Saved in:
Bibliographic Details
Main Authors: Deng, Shilong, Zheng, Zetao, He, Hongcai, Weng, Paul, Shao, Jie
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2501.07346
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912186177159168
author Deng, Shilong
Zheng, Zetao
He, Hongcai
Weng, Paul
Shao, Jie
author_facet Deng, Shilong
Zheng, Zetao
He, Hongcai
Weng, Paul
Shao, Jie
contents A major challenge in Reinforcement Learning (RL) is the difficulty of learning an optimal policy from sparse rewards. Prior works enhance online RL with conventional Imitation Learning (IL) via a handcrafted auxiliary objective, at the cost of restricting the RL policy to be sub-optimal when the offline data is generated by a non-expert policy. Instead, to better leverage valuable information in offline data, we develop Generalized Imitation Learning from Demonstration (GILD), which meta-learns an objective that distills knowledge from offline data and instills intrinsic motivation towards the optimal policy. Distinct from prior works that are exclusive to a specific RL algorithm, GILD is a flexible module intended for diverse vanilla off-policy RL algorithms. In addition, GILD introduces no domain-specific hyperparameter and minimal increase in computational cost. In four challenging MuJoCo tasks with sparse rewards, we show that three RL algorithms enhanced with GILD significantly outperform state-of-the-art methods.
format Preprint
id arxiv_https___arxiv_org_abs_2501_07346
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Enhancing Online Reinforcement Learning with Meta-Learned Objective from Offline Data
Deng, Shilong
Zheng, Zetao
He, Hongcai
Weng, Paul
Shao, Jie
Machine Learning
A major challenge in Reinforcement Learning (RL) is the difficulty of learning an optimal policy from sparse rewards. Prior works enhance online RL with conventional Imitation Learning (IL) via a handcrafted auxiliary objective, at the cost of restricting the RL policy to be sub-optimal when the offline data is generated by a non-expert policy. Instead, to better leverage valuable information in offline data, we develop Generalized Imitation Learning from Demonstration (GILD), which meta-learns an objective that distills knowledge from offline data and instills intrinsic motivation towards the optimal policy. Distinct from prior works that are exclusive to a specific RL algorithm, GILD is a flexible module intended for diverse vanilla off-policy RL algorithms. In addition, GILD introduces no domain-specific hyperparameter and minimal increase in computational cost. In four challenging MuJoCo tasks with sparse rewards, we show that three RL algorithms enhanced with GILD significantly outperform state-of-the-art methods.
title Enhancing Online Reinforcement Learning with Meta-Learned Objective from Offline Data
topic Machine Learning
url https://arxiv.org/abs/2501.07346