CSI: Novelty Detection via Contrastive Learning on Distributionally Shifted Instances

TL;DR

CSI introduces a contrastive learning approach that compares samples with distributionally-shifted augmentations to improve novelty detection across various scenarios and datasets.

Contribution

The paper proposes a novel contrastive learning method called CSI that contrasts samples with their distributionally-shifted augmentations for enhanced novelty detection.

Findings

CSI outperforms existing methods in multiple novelty detection benchmarks.

The proposed detection score improves identification of out-of-distribution samples.

CSI is effective in both labeled and unlabeled, one-class and multi-class scenarios.

Abstract

Novelty detection, i.e., identifying whether a given sample is drawn from outside the training distribution, is essential for reliable machine learning. To this end, there have been many attempts at learning a representation well-suited for novelty detection and designing a score based on such representation. In this paper, we propose a simple, yet effective method named contrasting shifted instances (CSI), inspired by the recent success on contrastive learning of visual representations. Specifically, in addition to contrasting a given sample with other instances as in conventional contrastive learning methods, our training scheme contrasts the sample with distributionally-shifted augmentations of itself. Based on this, we propose a new detection score that is specific to the proposed training scheme. Our experiments demonstrate the superiority of our method under various novelty detection scenarios, including unlabeled one-class, unlabeled multi-class and labeled multi-class settings, with various image benchmark datasets. Code and pre-trained models are available at https://github.com/alinlab/CSI.