Going Beyond Familiar Features for Deep Anomaly Detection

arXiv:2310.00797v3 Announce Type: replace
Abstract: Anomaly Detection (AD) is a critical task that involves identifying observations that do not conform to a learned model of normality. Prior work in deep AD is predominantly based on a familiarity hypothesis, where familiar features serve as the reference in a pre-trained embedding space. While this strategy has proven highly successful, it turns out that it causes consistent false negatives when anomalies consist of truly novel features that are not well captured by the …

abstract anomaly anomaly detection arxiv beyond cs.lg detection embedding features hypothesis normality prior reference serve space type work