# On the reliability of feature attribution methods for speech classification

arXiv: https://arxiv.org/abs/2505.16406 · HTML: https://mlsanity.com/paper/2505.16406/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 26 of its 31 references that are in the map.
- First posted on arXiv 22 May 2025.

## Builds on (most cited first, 26 of 26)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- ["Why Should I Trust You?": Explaining the Predictions of Any Classifier](https://mlsanity.com/paper/1602.04938/) (2016)
- [Robust Speech Recognition via Large-Scale Weak Supervision](https://mlsanity.com/paper/2212.04356/) (2022)
- [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://mlsanity.com/paper/2006.11477/) (2020)
- [Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps](https://mlsanity.com/paper/1312.6034/) (2013)
- [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://mlsanity.com/paper/2106.07447/) (2021)
- [Visualizing and Understanding Convolutional Networks](https://mlsanity.com/paper/1311.2901/) (2013)
- [Common Voice: A Massively-Multilingual Speech Corpus](https://mlsanity.com/paper/1912.06670/) (2019)
- [Captum: A unified and generic model interpretability library for PyTorch](https://mlsanity.com/paper/2009.07896/) (2020)
- [Explaining by Removing: A Unified Framework for Model Explanation](https://mlsanity.com/paper/2011.14878/) (2020)
- [Speech Model Pre-training for End-to-End Spoken Language Understanding](https://mlsanity.com/paper/1904.03670/) (2019)
- [The Disagreement Problem in Explainable Machine Learning: A Practitioner's Perspective](https://mlsanity.com/paper/2202.01602/) (2022)
- [The elephant in the interpretability room: Why use attention as explanation when we have saliency methods?](https://mlsanity.com/paper/2010.05607/) (2020)
- [Evaluating Explanations: How much do explanations from the teacher aid students?](https://mlsanity.com/paper/2012.00893/) (2020)
- [When Explanations Lie: Why Many Modified BP Attributions Fail](https://mlsanity.com/paper/1912.09818/) (2019)
- [Impossibility Theorems for Feature Attribution](https://mlsanity.com/paper/2212.11870/) (2022)
- [AudioMNIST: Exploring Explainable Artificial Intelligence for Audio Analysis on a Simple Benchmark](https://mlsanity.com/paper/1807.03418/) (2018)
- ["Will You Find These Shortcuts?" A Protocol for Evaluating the Faithfulness of Input Salience Methods for Text Classification](https://mlsanity.com/paper/2111.07367/) (2021)
- [Exploring the Role of BERT Token Representations to Explain Sentence Probing Results](https://mlsanity.com/paper/2104.01477/) (2021)
- [A Song of (Dis)agreement: Evaluating the Evaluation of Explainable Artificial Intelligence in Natural Language Processing](https://mlsanity.com/paper/2205.04559/) (2022)
- [Can We Trust Explainable AI Methods on ASR? An Evaluation on Phoneme Recognition](https://mlsanity.com/paper/2305.18011/) (2023)
- [Explanations for Automatic Speech Recognition](https://mlsanity.com/paper/2302.14062/) (2023)
- [Explaining Speech Classification Models via Word-Level Audio Segments and Paralinguistic Features](https://mlsanity.com/paper/2309.07733/) (2023)
- [SPES: Spectrogram Perturbation for Explainable Speech-to-Text Generation](https://mlsanity.com/paper/2411.01710/) (2024)
- [Investigating the Effectiveness of Explainability Methods in Parkinson's Detection from Speech](https://mlsanity.com/paper/2411.08013/) (2024)
- [Phoneme Discretized Saliency Maps for Explainable Detection of AI-Generated Voice](https://mlsanity.com/paper/2406.10422/) (2024)

## Built on by (most cited first, 1 of 1)

- [What Do Neural Networks Learn for TDOA Estimation? A Cross-Architecture Probing Study](https://mlsanity.com/paper/2606.22020/) (2026)

## Abstract (arXiv)

As the capabilities of large-scale pre-trained models evolve, understanding the determinants of their outputs becomes more important. Feature attribution aims to reveal which parts of the input elements contribute the most to model outputs. In speech processing, the unique characteristics of the input signal make the application of feature attribution methods challenging. We study how factors such as input type and aggregation and perturbation timespan impact the reliability of standard feature attribution methods, and how these factors interact with characteristics of each classification task. We find that standard approaches to feature attribution are generally unreliable when applied to the speech domain, with the exception of word-aligned perturbation methods when applied to word-based classification tasks.