Metraipolis › archive › dc_cdc52106bb35

Same Trajectory, Different Knowledge: Information-State-Aware On-Policy Distillation for Search Agents — arXiv:2611.01987

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.01987
License (as archived): CC0-1.0
Archived: 2026-11-03T11:39:41.314Z
SHA-256 of the archived text: b615bdbcf4e77a08a587018bd0c3a26d5ca689840c3e3fe82f7f3dfe2a161d1e
Stable URL of this page: https://metraipolis.com/documento/dc_cdc52106bb35

How to cite

Same Trajectory, Different Knowledge: Information-State-Aware On-Policy Distillation for Search Agents — arXiv:2611.01987 — arXiv (en), https://arxiv.org/abs/2611.01987 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_cdc52106bb35 (Metraipolis archive dc_cdc52106bb35, sha256 b615bdbcf4e77a08a587018bd0c3a26d5ca689840c3e3fe82f7f3dfe2a161d1e)

BibTeX

@misc{dccdc52106bb35,
  title = {Same Trajectory, Different Knowledge: Information-State-Aware On-Policy Distillation for Search Agents — arXiv:2611.01987},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.01987},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_cdc52106bb35; sha256 b615bdbcf4e77a08a587018bd0c3a26d5ca689840c3e3fe82f7f3dfe2a161d1e},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.01987. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

On-policy distillation trains search agents with teacher feedback on student-generated trajectories. Recent methods augment the teacher's context with supporting evidence, reference answers, or hindsight information. However, this privileged information may allow the teacher to rely on facts that the student cannot reliably recover from its current state. The resulting information-state mismatch can lead the teacher to undervalue retrieval that helps the student establish those facts. To address this mismatch, we introduce Information-State-Aware Distillation (ISA-Distill), which adapts teacher feedback to the student's factual recoverability. ISA-Distill estimates how reliably the student can recover action-relevant knowledge and derives teacher feedback for the same sampled action from two complementary views. The fact-use view supplies the teacher with the relevant fact, producing feedback that guides the student in using knowledge. The evidence-seeking view instead presents the information need, producing feedback that guides the student in acquiring supporting evidence. The two signals are weighted toward fact use when recoverability is high and toward evidence seeking when it is low, and the resulting feedback guides policy optimization. Experiments across diverse QA benchmarks and model scales demonstrate improved answer accuracy and retrieval efficiency. Our code and model checkpoints will be open-sourced.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.IR, nuovi lavori del giorno).
Categorie arXiv: Information Retrieval (cs.IR).
Pagina del lavoro: https://arxiv.org/abs/2611.01987 — 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 2026-11-03T11:06:22.191Z dalla pagina https://arxiv.org/list/cs.IR/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).