Metraipolis › archive › dc_242aedf91155
ResMerge: Residual Policy Learning and Merging for Continual Adaptation of Pre-trained Robot Policies — arXiv:2610.41663
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.41663
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T05:47:26.475Z
- SHA-256 of the archived text
3d8345042a2dff2728705fd76ea8703fe8a4f4ca7647f2eeea22e0e6c54b9370- Stable URL of this page
https://metraipolis.com/documento/dc_242aedf91155
How to cite
ResMerge: Residual Policy Learning and Merging for Continual Adaptation of Pre-trained Robot Policies — arXiv:2610.41663 — arXiv (en), https://arxiv.org/abs/2610.41663 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_242aedf91155 (Metraipolis archive dc_242aedf91155, sha256 3d8345042a2dff2728705fd76ea8703fe8a4f4ca7647f2eeea22e0e6c54b9370)
BibTeX
@misc{dc242aedf91155,
title = {ResMerge: Residual Policy Learning and Merging for Continual Adaptation of Pre-trained Robot Policies — arXiv:2610.41663},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.41663},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_242aedf91155; sha256 3d8345042a2dff2728705fd76ea8703fe8a4f4ca7647f2eeea22e0e6c54b9370},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41663. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Continual adaptation to new tasks is essential for robots deployed in the real world, but adapting pre-trained policies remains challenging due to parameter growth and the need for continual demonstration collection. Parameter-Efficient Continual Fine-Tuning (PECFT) isolates task-specific updates, but adapter-per-task designs require demonstrations and grow with the task set. We propose ResMerge, a two-stage framework that learns mergeable residual policies atop a frozen pre-trained policy: residual reinforcement learning (RL) acquires task-specific corrections without additional expert demonstrations during adaptation, and parameter-space merging consolidates compatible residuals to reduce stored parameters. We introduce Residual-Field Consistency (RFC), a simple, bounded, and interpretable metric that assesses compatibility in residual action-field space by comparing correction direction and magnitude. Experiments on LIBERO across two backbones and on a real-world Franka robot show improved continual learning performance over strong PECFT baselines, while selective merging reduces the residual-parameter footprint. Project website: this https URL . Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.RO, nuovi lavori del giorno). Categorie arXiv: Robotics (cs.RO). Pagina del lavoro: https://arxiv.org/abs/2610.41663 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-11-03T05:33:48.023Z dalla pagina https://arxiv.org/list/cs.RO/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).