ProME：面向组鲁棒学习的带修复感知选择的原型-边际环境
ProME: Prototype-Margin Environments with Repair-Aware Selection for Group-Robust Learning
浏览论文内容
中文总结 AI 辅助
针对无训练组标签的组鲁棒学习问题，提出ProME方法，通过划分原型边际构建平衡环境并对齐决策与部署预测器，实验显示其平均最坏组准确性优于对比方法。
中文摘要 AI 辅助
组鲁棒学习在训练组标签不可用时，对维持稀有子群体的准确性至关重要。然而，现有方法通常从单独的参考模型推断环境，并在拟合部署时使用的分类器之前选择表示，使得这两个决策与部署的预测器不一致。在本研究中，我们将无训练组标签的组鲁棒学习形式化为带修复感知选择的内生环境（ERAS）问题，并提出ProME（原型-边际环境），以使这两个决策与部署的预测器对齐。ProME沿训练轨迹将原型边际以中位数划分，以构建近似平衡的环境，并在带有组注释的验证数据上拟合组平衡线性头，以通过验证最坏组准确性对所得预测器进行排名。我们从理论上为固定预测器和划分的推断环境上的最坏风险提供了界，表明该界在显式对齐条件下可迁移到 oracle 组。大量实验表明，原型边际丰富了捷径冲突示例，分类器修复重塑了候选评估，且在相同组标签访问权限下，ProME 在对比方法中实现了最高的平均最坏组准确性。
英文摘要
Group-robust learning is crucial for maintaining accuracy on rare subpopulations when training-group labels are unavailable. However, existing methods often infer environments from a separate reference model and select representations before fitting the classifier used at deployment, leaving both decisions misaligned with the deployed predictor. In this work, we formulate group robustness without training-group labels as the endogenous environments with repair-aware selection (ERAS) problem, and propose ProME (Prototype-Margin Environments) to align both decisions with the deployed predictor. ProME splits prototype margins at their median to construct approximately balanced environments along the training trajectory, and fits a group-balanced linear head on group-annotated validation data to rank the resulting predictors by validation worst-group accuracy. We theoretically bound the worst risk across the inferred environments for a fixed predictor and partition, showing that this bound transfers to the oracle groups under an explicit alignment condition. Extensive experiments show that prototype margins enrich shortcut-conflicting examples, classifier repair reshapes candidate evaluation, and ProME achieves the highest average worst-group accuracy among the compared methods with the same group-label access.
发表机构
- Shenzhen Key Laboratory of Safety and Security for Next Generation of Industrial Internet（深圳市下一代工业互联网安全重点实验室）
- Department of Statistics and Data Science, Southern University of Science and Technology（南方科技大学统计与数据科学系）
机构由 AI 辅助整理，请以论文原文为准。
