Saved in:
Bibliographic Details
Main Authors: Xu, Feng, Zhai, Guangyao, Kong, Xin, Fu, Tingzhong, Gordon, Daniel F. N., An, Xueli, Busam, Benjamin
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.05107
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917167281209344
author Xu, Feng
Zhai, Guangyao
Kong, Xin
Fu, Tingzhong
Gordon, Daniel F. N.
An, Xueli
Busam, Benjamin
author_facet Xu, Feng
Zhai, Guangyao
Kong, Xin
Fu, Tingzhong
Gordon, Daniel F. N.
An, Xueli
Busam, Benjamin
contents Recent advances in Vision-Language-Action (VLA) models, powered by large language models and reinforcement learning-based fine-tuning, have shown remarkable progress in robotic manipulation. Existing methods often treat long-horizon actions as linguistic sequences and apply trajectory-level optimization methods such as Trajectory-wise Preference Optimization (TPO) or Proximal Policy Optimization (PPO), leading to coarse credit assignment and unstable training. However, unlike language, where a unified semantic meaning is preserved despite flexible sentence order, action trajectories progress through causally chained stages with different learning difficulties. This motivates progressive stage optimization. Thereby, we present Stage-Aware Reinforcement (STARE), a module that decomposes a long-horizon action trajectory into semantically meaningful stages and provides dense, interpretable, and stage-aligned reinforcement signals. Integrating STARE into TPO and PPO, we yield Stage-Aware TPO (STA-TPO) and Stage-Aware PPO (STA-PPO) for offline stage-wise preference and online intra-stage interaction, respectively. Further building on supervised fine-tuning as initialization, we propose the Imitation -> Preference -> Interaction (IPI), a serial fine-tuning pipeline for improving action accuracy in VLA models. Experiments on SimplerEnv and ManiSkill3 demonstrate substantial gains, achieving state-of-the-art success rates of 98.0 percent on SimplerEnv and 96.4 percent on ManiSkill3 tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2512_05107
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle STARE-VLA: Progressive Stage-Aware Reinforcement for Fine-Tuning Vision-Language-Action Models
Xu, Feng
Zhai, Guangyao
Kong, Xin
Fu, Tingzhong
Gordon, Daniel F. N.
An, Xueli
Busam, Benjamin
Robotics
Recent advances in Vision-Language-Action (VLA) models, powered by large language models and reinforcement learning-based fine-tuning, have shown remarkable progress in robotic manipulation. Existing methods often treat long-horizon actions as linguistic sequences and apply trajectory-level optimization methods such as Trajectory-wise Preference Optimization (TPO) or Proximal Policy Optimization (PPO), leading to coarse credit assignment and unstable training. However, unlike language, where a unified semantic meaning is preserved despite flexible sentence order, action trajectories progress through causally chained stages with different learning difficulties. This motivates progressive stage optimization. Thereby, we present Stage-Aware Reinforcement (STARE), a module that decomposes a long-horizon action trajectory into semantically meaningful stages and provides dense, interpretable, and stage-aligned reinforcement signals. Integrating STARE into TPO and PPO, we yield Stage-Aware TPO (STA-TPO) and Stage-Aware PPO (STA-PPO) for offline stage-wise preference and online intra-stage interaction, respectively. Further building on supervised fine-tuning as initialization, we propose the Imitation -> Preference -> Interaction (IPI), a serial fine-tuning pipeline for improving action accuracy in VLA models. Experiments on SimplerEnv and ManiSkill3 demonstrate substantial gains, achieving state-of-the-art success rates of 98.0 percent on SimplerEnv and 96.4 percent on ManiSkill3 tasks.
title STARE-VLA: Progressive Stage-Aware Reinforcement for Fine-Tuning Vision-Language-Action Models
topic Robotics
url https://arxiv.org/abs/2512.05107