CADet: Fully Self-Supervised Out-Of-Distribution Detection With Contrastive Learning

TL;DR

This paper introduces CADet, a fully self-supervised contrastive learning method for detecting out-of-distribution samples, including unseen classes and adversarial attacks, outperforming existing methods on multiple benchmarks.

Contribution

The work presents CADet, a novel self-supervised contrastive learning approach for OOD detection that does not require labeled data or OOD examples, improving robustness and performance.

Findings

CADet outperforms existing adversarial detection methods on ImageNet.

CADet achieves comparable results to label-based methods on ImageNet-O and iNaturalist.

The combined MMD and contrastive learning approach enhances OOD detection confidence.

Abstract

Handling out-of-distribution (OOD) samples has become a major stake in the real-world deployment of machine learning systems. This work explores the use of self-supervised contrastive learning to the simultaneous detection of two types of OOD samples: unseen classes and adversarial perturbations. First, we pair self-supervised contrastive learning with the maximum mean discrepancy (MMD) two-sample test. This approach enables us to robustly test whether two independent sets of samples originate from the same distribution, and we demonstrate its effectiveness by discriminating between CIFAR-10 and CIFAR-10.1 with higher confidence than previous work. Motivated by this success, we introduce CADet (Contrastive Anomaly Detection), a novel method for OOD detection of single samples. CADet draws inspiration from MMD, but leverages the similarity between contrastive transformations of a same sample. CADet outperforms existing adversarial detection methods in identifying adversarially perturbed samples on ImageNet and achieves comparable performance to unseen label detection methods on two challenging benchmarks: ImageNet-O and iNaturalist. Significantly, CADet is fully self-supervised and requires neither labels for in-distribution samples nor access to OOD examples.