Metraipolis › archive › dc_d554fddd9f14
Canopy: Exploiting Piecewise Smooth Tree Priors for Multi-Fidelity Bandits — arXiv:2609.30017
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.30017
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T04:27:58.435Z
- SHA-256 of the archived text
bc6c7c7a9bd1976df959d475cba1918417c4fecf1e8e3367cadee9015dc72017- Stable URL of this page
https://metraipolis.com/documento/dc_d554fddd9f14
How to cite
Canopy: Exploiting Piecewise Smooth Tree Priors for Multi-Fidelity Bandits — arXiv:2609.30017 — arXiv (en), https://arxiv.org/abs/2609.30017 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d554fddd9f14 (Metraipolis archive dc_d554fddd9f14, sha256 bc6c7c7a9bd1976df959d475cba1918417c4fecf1e8e3367cadee9015dc72017)
BibTeX
@misc{dcd554fddd9f14,
title = {Canopy: Exploiting Piecewise Smooth Tree Priors for Multi-Fidelity Bandits — arXiv:2609.30017},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.30017},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d554fddd9f14; sha256 bc6c7c7a9bd1976df959d475cba1918417c4fecf1e8e3367cadee9015dc72017},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.30017. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Many LLM inference problems, including model routing, prefix-cache management, prompt trimming, and test-time search, can be viewed as optimization over a tree. This structure arises naturally from autoregressive generation: every prefix defines a node, and its continuations form a subtree below it. Internal nodes of the tree provide cheap but biased estimates of a region's value, while leaf evaluations are expensive but accurate. Hierarchical bandit methods can exploit this structure, but typically require a specific smoothness schedule to be specified in advance, even though real objectives are often only piecewise smooth and their optima may lie near sharp boundaries. We introduce CANOPY, a multi-fidelity tree bandit that learns where the smoothness prior is valid rather than assuming it globally. CANOPY uses cheap random-path probes to construct an online certificate of local aggregation bias, then directs expensive leaf evaluations toward cells where the certificate detects a smoothness violation. We prove fixed-budget and regret guarantees whose additional cost is additive in the number of discontinuities, recovering the smooth-tree rate when no violations are present and approaching structure-blind search as violations become dense. Across routing, top-$k$ identification, test-time search, caching, and prompt trimming, CANOPY consistently improves matched-budget performance, including $2.9\times$ higher top-10 recall on a 1000-model pool, $1.6\times$ more SWE-bench Verified issues resolved than best-of-$N$, and $3.6\times$ lower median time-to-first-token with prefix caching. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-25 in cs.AI). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.30017 — 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 2026-10-08T03:50:58.219Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).