards of Learning for Virginia Public Schools in early 2025.
Beyond engineering and computation, Barto and Sutton’s work has bridged AI and neuroscience. In 1981, they demonstrated that temporal difference (TD) learning could account for learning behaviors unexplained by the Rescorla-Wagner model, reshaping theoretical frameworks in psychology. A 1995 study linked the TD algorithm to the activity of dopamine neurons, revealing a biological parallel to computational reward prediction. Subsequent experiments confirmed that TD learning accurately models how dopamine influences reward-based learning, deepening the understanding of both machine and human cognition.
The recognition of Barto and Sutton with the 2024 Turing Award underscores the enduring value of sustained federal investment in fundamental research. Their contributions have not only advanced the technical capabilities of reinforcement learning systems but also illuminated the mechanisms of learning itself, influencing fields from robotics to neuroscience and shaping the trajectory of intelligent systems development.
