Metraipolis › archive › dc_e5f502d06ec9
Self Improvement via Fast Tree-search — arXiv:2609.19526
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.19526
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-10T03:20:02.441Z
- SHA-256 of the archived text
04a5dfd61c676471106f1c69357b5911355ae5d96923d811a115dc468844dddf- Stable URL of this page
https://metraipolis.com/documento/dc_e5f502d06ec9
How to cite
Self Improvement via Fast Tree-search — arXiv:2609.19526 — arXiv (en), https://arxiv.org/abs/2609.19526 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e5f502d06ec9 (Metraipolis archive dc_e5f502d06ec9, sha256 04a5dfd61c676471106f1c69357b5911355ae5d96923d811a115dc468844dddf)
BibTeX
@misc{dce5f502d06ec9,
title = {Self Improvement via Fast Tree-search — arXiv:2609.19526},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.19526},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e5f502d06ec9; sha256 04a5dfd61c676471106f1c69357b5911355ae5d96923d811a115dc468844dddf},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.19526. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Coding agents can recursively modify their own implementations, forming a loop of self-improvement. While prior work shows this can boost performance on coding benchmarks, existing approaches are costly and compute-intensive. We introduce a simple, sample-efficient self-improvement framework that significantly improves coding performance under strict budget constraints. We identify evaluation of candidate self-modifications as the main runtime bottleneck since prior approaches estimate their effectiveness by re-running a subset of benchmark tasks with the modified agent, which is time-consuming. We introduce Recursive Self Improvement via Fast Tree-search (SIFT), which augments these downstream task evaluations with an LLM-as-a-judge signal that performs pairwise comparisons between candidate patches, where the win-loss record is aggregated with a regularized Bradley-Terry model, and the resulting strength scores drive rank-based parent sampling inside a lightweight disaggregated tree search. Expensive downstream task evaluations are reserved only for the most promising nodes. Using a fully disaggregated tree search pipeline, the judge scores provide intermediate signal to guide exploration on promising candidate patches without being bottlenecked by slow evaluation runs. SIFT outperforms existing tree-search based self-evolution frameworks on the full Polyglot benchmark with significantly lower resource requirements in terms of CPU hours, wall clock time, and API cost. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-18 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2609.19526 — 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 2026-10-10T03:04:13.613Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).