arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2503.10468cs.CVcs.LG

OODD:使用动态字典的测试时分布外检测

OODD: Test-time Out-of-Distribution Detection with Dynamic Dictionary

Shanghai Jiao Tong University(上海交通大学);Tianjin University(天津大学);The Chinese University of Hong Kong(香港中文大学);Shanghai Artificial Intelligence Laboratory(上海人工智能实验室)

Yifeng Yang, Lin Zhu, Zewen Sun, Hengyu Liu, Qinying Gu, Nanyang Ye

AI总结:

针对测试时 OOD 样本与训练异常差异较大的问题,本文提出无需微调的动态字典方法 OODD,并通过双重稳定机制与 KNN 优化,在 OpenOOD 上显著提升检测性能与速度。

摘要:

分布外(OOD)检测对深度学习模型而言仍具挑战性,尤其是当测试时的 OOD 样本与训练异常样本存在显著差异时。我们提出 OODD,这是一种新颖的测试时 OOD 检测方法,可在不进行微调的情况下动态维护并更新 OOD 字典。该方法采用基于优先队列的字典,在测试过程中积累具有代表性的 OOD 特征,并结合用于分布内(ID)样本的信息丰富的 inlier 采样策略。为确保测试早期的稳定性能,我们提出一种双重 OOD 稳定机制,该机制利用从 ID 数据中有策略地生成的异常样本。据我们所知,在 OpenOOD 基准上的大量实验表明,OODD 显著优于现有方法;在 CIFAR-100 Far OOD 检测中,其 FPR95 相比最先进方法提升了 26.0%。此外,我们提出了基于 KNN 的 OOD 检测框架的优化变体,在保持检测性能的同时实现了 3 倍加速。

英文摘要:

Out-of-distribution (OOD) detection remains challenging for deep learning models, particularly when test-time OOD samples differ significantly from training outliers. We propose OODD, a novel test-time OOD detection method that dynamically maintains and updates an OOD dictionary without fine-tuning. Our approach leverages a priority queue-based dictionary that accumulates representative OOD features during testing, combined with an informative inlier sampling strategy for in-distribution (ID) samples. To ensure stable performance during early testing, we propose a dual OOD stabilization mechanism that leverages strategically generated outliers derived from ID data. To our best knowledge, extensive experiments on the OpenOOD benchmark demonstrate that OODD significantly outperforms existing methods, achieving a 26.0% improvement in FPR95 on CIFAR-100 Far OOD detection compared to the state-of-the-art approach. Furthermore, we present an optimized variant of the KNN-based OOD detection framework that achieves a 3x speedup while maintaining detection performance.

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。

↑