Policy Similarity Metric

2000년 도입 · 논문 2편에서 사용

Policy Similarity Metric, or PSM, is a similarity metric for measuring behavioral similarity between states in reinforcement learning. It assigns high similarity to states for which the optimal policies in those states as well as in future states are similar. PSM is reward-agnostic, making it more robust for generalization compared to approaches that rely on reward information.

출처: Contrastive Behavioral Similarity Embeddings for Generalization in Reinforcement Learning

소개 논문: Contrastive Behavioral Similarity Embeddings for Generalization in Reinforcement Learning