arXiv cs.AI· Bo Xue, Ji Cheng, Shen-Huan Lyu, Yuanyu Wan, Shuang Qiu·· 3 天前AI 评分22
BMA-GAI：预算约束下的多属性验证实现测试时扩展
Test-Time Scaling via Budgeted Multi-Attribute Verification
AI 导读
研究者提出 BMA-GAI 算法，将 LLM 答案验证建模为全局预算下的多属性优质臂识别问题，通过成本感知的候选选择与属性自适应采样，在每次观测中同时完成分配引导与随时有效认证，无需单独确认阶段。该算法获得渐近覆盖保证，并推导出匹配的信息论逆定理，证明其在临界预算之外具有一阶最优性。合成基准与 LLM 答案验证任务实验显示，BMA-GAI 比对比方法更高效地分配验证预算，认证出更多高质量候选答案。
正文 · 原文
Abstract:Verifying LLM-generated answers under a shared computational budget requires jointly deciding which candidates to inspect and which verification attributes to evaluate. We formulate this problem as multi-attribute good-arm identification under a global budget: each candidate is an arm evaluated along several costly attributes, and the goal is to certify as many candidates as possible whose mean scores exceed the prescribed thresholds on all attributes. We propose \textsc{BMA-GAI}, an algorithm that combines cost-aware arm selection with adaptive sampling of attributes. Every observation serves both to guide adaptive allocation and to support anytime-valid certification, which removes the need for a separate confirmation stage. We establish an asymptotic coverage guarantee for \textsc{BMA-GAI} and derive a matching information-theoretic converse that characterizes the intrinsic complexity of the problem, thereby proving that \textsc{BMA-GAI} is first-order optimal away from critical budget levels. Experiments on synthetic benchmarks and an LLM answer-verification task show that \textsc{BMA-GAI} allocates the verification budget more efficiently and certifies more high-quality candidates than competing methods.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2609.34322 [cs.AI] |
| (or arXiv:2609.34322v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2609.34322 arXiv-issued DOI via DataCite |
Submission history
From: Bo Xue [view email]
[v1]
Mon, 28 Sep 2026 05:01:09 UTC (213 KB)
[v2]
Mon, 12 Oct 2026 14:04:09 UTC (210 KB)
来源：arXiv cs.AI · arxiv.org