Metraipolis › archive › dc_79758fb73ab7

From Local Skills to Long-Horizon Web Tasks: Progressive Skill Exploration for LLM Web Agents — arXiv:2611.03116

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.03116
License (as archived)
CC0-1.0
Archived
2026-11-03T10:54:29.269Z
SHA-256 of the archived text
5965a43bba6b6e019d1f9dbc666facfcc514a3143b2bc31e36ac1c0a8890974c
Stable URL of this page
https://metraipolis.com/documento/dc_79758fb73ab7

How to cite

From Local Skills to Long-Horizon Web Tasks: Progressive Skill Exploration for LLM Web Agents — arXiv:2611.03116 — arXiv (en), https://arxiv.org/abs/2611.03116 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_79758fb73ab7 (Metraipolis archive dc_79758fb73ab7, sha256 5965a43bba6b6e019d1f9dbc666facfcc514a3143b2bc31e36ac1c0a8890974c)
BibTeX
@misc{dc79758fb73ab7,
  title = {From Local Skills to Long-Horizon Web Tasks: Progressive Skill Exploration for LLM Web Agents — arXiv:2611.03116},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.03116},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_79758fb73ab7; sha256 5965a43bba6b6e019d1f9dbc666facfcc514a3143b2bc31e36ac1c0a8890974c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.03116. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language models (LLMs) are driving autonomous agents toward real-world environments, with the Web emerging as one of the most representative interaction scenarios. However, training Web agents with naive exploration remains challenging: successful trajectories are rare, and the sparse terminal rewards make it difficult for agents to obtain useful feedback from failed trajectories. Inspired by skill acquisition theory, we propose Skill-Composed Progressive Exploration (SCOPE), a short-to-long learning framework for long-horizon Web agents that decomposes complex task exploration into a gradual expansion process from local skills to complete task-solving capability. Specifically, SCOPE iteratively explores around local skills: Skill Discovery extracts reusable local skills and failure experiences from historical rollouts; Skill Orchestration composes discovered skills into high-level skill chains; and Skill Expansion leverages failure experiences aligned with the current skill chain to guide the agent toward exploring new subskills. Through this process, SCOPE progressively constructs task-completing skill chains across rollouts and internalizes them into the policy parameters through agentic on-policy distillation. Extensive experiments on WebArena, WebVoyager, and WebShop show that SCOPE consistently improves exploration dynamics and achieves 3.9% to 18.6% relative gains over the strongest baselines in each setting.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2611.03116 — 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 2026-11-03T10:49:38.411Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).