DSV: An Alignment Validation Loss for Self-supervised Outlier Model Selection

TL;DR

This paper introduces DSV, an unsupervised validation loss for selecting effective self-supervised anomaly detection models by evaluating augmentation alignment, leading to improved detection accuracy across multiple real-world tasks.

Contribution

The paper proposes DSV, a novel unsupervised validation method that guides hyperparameter tuning for self-supervised anomaly detection models, enhancing their performance.

Findings

DSV outperforms baseline methods on 21 real-world tasks

Theoretical analysis confirms the effectiveness of surrogate losses

Improved model selection leads to higher anomaly detection accuracy

Abstract

Self-supervised learning (SSL) has proven effective in solving various problems by generating internal supervisory signals. Unsupervised anomaly detection, which faces the high cost of obtaining true labels, is an area that can greatly benefit from SSL. However, recent literature suggests that tuning the hyperparameters (HP) of data augmentation functions is crucial to the success of SSL-based anomaly detection (SSAD), yet a systematic method for doing so remains unknown. In this work, we propose DSV (Discordance and Separability Validation), an unsupervised validation loss to select high-performing detection models with effective augmentation HPs. DSV captures the alignment between an augmentation function and the anomaly-generating mechanism with surrogate losses, which approximate the discordance and separability of test data, respectively. As a result, the evaluation via DSV leads to selecting an effective SSAD model exhibiting better alignment, which results in high detection accuracy. We theoretically derive the degree of approximation conducted by the surrogate losses and empirically show that DSV outperforms a wide range of baselines on 21 real-world tasks.