Reliable Critics: Monotonic Improvement and Convergence Guarantees for Reinforcement Learning
By: Eshwar S. R. , Gugan Thoppe , Aditya Gopalan and more
Potential Business Impact:
Makes learning robots more reliable and predictable.
Despite decades of research, it remains challenging to correctly use Reinforcement Learning (RL) algorithms with function approximation. A prime example is policy iteration, whose fundamental guarantee of monotonic improvement collapses even under linear function approximation. To address this issue, we introduce Reliable Policy Iteration (RPI). It replaces the common projection or Bellman-error minimization during policy evaluation with a Bellman-based constrained optimization. We prove that not only does RPI confer textbook monotonicity on its value estimates but these estimates also lower bound the true return. Also, their limit partially satisfies the unprojected Bellman equation, emphasizing RPI's natural fit within RL. RPI is the first algorithm with such monotonicity and convergence guarantees under function approximation. For practical use, we provide a model-free variant of RPI that amounts to a novel critic. It can be readily integrated into primary model-free PI implementations such as DQN and DDPG. In classical control tasks, such RPI-enhanced variants consistently maintain their lower-bound guarantee while matching or surpassing the performance of all baseline methods.
Similar Papers
Reliable Policy Iteration: Performance Robustness Across Architecture and Environment Perturbations
Artificial Intelligence
Makes robots learn faster and more reliably.
Implicit Constraint-Aware Off-Policy Correction for Offline Reinforcement Learning
Systems and Control
Teaches computers to follow rules for better learning.
Adaptive Resolving Methods for Reinforcement Learning with Function Approximations
Machine Learning (CS)
Teaches computers to learn from experience faster.