arXiv：cs.LG· Utkarsh Grover, Ravi Ranjan, Agoritsa Polyzou, Wyatt Mackey, J. Morris Chang, Leonardo Bobadilla, Xiaomin Lin·· 7 小时前AI 评分33
DePICT：面向受限下游任务的决策保持接口
DePICT: Decision-Preserving Interface for Constrained Downstream Tasks
AI 导读
DePICT 通过按优化器解的敏感度对上下文方向排序，构建决策保持接口，用于约束优化场景中筛选下游智能体真正依赖的输入方向。基于 KKT 条件的分析表明，出现在活跃优化问题中不等于影响最终决策，部分方向的效应会被对偶变量吸收。在受控诊断中，它精确恢复决策相关接口，将线性预测器 regret 降至 0.009，最强基线为 0.475。
正文
Abstract:A constrained optimization problem may involve a parameter in its objective and active constraints, yet the final decision may remain insensitive to small changes in that parameter. This raises a fundamental question: which inputs does a decision making system truly depend on? Building on this question, we introduce DePICT, a procedure for constructing decision preserving interfaces by ranking context directions according to the optimizer's solution sensitivity and aggregating them across an operating regime. We study this problem in a high dimensional setting where primitive context parameterizes a constrained task and the downstream agent observes only a selected subset of context directions. For locally regular constrained programs, we derive a Karush Kuhn Tucker (KKT) based characterization of when a context direction is optimizer relevant. Our analysis shows that appearing in the active optimization problem does not necessarily imply that a variable affects the final decision. Some context directions can alter the KKT conditions while leaving the optimal solution unchanged because their effect is absorbed by the dual variables. DePICT is designed to remove exactly these directions. In a controlled diagnosis, it recovers the decision relevant interface exactly and reduces linear predictor regret to 0.009, compared with 0.475 for the strongest competing baseline.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.03945 [cs.LG] |
| (or arXiv:2610.03945v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.03945 arXiv-issued DOI via DataCite |
Submission history
From: Utkarsh Grover [view email]
[v1]
Fri, 2 Oct 2026 19:00:11 UTC (1,396 KB)
[v2]
Tue, 6 Oct 2026 04:10:14 UTC (1,401 KB)
来源：arXiv：cs.LG · arxiv.org