Metraipolis › archive › dc_8b920ff297ef

MDSKILL: Learning to Internalize Skill Memory through Multi-View Distillation for Efficient Dynamic Agent Self-Evolution — arXiv:2611.05313

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.05313
License (as archived)
CC0-1.0
Archived
2026-11-04T21:43:41.490Z
SHA-256 of the archived text
fb8f8cff1d2179c9ccc72a2a0b6080b83b80d1fbe4450f4f49d07ba370a0ceb5
Stable URL of this page
https://metraipolis.com/documento/dc_8b920ff297ef

How to cite

MDSKILL: Learning to Internalize Skill Memory through Multi-View Distillation for Efficient Dynamic Agent Self-Evolution — arXiv:2611.05313 — arXiv (en), https://arxiv.org/abs/2611.05313 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_8b920ff297ef (Metraipolis archive dc_8b920ff297ef, sha256 fb8f8cff1d2179c9ccc72a2a0b6080b83b80d1fbe4450f4f49d07ba370a0ceb5)
BibTeX
@misc{dc8b920ff297ef,
  title = {MDSKILL: Learning to Internalize Skill Memory through Multi-View Distillation for Efficient Dynamic Agent Self-Evolution — arXiv:2611.05313},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.05313},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_8b920ff297ef; sha256 fb8f8cff1d2179c9ccc72a2a0b6080b83b80d1fbe4450f4f49d07ba370a0ceb5},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.05313. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Skill memories encode past experiences as textual prompts to improve reinforcement learning (RL) for self-evolving agents. Complementing RL, On-Policy Self-Distillation (OPSD) provides dense, fine-grained supervision to internalize knowledge from skill-guided trajectories and accelerate policy learning. However, self-generated memories often contain noisy supervision and functionally entangled, redundant, or conflicting knowledge, making direct distillation difficult. Under limited training budgets, these issues can hinder knowledge internalization and even degrade performance. To address this problem, we propose MDSKILL, a multi-view distillation approach for skill memory internalization. MDSKILL organizes skill memories into three complementary functional views (state grounding, action planning, and execution control) and selects high-quality skills using dynamic metrics to construct view-specific and integrated reasoning branches. OPSD then transfers and consolidates complementary experiential knowledge across these branches into a shared policy, enabling efficient skill internalization and continuous agent self-evolution. We evaluate MDSKILL on three agent benchmarks against RL, skill-augmented, and self-distillation baselines that do not rely on stronger external teacher models. Without additional computational overhead, MDSKILL improves upon SOTA performance by up to 7.1% and 8.1% on ALFWorld and WebShop. Extensive ablations and evolution analyses further validate its effectiveness in knowledge internalization and policy improvement. The code is available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2611.05313 — 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 2026-11-04T21:36:04.444Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).