LINe: Out-of-Distribution Detection by Leveraging Important Neurons

It is important to quantify the uncertainty of input samples, especially in mission-critical domains such as autonomous driving and healthcare, where failure predictions on out-of-distribution (OOD) data are likely to cause big problems. OOD detection problem fundamentally begins in that the model cannot express what it is not aware of. Post-hoc OOD detection approaches are widely explored because they do not require an additional re-training process which might degrade the model's performance and increase the training cost. In this study, from the perspective of neurons in the deep layer of the model representing high-level features, we introduce a new aspect for analyzing the difference in model outputs between in-distribution data and OOD data. We propose a novel method, Leveraging Important Neurons (LINe), for post-hoc Out of distribution detection. Shapley value-based pruning reduces the effects of noisy outputs by selecting only high-contribution neurons for predicting specific classes of input data and masking the rest. Activation clipping fixes all values above a certain threshold into the same value, allowing LINe to treat all the class-specific features equally and just consider the difference between the number of activated feature differences between in-distribution and OOD data. Comprehensive experiments verify the effectiveness of the proposed method by outperforming state-of-the-art post-hoc OOD detection methods on CIFAR-10, CIFAR-100, and ImageNet datasets.

PDF Abstract CVPR 2023 PDF CVPR 2023 Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Out-of-Distribution Detection ImageNet-1k vs Curated OODs (avg.) LINe (ResNet-50) AUROC 95.03 # 5
FPR95 20.70 # 4
Out-of-Distribution Detection ImageNet-1k vs iNaturalist LINe (ResNet-50) FPR95 12.26 # 7
AUROC 97.56 # 8
Out-of-Distribution Detection ImageNet-1k vs Places LINe (ResNet50) FPR95 28.52 # 1
AUROC 92.85 # 1
Out-of-Distribution Detection ImageNet-1k vs SUN LINe (ResNet50) FPR95 19.48 # 1
AUROC 95.26 # 1
Out-of-Distribution Detection ImageNet-1k vs Textures LINe (ResNet-50) FPR95 22.54 # 9
AUROC 94.44 # 10