PRIMT: Preference-based Reinforcement Learning with Multimodal Feedback and Trajectory Synthesis from Foundation Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Ruiqi, Zhao, Dezhong, Yuan, Ziqin, Shao, Tianyu, Chen, Guohua, Kao, Dominic, Hong, Sungeun, Min, Byung-Cheol
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912739372302336
author Wang, Ruiqi
Zhao, Dezhong
Yuan, Ziqin
Shao, Tianyu
Chen, Guohua
Kao, Dominic
Hong, Sungeun
Min, Byung-Cheol
author_facet Wang, Ruiqi
Zhao, Dezhong
Yuan, Ziqin
Shao, Tianyu
Chen, Guohua
Kao, Dominic
Hong, Sungeun
Min, Byung-Cheol
contents Preference-based reinforcement learning (PbRL) has emerged as a promising paradigm for teaching robots complex behaviors without reward engineering. However, its effectiveness is often limited by two critical challenges: the reliance on extensive human input and the inherent difficulties in resolving query ambiguity and credit assignment during reward learning. In this paper, we introduce PRIMT, a PbRL framework designed to overcome these challenges by leveraging foundation models (FMs) for multimodal synthetic feedback and trajectory synthesis. Unlike prior approaches that rely on single-modality FM evaluations, PRIMT employs a hierarchical neuro-symbolic fusion strategy, integrating the complementary strengths of large language models and vision-language models in evaluating robot behaviors for more reliable and comprehensive feedback. PRIMT also incorporates foresight trajectory generation, which reduces early-stage query ambiguity by warm-starting the trajectory buffer with bootstrapped samples, and hindsight trajectory augmentation, which enables counterfactual reasoning with a causal auxiliary loss to improve credit assignment. We evaluate PRIMT on 2 locomotion and 6 manipulation tasks on various benchmarks, demonstrating superior performance over FM-based and scripted baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2509_15607
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle PRIMT: Preference-based Reinforcement Learning with Multimodal Feedback and Trajectory Synthesis from Foundation Models
Wang, Ruiqi
Zhao, Dezhong
Yuan, Ziqin
Shao, Tianyu
Chen, Guohua
Kao, Dominic
Hong, Sungeun
Min, Byung-Cheol
Robotics
Preference-based reinforcement learning (PbRL) has emerged as a promising paradigm for teaching robots complex behaviors without reward engineering. However, its effectiveness is often limited by two critical challenges: the reliance on extensive human input and the inherent difficulties in resolving query ambiguity and credit assignment during reward learning. In this paper, we introduce PRIMT, a PbRL framework designed to overcome these challenges by leveraging foundation models (FMs) for multimodal synthetic feedback and trajectory synthesis. Unlike prior approaches that rely on single-modality FM evaluations, PRIMT employs a hierarchical neuro-symbolic fusion strategy, integrating the complementary strengths of large language models and vision-language models in evaluating robot behaviors for more reliable and comprehensive feedback. PRIMT also incorporates foresight trajectory generation, which reduces early-stage query ambiguity by warm-starting the trajectory buffer with bootstrapped samples, and hindsight trajectory augmentation, which enables counterfactual reasoning with a causal auxiliary loss to improve credit assignment. We evaluate PRIMT on 2 locomotion and 6 manipulation tasks on various benchmarks, demonstrating superior performance over FM-based and scripted baselines.
title PRIMT: Preference-based Reinforcement Learning with Multimodal Feedback and Trajectory Synthesis from Foundation Models
topic Robotics
url https://arxiv.org/abs/2509.15607