跳到正文
arXiv cs.AI· Kenton White·· 4 小时前AI 评分20

FLEX:面向多标签可解释性的带符号特征归因方法

Right Labels, Wrong Reasons: Signed Feature Attribution for Multi-Label Explainability

AI 导读

论文提出 FLEX(Fisher LDA Explainer),一种在完整二值标签矩阵上拟合单一直接多标签 Fisher LDA 的全局代理模型,可在统一坐标系中为每个标签生成带符号特征归因列表。

正文 · 原文

View PDF HTML (experimental)

Abstract:When a multi-label classifier assigns the right labels, it may do so for the wrong reasons. Diffuse feature combinations may vary across classifier families and lack a dominant feature. This paper proposes FLEX (Fisher LDA Explainer), a global surrogate that fits a single direct multi-label Fisher LDA over the full binary label matrix, producing signed feature attribution lists for each label in a single coherent coordinate system. Across the evaluated benchmarks and classifier families, FLEX reproduces the black box's predictions with consistently high fidelity. Pooling covariance across labels outperforms a per-label alternative on most datasets, with the advantage growing largest where labels are numerous and positive examples are scarce. FLEX's fidelity-complexity Pareto curve is monotonically increasing, giving practitioners a predictable path to higher-fidelity explanations as more features are used, and its in-sample fidelity matches or leads Binary Relevance alternatives once a modest feature budget is allowed. The magnitude of each discriminant weight vector bounds its label's separability, flagging labels that may be predicted correctly for the wrong reasons. The evaluated classifiers largely agree on which predictions to make, but reach those decisions through different feature subsets when no single feature dominates.
Comments: 7 pages, 1 figure, 5 tables. Accepted at the 2027 IEEE Symposium Series on Computational Intelligence (IEEE SSCI 2027), Gold Coast, Australia, February 2027
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
MSC classes: 68T05, 62H30
ACM classes: I.2.6; I.5.2
Cite as: arXiv:2611.22515 [cs.LG]
  (or arXiv:2611.22515v1 [cs.LG] for this version)
  https://doi.org/10.48550/arXiv.2611.22515

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Kenton White [view email]
[v1] Tue, 17 Nov 2026 10:07:21 UTC (43 KB)

来源:arXiv cs.AI · arxiv.org