Refinery: Active Fine-tuning and Deployment-time Optimization for Contact-Rich Policies

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Tang, Bingjie, Akinola, Iretiayo, Xu, Jie, Wen, Bowen, Fox, Dieter, Sukhatme, Gaurav S., Ramos, Fabio, Gupta, Abhishek, Narang, Yashraj
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911206866944000
author Tang, Bingjie
Akinola, Iretiayo
Xu, Jie
Wen, Bowen
Fox, Dieter
Sukhatme, Gaurav S.
Ramos, Fabio
Gupta, Abhishek
Narang, Yashraj
author_facet Tang, Bingjie
Akinola, Iretiayo
Xu, Jie
Wen, Bowen
Fox, Dieter
Sukhatme, Gaurav S.
Ramos, Fabio
Gupta, Abhishek
Narang, Yashraj
contents Simulation-based learning has enabled policies for precise, contact-rich tasks (e.g., robotic assembly) to reach high success rates (~80%) under high levels of observation noise and control error. Although such performance may be sufficient for research applications, it falls short of industry standards and makes policy chaining exceptionally brittle. A key limitation is the high variance in individual policy performance across diverse initial conditions. We introduce Refinery, an effective framework that bridges this performance gap, robustifying policy performance across initial conditions. We propose Bayesian Optimization-guided fine-tuning to improve individual policies, and Gaussian Mixture Model-based sampling during deployment to select initializations that maximize execution success. Using Refinery, we improve mean success rates by 10.98% over state-of-the-art methods in simulation-based learning for robotic assembly, reaching 91.51% in simulation and comparable performance in the real world. Furthermore, we demonstrate that these fine-tuned policies can be chained to accomplish long-horizon, multi-part assembly$\unicode{x2013}$successfully assembling up to 8 parts without requiring explicit multi-step training.
format Preprint
id arxiv_https___arxiv_org_abs_2510_11019
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Refinery: Active Fine-tuning and Deployment-time Optimization for Contact-Rich Policies
Tang, Bingjie
Akinola, Iretiayo
Xu, Jie
Wen, Bowen
Fox, Dieter
Sukhatme, Gaurav S.
Ramos, Fabio
Gupta, Abhishek
Narang, Yashraj
Robotics
Simulation-based learning has enabled policies for precise, contact-rich tasks (e.g., robotic assembly) to reach high success rates (~80%) under high levels of observation noise and control error. Although such performance may be sufficient for research applications, it falls short of industry standards and makes policy chaining exceptionally brittle. A key limitation is the high variance in individual policy performance across diverse initial conditions. We introduce Refinery, an effective framework that bridges this performance gap, robustifying policy performance across initial conditions. We propose Bayesian Optimization-guided fine-tuning to improve individual policies, and Gaussian Mixture Model-based sampling during deployment to select initializations that maximize execution success. Using Refinery, we improve mean success rates by 10.98% over state-of-the-art methods in simulation-based learning for robotic assembly, reaching 91.51% in simulation and comparable performance in the real world. Furthermore, we demonstrate that these fine-tuned policies can be chained to accomplish long-horizon, multi-part assembly$\unicode{x2013}$successfully assembling up to 8 parts without requiring explicit multi-step training.
title Refinery: Active Fine-tuning and Deployment-time Optimization for Contact-Rich Policies
topic Robotics
url https://arxiv.org/abs/2510.11019