UNSUPERVISED ANOMALY DETECTION FROM SEMANTIC SIMILARITY SCORES

Nima Rafiee, Rahil Gholamipoor, Markus Kollmann

Keywords: Anomaly Detection, Out-of-Distribution Detection, Novelty Detection

Abstract: In this paper we present SemSAD, a simple and generic framework for detecting examples that lie out-of-distribution (OOD) for a given training set. The approach is based on learning a semantic similarity measure to find for a given test example the semantically closest example in the training set and then using a discriminator to classify whether the two examples show sufficient semantic dissimilarity such that the test example can be rejected as OOD. We are able to outperform previous approaches for anomaly, novelty, or out-of-distribution detection in the visual domain by a large margin. In particular we obtain AUROC values close to one for the challenging task of detecting examples from CIFAR-10 as out-of-distribution given CIFAR-100 as in-distribution, without making use of label information.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

One-sentence Summary: Combining Contrastive Learning and Discriminative Learning for unsupervised Anomaly Detection

Community Implementations: [ 1 code implementation](https://www.catalyzex.com/paper/arxiv:2012.00461/code)

Reviewed Version (pdf): https://openreview.net/references/pdf?id=KV1wEo8vEy