A general-purpose method for applying Explainable AI for Anomaly Detection

TL;DR

This paper introduces a general-purpose explainability method for unsupervised anomaly detection, leveraging Integrated Gradients to improve interpretability and diagnosis in real-world datasets.

Contribution

It presents a novel approach that applies explainability to unsupervised anomaly detection, bridging algorithmic and cognitive aspects for practical diagnosis.

Findings

Integrated Gradients reduces attribution errors compared to alternatives

The method is effective on real-world labeled datasets

Provides a principled approach to explainability in unsupervised anomaly detection

Abstract

The need for explainable AI (XAI) is well established but relatively little has been published outside of the supervised learning paradigm. This paper focuses on a principled approach to applying explainability and interpretability to the task of unsupervised anomaly detection. We argue that explainability is principally an algorithmic task and interpretability is principally a cognitive task, and draw on insights from the cognitive sciences to propose a general-purpose method for practical diagnosis using explained anomalies. We define Attribution Error, and demonstrate, using real-world labeled datasets, that our method based on Integrated Gradients (IG) yields significantly lower attribution errors than alternative methods.