# A Framework for Evaluating Faithfulness in Explainable AI for Machine Anomalous Sound Detection Using Frequency-Band Perturbation

arXiv: https://arxiv.org/abs/2601.19017 · HTML: https://mlsanity.com/paper/2601.19017/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 11 of its 27 references that are in the map.
- First posted on arXiv 26 Jan 2026.

## Builds on (most cited first, 11 of 11)

- [Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization](https://mlsanity.com/paper/1610.02391/) (2016)
- [Axiomatic Attribution for Deep Networks](https://mlsanity.com/paper/1703.01365/) (2017)
- [SmoothGrad: removing noise by adding noise](https://mlsanity.com/paper/1706.03825/) (2017)
- [Interpretable Explanations of Black Boxes by Meaningful Perturbation](https://mlsanity.com/paper/1704.03296/) (2017)
- [ToyADMOS2: Another dataset of miniature-machine operating sounds for anomalous sound detection under domain shift conditions](https://mlsanity.com/paper/2106.02369/) (2021)
- [MIMII DG: Sound Dataset for Malfunctioning Industrial Machine Investigation and Inspection for Domain Generalization Task](https://mlsanity.com/paper/2205.13879/) (2022)
- [Description and Discussion on DCASE 2023 Challenge Task 2: First-Shot Unsupervised Anomalous Sound Detection for Machine Condition Monitoring](https://mlsanity.com/paper/2305.07828/) (2023)
- [Self-Supervised Learning for Anomalous Sound Detection](https://mlsanity.com/paper/2312.09578/) (2023)
- [Faithful and Plausible Natural Language Explanations for Image Classification: A Pipeline Approach](https://mlsanity.com/paper/2407.20899/) (2024)
- [Exploring Self-Supervised Audio Models for Generalized Anomalous Sound Detection](https://mlsanity.com/paper/2508.12230/) (2025)
- [Solving the enigma: Enhancing faithfulness and comprehensibility in explanations of deep networks](https://mlsanity.com/paper/2405.10008/) (2024)

## Abstract (arXiv)

Explainable AI (XAI) is commonly applied to anomalous sound detection (ASD) models to identify which time-frequency regions of an audio signal contribute to an anomaly decision. However, most audio explanations rely on qualitative inspection of saliency maps, leaving open the question of whether these attributions accurately reflect the spectral cues the model uses. In this work, we introduce a new quantitative framework for evaluating XAI faithfulness in machine-sound analysis by directly linking attribution relevance to model behaviour through systematic frequency-band removal. This approach provides an objective measure of whether an XAI method for machine ASD correctly identifies frequency regions that influence an ASD model's predictions. By using four widely adopted methods, namely Integrated Gradients, Occlusion, Grad-CAM and SmoothGrad, we show that XAI techniques differ in reliability, with Occlusion demonstrating the strongest alignment with true model sensitivity and gradient-+based methods often failing to accurately capture spectral dependencies. The proposed framework offers a reproducible way to benchmark audio explanations and enables more trustworthy interpretation of spectrogram-based ASD systems.