用输入空间扰动半径提升模型误判检测能力，无需额外数据。
RAT: Boosting Misclassification Detection Ability without Extra Data
- 以输入空间扰动半径作为置信度指标，设计高效估计算法。
- 相比之前方法，AURC降低29.3%，FPR@95TPR降低21.62%。
- 适合高风险场景下需可靠错误检测的模型部署。
随着深度神经网络在自动驾驶、医疗等高风险领域广泛应用，模型错误预测的检测与干预变得至关重要。本文从对抗扰动视角出发，提出以鲁棒半径（即输入空间边际）作为置信度指标，并设计两种高效估计算法：RR-BS 和 RR-Fast，用于误分类检测。进一步提出一种名为半径感知训练（Radius Aware Training, RAT）的训练方法，增强模型识别错误预测的能力。大量实验表明，该方法相较先前方法，在 AURC 上最多降低 29.3%，在 FPR@95TPR 上降低 21.62%。
原文摘要 · Abstract (English)
As deep neural networks(DNN) become increasingly prevalent, particularly in high-stakes areas such as autonomous driving and healthcare, the ability to detect incorrect predictions of models and intervene accordingly becomes crucial for safety. In this work, we investigate the detection of misclassified inputs for image classification models from the lens of adversarial perturbation: we propose to use robust radius (a.k.a. input-space margin) as a confidence metric and design two efficient estimation algorithms, RR-BS and RR-Fast, for misclassification detection. Furthermore, we design a training method called Radius Aware Training (RAT) to boost models' ability to identify mistakes. Extensive experiments show our method could achieve up to 29.3% reduction on AURC and 21.62% reduction in FPR@95TPR, compared with previous methods.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。