发表机构
University of Chinese Academy of Sciences; Institute of Information Engineering, Chinese Academy of Sciences（中国科学院大学; 中国科学院信息工程研究所）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
VStress通过相关性感知审计和自适应预算分配，将重复验证器的相关性从事后警告转为可审计决策，提升平衡准确率并控制成本。
AI 中文摘要
重复的验证器调用只有在贡献条件信息时才有用。我们引入了VStress，一个可审计的重放合约，以及VStress-CA，一种相关性感知的分配策略，该策略在密封的校准分割上估计未查询验证器的条件边际信息，对不确定性进行折扣，按调用成本归一化，并在下一次调用不具信息性时停止或弃权（不执行）。控制器在加入干净预言机之前冻结其决策和成本账本；依赖转移警报禁用通道偏好并回退到精确停止。受控审计给出了机制边界：在35%对称腐败下，多数-5将平衡准确率从0.6578提高到0.7739，而在65%时则损失0.1226点。在匹配的固定预算比较中，广度、冗余和自适应分配分别获得0.6048、0.6375和0.6538的平衡准确率，每次调用3.4216次，VStress-CA的RLVR得分为0.6417。依赖诊断也从同模型重复增加到跨家族通道，条件边际增益分别为0.0126、0.0462和0.0913。这些测量将相关性从事后警告转变为可审计的分配决策。
英文摘要
Repeated verifier calls are useful only when they contribute conditional information. We introduce VStress, an auditable replay contract, and VStress-CA, a correlation-aware allocation policy that estimates the conditional marginal information of an unqueried verifier on a sealed calibration split, discounts uncertainty, normalizes by call cost, and stops or abstains when the next call is not informative. The controller freezes its decision and cost ledger before joining the clean oracle; a dependence-shift alarm disables channel preference and falls back to exact-stop. The controlled audit gives the mechanism boundary: at 35% symmetric corruption, majority-5 improves balanced accuracy from 0.6578 to 0.7739, whereas at 65% it loses 0.1226 points. In the matched fixed-budget comparison, breadth, redundancy, and adaptive allocation obtain balanced accuracies 0.6048, 0.6375, and 0.6538, with 3.4216 calls per item and an RLVR score of 0.6417 for VStress-CA. Dependence diagnostics also increase from same-model repeats to cross-family channels, with conditional marginal gains of 0.0126, 0.0462, and 0.0913. These measurements turn correlation from a post-hoc warning into an auditable allocation decision.
Comments27 pages, 5 figures
