Metraipolis › archive › dc_0bc62222d67d
Distillation for Efficient Multitask Manipulation Policies via Conditional Flow Matching — arXiv:2609.28107
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.28107
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T09:13:20.995Z
- SHA-256 of the archived text
e6ce29c73c68a8200c98075d7f8aedd1a4fcf2fe455663c780bf90af4dd5b9f5- Stable URL of this page
https://metraipolis.com/documento/dc_0bc62222d67d
How to cite
Distillation for Efficient Multitask Manipulation Policies via Conditional Flow Matching — arXiv:2609.28107 — arXiv (en), https://arxiv.org/abs/2609.28107 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0bc62222d67d (Metraipolis archive dc_0bc62222d67d, sha256 e6ce29c73c68a8200c98075d7f8aedd1a4fcf2fe455663c780bf90af4dd5b9f5)
BibTeX
@misc{dc0bc62222d67d,
title = {Distillation for Efficient Multitask Manipulation Policies via Conditional Flow Matching — arXiv:2609.28107},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.28107},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0bc62222d67d; sha256 e6ce29c73c68a8200c98075d7f8aedd1a4fcf2fe455663c780bf90af4dd5b9f5},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.28107. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Advances in generative modeling have recently been extensively employed in robotics for policy learning. In particular, Conditional Flow Matching (CFM) trained with expert demonstrations has been shown to outperform existing methods on robot manipulation benchmarks. While prior work has mainly focused on single-task settings, we study the problem from a multi-task perspective, as training independent models for each task is computationally expensive. Multi-Task policy learning comes with its own set of challenges, as naively training on a concatenated dataset of demonstrations would either require increased model capacity to accommodate the added complexity or result in drops in performance. We propose to distill knowledge from single-task CFM experts into a shared multi-task policy by transferring their learned velocity fields. We combine this distillation signal with the original CFM objective to retain fidelity to the demonstrations. Experiments on RLBench show that our approach improves multi-task policy performance over naive training while maintaining a fixed model size. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-24 in cs.AI). Categorie arXiv: Robotics (cs.RO) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2609.28107 — 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 2026-10-08T08:13:26.750Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-24?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).