Env-Aware Anomaly Detection: Ignore Style Changes, Stay True to Content!Download PDF

Published: 21 Oct 2022, Last Modified: 05 May 2023NeurIPS 2022 Workshop DistShift PosterReaders: Everyone
Keywords: distribution shift, anomaly, non-stationary, unsupervised, natural distribution shifts, domains, robustness
TL;DR: Env-aware pretraining for unsupervised anomaly detection, under style distribution shifts.
Abstract: We introduce a formalization and benchmark for the unsupervised anomaly detection task in the distribution-shift scenario. Our work builds upon the iWildCam dataset, and, to the best of our knowledge, we are the first to propose such an approach for visual data. We empirically validate that environment-aware methods perform better in such cases when compared with the basic Empirical Risk Minimization (ERM). We next propose an extension for generating positive samples for contrastive methods that considers the environment labels when training, improving the ERM baseline score by 8.7%.
0 Replies