Metraipolis › archive › dc_610c5fdfdfcf
Fast A/B/n Testing: Exact Multi-Policy Comparison via Tree-Coupled Feedback Sharing — arXiv:2608.12831
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.12831
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T04:18:09.710Z
- SHA-256 of the archived text
9c8b4eaf576e49ad40f7c294fcda1ccf890d1a6c06fe5c6574c72143a830dae2- Stable URL of this page
https://metraipolis.com/documento/dc_610c5fdfdfcf
How to cite
Fast A/B/n Testing: Exact Multi-Policy Comparison via Tree-Coupled Feedback Sharing — arXiv:2608.12831 — arXiv (en), https://arxiv.org/abs/2608.12831 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_610c5fdfdfcf (Metraipolis archive dc_610c5fdfdfcf, sha256 9c8b4eaf576e49ad40f7c294fcda1ccf890d1a6c06fe5c6574c72143a830dae2)
BibTeX
@misc{dc610c5fdfdfcf,
title = {Fast A/B/n Testing: Exact Multi-Policy Comparison via Tree-Coupled Feedback Sharing — arXiv:2608.12831},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.12831},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_610c5fdfdfcf; sha256 9c8b4eaf576e49ad40f7c294fcda1ccf890d1a6c06fe5c6574c72143a830dae2},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.12831. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Online platforms increasingly compare many adaptive decision policies---ranking systems, recommendation algorithms, pricing rules, and language-model agents---while each reward-bearing interaction can be costly or risky. A direct A/B/n design gives each of $J\ge 2$ policies its own horizon-$T$ trajectory and therefore uses $JT$ outcomes. We introduce Tree-Coupled A/B Testing (\TCAB), an exact feedback-sharing design for arbitrary history-dependent contextual-bandit policies. At each round, a predictable tree connects the current policy histories; every parent--child context--action law is maximally coupled, and one reward is shared within each component of matched tree edges. Every policy retains exactly its standalone finite-horizon trajectory law, even though the policies are deliberately dependent. If $D_{e,t}$ records a mismatch on tree edge $e$ at round $t$, the number of reward queries satisfies the pathwise identity $N(T)=T+\sum_{t,e}D_{e,t}$ and hence equals $T$ plus cumulative tree-edge total variation in expectation. This cost is conditionally optimal among exact edge-local designs on the selected tree, and a current-round minimum-spanning tree is myopically optimal among tree designs. For fixed $J$, sublinear pseudo-regret of every policy and almost-sure uniqueness of the oracle action imply $\mathbb{E}[N(T)]=T+o(T)$, versus $JT$ for independent runs. We also obtain finite-sample variance bounds for pairwise policy contrasts. Experiments on reward-model evaluation, multiple-choice language-model evaluation, and adaptive search policies demonstrate substantial improvements in the cost--precision frontier.
Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-14 in cs.AI).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.12831 — 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 2026-10-19T04:13:06.706Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-14?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).