Using Contrastive Samples for Identifying and Leveraging Possible Causal Relationships in Reinforcement Learning

2022/10/28 by Harshad Khadilkar, Khadilkar, Harshad, Hardik Meisheri +1
Computer Science · #Advanced Multi-Objective Optimization Algorithms #Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Reinforcement Learning in Robotics #Software Engineering Research

paper · pdf · doi:10.48550/arxiv.2210.17296

openalex publication_date 2022/10/28 · openalex created_date 2022/11/06 · openalex updated_date 2026/07/28

Abstract

A significant challenge in reinforcement learning is quantifying the complex relationship between actions and long-term rewards. The effects may manifest themselves over a long sequence of state-action pairs, making them hard to pinpoint. In this paper, we propose a method to link transitions with significant deviations in state with unusually large variations in subsequent rewards. Such transitions are marked as possible causal effects, and the corresponding state-action pairs are added to a separate replay buffer. In addition, we include contrastive samples corresponding to transitions from a similar state but with differing actions. Including this Contrastive Experience Replay (CER) during training is shown to outperform standard value-based methods on 2D navigation tasks. We believe that CER can be useful for a broad class of learning tasks, including for any off-policy reinforcement learning algorithm.

Related

• CIER: A Novel Experience Replay Approach with Causal Inference in Deep Reinforcement Learning
• Toward Causal-Aware RL: State-Wise Action-Refined Temporal Difference
• Mixing Human Demonstrations with Self-Exploration in Experience Replay for Deep Reinforcement Learning
• Return-Based Contrastive Representation Learning for Reinforcement Learning
• CUER: Corrected Uniform Experience Replay for Off-Policy Continuous Deep Reinforcement Learning Algorithms