Consistency-based Abductive Reasoning over Perceptual Errors of Multiple Pre-trained Models in Novel Environments

该论文提出了一种基于一致性的测试时溯因推理框架,通过逻辑编程将多个预训练模型的预测及其错误检测规则编码,利用整数规划或启发式搜索算法在满足逻辑一致性约束的前提下最大化预测覆盖率,从而在分布偏移的新环境中有效缓解单一模型性能下降并提升整体精度与召回率。

Mario Leiva, Noel Ngu, Joshua Shay Kricheli, Aditya Taparia, Ransalu Senanayake, Paulo Shakarian, Nathaniel Bastian, John Corcoran, Gerardo SimariThu, 12 Ma🤖 cs.AI