用于错位二元分类器的逻辑警报器
A logical alarm for misaligned binary classifiers
AI总结:
本文针对二元分类智能体的决策分歧问题,基于通用公理构建仅需无标注数据的逻辑警报器,可检测集成模型的成员故障,为安全可保障AI研究提供新方法。
摘要:
若两个智能体的决策存在分歧,我们可推测二者并非都正确。本文将这一直觉形式化,用于评估执行二元分类任务的智能体。二者在联合测试上的一致与分歧情况,可帮助我们确立与其响应逻辑一致的唯一群体评估结果。该方法通过构建一组公理(代数关系)实现,这些公理是所有二元响应器评估都必须普遍遵循的。对于规模为N的任意集成模型,都可得到一套完整的此类公理。本文利用N=1、2时的公理构建了一个完全逻辑化的警报器——它仅需无标注数据,就能证明集成模型中至少有一个成员出现故障。文章还讨论了该方法与形式化软件验证的相似性,以及它在当前安全可保障AI相关研究议程中的实用价值。
英文摘要:
If two agents disagree in their decisions, we may suspect they are not both correct. This intuition is formalized for evaluating agents that have carried out a binary classification task. Their agreements and disagreements on a joint test allow us to establish the only group evaluations logically consistent with their responses. This is done by establishing a set of axioms (algebraic relations) that must be universally obeyed by all evaluations of binary responders. A complete set of such axioms are possible for each ensemble of size N. The axioms for $N = 1, 2$ are used to construct a fully logical alarm - one that can prove that at least one ensemble member is malfunctioning using only unlabeled data. The similarities of this approach to formal software verification and its utility for recent agendas of safe guaranteed AI are discussed.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。