Accelerating Multi-Task Temporal Difference Learning under Low-Rank Representation

2025/03/03 by Bai, Yitao, Zeng, Sihan, Romberg, Justin +1
#FOS: Computer and information sciences #FOS: Electrical engineering #Machine Learning (cs.LG) #Systems and Control (eess.SY) #electronic engineering #information engineering

paper · doi:10.48550/arxiv.2503.02030

Abstract

We study policy evaluation problems in multi-task reinforcement learning (RL) under a low-rank representation setting. In this setting, we are given N learning tasks where the corresponding value function of these tasks lie in an r-dimensional subspace, with r

Related

• A Tensor Low-Rank Approximation for Value Functions in Multi-Task Reinforcement Learning
• Provable Benefit of Multitask Representation Learning in Reinforcement Learning
• Learning Shared Representations in Multi-task Reinforcement Learning
• Offline Multitask Representation Learning for Reinforcement Learning
• Hierarchical Reinforcement Learning with Hindsight