Loading timeline…
19861980s
Ronald J. Williams
Third author of the 1986 backpropagation paper and inventor of REINFORCE, the basis of policy-gradient reinforcement learning.
Organizations
Northeastern UniversityUniversity of California, San Diego
Major Achievements
- •Co-authored the 1986 Nature paper on backpropagation with David Rumelhart and Geoffrey Hinton that reignited neural network research.
- •Introduced REINFORCE (1992), the policy-gradient method underpinning modern reinforcement learning and RLHF.
- •Co-developed real-time recurrent learning for training recurrent networks on streaming sequences.
Key Papers
- Learning Representations by Back-Propagating Errors
David E. Rumelhart, Geoffrey E. Hinton, Ronald J. Williams · Nature · 1986
- Simple Statistical Gradient-Following Algorithms for Connectionist Reinforcement Learning
Ronald J. Williams · Machine Learning · 1992