Transfer Learning via Temporal Contrastive Learning

Abstract

This thesis introduces a novel transfer learning framework for deep reinforcement learning. The approach automatically combines goalconditioned policies with temporal contrastive learning to discover meaningful sub-goals. The approach involves pre-training a goal-conditioned agent, finetuning it on the target domain, and using contrastive learning to construct a planning graph that guides the agent via sub-goals. Experiments on PointMaze and multi-agent coordination Overcooked tasks demonstrate improved sample efficiency, the ability to solve sparse-reward and long-horizon problems, and enhanced interpretability compared to baselines. The results highlight the effectiveness of integrating goalconditioned policies with unsupervised temporal abstraction learning for complex multi-agent transfer learning. Compared to state-of-the-art baselines, our method achieves the same or better performances while requiring only 23.4% of the training samples.