Toward Faithful Explanations in Acoustic Anomaly Detection
This work addresses the need for interpretability in acoustic anomaly detection for industrial applications, offering an incremental improvement in explanation quality.
The study compared autoencoder (AE) and mask autoencoder (MAE) models for audio anomaly detection, finding that MAE provides more faithful and temporally precise explanations despite slightly lower detection performance, as evaluated using attribution methods and a proposed perturbation-based faithfulness metric in a real industrial scenario.
Interpretability is essential for user trust in real-world anomaly detection applications. However, deep learning models, despite their strong performance, often lack transparency. In this work, we study the interpretability of autoencoder-based models for audio anomaly detection, by comparing a standard autoencoder (AE) with a mask autoencoder (MAE) in terms of detection performance and interpretability. We applied several attribution methods, including error maps, saliency maps, SmoothGrad, Integrated Gradients, GradSHAP, and Grad-CAM. Although MAE shows a slightly lower detection, it consistently provides more faithful and temporally precise explanations, suggesting a better alignment with true anomalies. To assess the relevance of the regions highlighted by the explanation method, we propose a perturbation-based faithfulness metric that replaces them with their reconstructions to simulate normal input. Our findings, based on experiments in a real industrial scenario, highlight the importance of incorporating interpretability into anomaly detection pipelines and show that masked training improves explanation quality without compromising performance.