Efficient Model-Based Reinforcement Learning Through Optimistic Thompson Sampling

Abstract

An efficient model-based reinforcement learning method using optimistic Thompson sampling to balance exploration and exploitation in complex sequential decision-making problems.

Paper