Knowledge-Augmented Explainable and Interpretable Learning for Anomaly Detection and Diagnosis

TL;DR

This paper discusses how combining knowledge-based and data-driven methods enhances explainability and interpretability in anomaly detection and diagnosis, emphasizing transparency and computational sensemaking.

Contribution

It introduces a framework for knowledge-augmented explainable learning, showcasing various approaches from simple interpretable models to advanced neuro-symbolic techniques.

Findings

Enhanced understandability and transparency in anomaly detection.

Improved interpretability through neuro-symbolic approaches.

Demonstrated effectiveness across different domains.

Abstract

Knowledge-augmented learning enables the combination of knowledge-based and data-driven approaches. For anomaly detection and diagnosis, understandability is typically an important factor, especially in high-risk areas. Therefore, explainability and interpretability are also major criteria in such contexts. This chapter focuses on knowledge-augmented explainable and interpretable learning to enhance understandability, transparency and ultimately computational sensemaking. We exemplify different approaches and methods in the domains of anomaly detection and diagnosis - from comparatively simple interpretable methods towards more advanced neuro-symbolic approaches.