打破预测还不够：面向时序图的指定反事实解释

中文摘要 AI 辅助

时序图反事实解释通常通过改变过去的事件来改变或推翻原始预测，但未指定替代预测。然而，面对预测结果的用户往往想知道，哪些过去条件会导致特定替代结果发生。我们将这种目标特定的问题形式化为指定反事实：给定原始预测A和搜索前固定的反事实目标B，寻找一个低成本的过去事件干预，使得同一预测器将B列为最高排名。我们的轨迹引导干预搜索将A的完整执行与B的重建不完整执行进行对比，将差异映射为删除、插入、重连、重新标记和移动操作，并通过精确重放验证B。我们在连续时间动态图上使用LiFTER，在时序知识图谱上使用TLogic实例化该原则。在CTDG上，该方法保留了黑盒贪心成功率的85.7%-93.6%，同时将预测器评估减少了75.0%-80.0%；在TKG上，在600次比较中，有74.8%达到了指定反事实目标。因此，可执行轨迹成为构建未选择替代条件条件的计算结构，而不仅仅是用于解释已有预测的记录。

英文摘要

Can a trace explaining model execution also compute the changes needed for a specified alternative prediction? We propose trace-guided intervention search, which uses executable reasoning traces as an intermediate representation for intervention synthesis. A Specified-Foil Counterfactual edits past events so that a frozen temporal predictor selects a designated foil. Our method constructs facts and replacement values from completed original and foil executions and recovered unmet conditions. Proposal generation constructs edits and selects candidates within a fixed cap; exact replay verifies foil top-1 outcomes among retained edits and compositions. Implemented in LiFTER for continuous-time dynamic graphs (CTDGs) and TLogic for temporal knowledge graphs (TKGs), the method improves success over coordinate-based proposal generation by 13.7-34.7 percentage points on four CTDG datasets and 60.0-83.3 points on two TKG datasets under matched downstream search and a proposal cap of 32. Separate shared-candidate comparisons retain 85.7-93.6% of black-box greedy's CTDG success rate with 75.0-80.0% fewer predictor evaluations. A Pulse case study confirms simulator-level survival for five of six interventions. Executable traces thus provide both explanatory evidence and a reusable computational representation for constructing and testing specified alternatives.

发表机构

• Konkuk University（建国大学）

机构由 AI 辅助整理，请以论文原文为准。