Robot Policy Learning with Temporal Optimal Transport Reward

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Fu, Yuwei, Zhang, Haichao, Wu, Di, Xu, Wei, Boulet, Benoit
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866929573747228672
author Fu, Yuwei
Zhang, Haichao
Wu, Di
Xu, Wei
Boulet, Benoit
author_facet Fu, Yuwei
Zhang, Haichao
Wu, Di
Xu, Wei
Boulet, Benoit
contents Reward specification is one of the most tricky problems in Reinforcement Learning, which usually requires tedious hand engineering in practice. One promising approach to tackle this challenge is to adopt existing expert video demonstrations for policy learning. Some recent work investigates how to learn robot policies from only a single/few expert video demonstrations. For example, reward labeling via Optimal Transport (OT) has been shown to be an effective strategy to generate a proxy reward by measuring the alignment between the robot trajectory and the expert demonstrations. However, previous work mostly overlooks that the OT reward is invariant to temporal order information, which could bring extra noise to the reward signal. To address this issue, in this paper, we introduce the Temporal Optimal Transport (TemporalOT) reward to incorporate temporal order information for learning a more accurate OT-based proxy reward. Extensive experiments on the Meta-world benchmark tasks validate the efficacy of the proposed method. Code is available at: https://github.com/fuyw/TemporalOT
format Preprint
id arxiv_https___arxiv_org_abs_2410_21795
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Robot Policy Learning with Temporal Optimal Transport Reward
Fu, Yuwei
Zhang, Haichao
Wu, Di
Xu, Wei
Boulet, Benoit
Artificial Intelligence
Machine Learning
Robotics
Reward specification is one of the most tricky problems in Reinforcement Learning, which usually requires tedious hand engineering in practice. One promising approach to tackle this challenge is to adopt existing expert video demonstrations for policy learning. Some recent work investigates how to learn robot policies from only a single/few expert video demonstrations. For example, reward labeling via Optimal Transport (OT) has been shown to be an effective strategy to generate a proxy reward by measuring the alignment between the robot trajectory and the expert demonstrations. However, previous work mostly overlooks that the OT reward is invariant to temporal order information, which could bring extra noise to the reward signal. To address this issue, in this paper, we introduce the Temporal Optimal Transport (TemporalOT) reward to incorporate temporal order information for learning a more accurate OT-based proxy reward. Extensive experiments on the Meta-world benchmark tasks validate the efficacy of the proposed method. Code is available at: https://github.com/fuyw/TemporalOT
title Robot Policy Learning with Temporal Optimal Transport Reward
topic Artificial Intelligence
Machine Learning
Robotics
url https://arxiv.org/abs/2410.21795