arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.10062cs.LGcs.CV
收藏

Vendi Novelty Scores for Out-of-Distribution Detection

Vendi Novelty Scores for Out-of-Distribution Detection

Lewis-Sigler Institute For Integrative Genomics, Princeton University(普林斯顿大学整合基因组学研究所);Department of Computer Science, Princeton University(普林斯顿大学计算机科学系)

Amey P. Pasarkar, Adji Bousso Dieng

AI总结:

本文提出了一种基于Vendi Scores的Vendi Novelty Score(VNS)方法,从多样性角度解决分布外检测问题,该方法无需密度建模,具有线性时间复杂度和非参数特性,并在多个图像分类基准上实现了最先进的OOD检测性能。

摘要:

Out-of-distribution (OOD) detection is critical for the safe deployment of machine learning systems. Existing post-hoc detectors typically rely on model confidence scores or likelihood estimates in feature space, often under restrictive distributional assumptions. In this work, we introduce a third paradigm and formulate OOD detection from a diversity perspective. We propose the Vendi Novelty Score (VNS), an OOD detector based on the Vendi Scores (VS), a family of similarity-based diversity metrics. VNS quantifies how much a test sample increases the VS of the in-distribution feature set, providing a principled notion of novelty that does not require density modeling. VNS is linear-time, non-parametric, and naturally combines class-conditional (local) and dataset-level (global) novelty signals. Across multiple image classification benchmarks and network architectures, VNS achieves state-of-the-art OOD detection performance. Remarkably, VNS retains this performance when computed using only 1% of the training data, enabling deployment in memory- or access-constrained settings.

英文摘要:

Out-of-distribution (OOD) detection is critical for the safe deployment of machine learning systems. Existing post-hoc detectors typically rely on model confidence scores or likelihood estimates in feature space, often under restrictive distributional assumptions. In this work, we introduce a third paradigm and formulate OOD detection from a diversity perspective. We propose the Vendi Novelty Score (VNS), an OOD detector based on the Vendi Scores (VS), a family of similarity-based diversity metrics. VNS quantifies how much a test sample increases the VS of the in-distribution feature set, providing a principled notion of novelty that does not require density modeling. VNS is linear-time, non-parametric, and naturally combines class-conditional (local) and dataset-level (global) novelty signals. Across multiple image classification benchmarks and network architectures, VNS achieves state-of-the-art OOD detection performance. Remarkably, VNS retains this performance when computed using only 1% of the training data, enabling deployment in memory- or access-constrained settings.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑