Optimistic Reinforcement Learning by Forward Kullback-Leibler Divergence Optimization

Abstract

This paper addresses a new interpretation of the traditional optimization method in reinforcement learning (RL) as optimization problems using reverse Kullback-Leibler (KL) divergence, and derives a new optimization method using forward KL divergence, instead of reverse KL divergence in the optimization problems. Although RL originally aims to maximize return indirectly through optimization of policy, the recent work by Levine has proposed a different derivation process with explicit consideration of optimality as stochastic variable. This paper follows this concept and formulates the traditional learning laws for both value function and policy as the optimization problems with reverse KL divergence including optimality. Focusing on the asymmetry of KL divergence, the new optimization problems with forward KL divergence are derived. Remarkably, such new optimization problems can be regarded as optimistic RL. That optimism is intuitively specified by a hyperparameter converted from an uncertainty parameter. In addition, it can be enhanced when it is integrated with prioritized experience replay and eligibility traces, both of which accelerate learning. The effects of this expected optimism was investigated through learning tendencies on numerical simulations using Pybullet. As a result, moderate optimism accelerated learning and yielded higher rewards. In a realistic robotic simulation, the proposed method with the moderate optimism outperformed one of the state-of-the-art RL method.

14 pages, 6 figures

References in corpus (5)

• Model-Augmented Actor-Critic: Backpropagating through Paths
• Domain-Independent Optimistic Initialization for Reinforcement Learning
• Effective Multi-step Temporal-Difference Learning for Non-Linear Function Approximation
• Optimistic Exploration even with a Pessimistic Initialisation
• Uncertainty-aware Model-based Policy Optimization

Cited by in corpus (6)

• Reward Bonuses with Gain Scheduling Inspired by Iterative Deepening Search
• Optimization Algorithm for Feedback and Feedforward Policies towards Robot Control Robust to Sensing Failures
• Consolidated Adaptive T-soft Update for Deep Reinforcement Learning
• Revisiting Experience Replayable Conditions
• A unified framework for divergences, free energies, and Fokker-Planck equations
• Weber-Fechner Law in Temporal Difference learning derived from Control as Inference