One Wave To Explain Them All: A Unifying Perspective On Feature Attribution
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908393821700096 |
|---|---|
| author | Kasmi, Gabriel Brunetto, Amandine Fel, Thomas Parekh, Jayneel |
| author_facet | Kasmi, Gabriel Brunetto, Amandine Fel, Thomas Parekh, Jayneel |
| contents | Feature attribution methods aim to improve the transparency of deep neural networks by identifying the input features that influence a model's decision. Pixel-based heatmaps have become the standard for attributing features to high-dimensional inputs, such as images, audio representations, and volumes. While intuitive and convenient, these pixel-based attributions fail to capture the underlying structure of the data. Moreover, the choice of domain for computing attributions has often been overlooked. This work demonstrates that the wavelet domain allows for informative and meaningful attributions. It handles any input dimension and offers a unified approach to feature attribution. Our method, the Wavelet Attribution Method (WAM), leverages the spatial and scale-localized properties of wavelet coefficients to provide explanations that capture both the where and what of a model's decision-making process. We show that WAM quantitatively matches or outperforms existing gradient-based methods across multiple modalities, including audio, images, and volumes. Additionally, we discuss how WAM bridges attribution with broader aspects of model robustness and transparency. Project page: https://gabrielkasmi.github.io/wam/ |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_01482 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | One Wave To Explain Them All: A Unifying Perspective On Feature Attribution Kasmi, Gabriel Brunetto, Amandine Fel, Thomas Parekh, Jayneel Machine Learning Artificial Intelligence Feature attribution methods aim to improve the transparency of deep neural networks by identifying the input features that influence a model's decision. Pixel-based heatmaps have become the standard for attributing features to high-dimensional inputs, such as images, audio representations, and volumes. While intuitive and convenient, these pixel-based attributions fail to capture the underlying structure of the data. Moreover, the choice of domain for computing attributions has often been overlooked. This work demonstrates that the wavelet domain allows for informative and meaningful attributions. It handles any input dimension and offers a unified approach to feature attribution. Our method, the Wavelet Attribution Method (WAM), leverages the spatial and scale-localized properties of wavelet coefficients to provide explanations that capture both the where and what of a model's decision-making process. We show that WAM quantitatively matches or outperforms existing gradient-based methods across multiple modalities, including audio, images, and volumes. Additionally, we discuss how WAM bridges attribution with broader aspects of model robustness and transparency. Project page: https://gabrielkasmi.github.io/wam/ |
| title | One Wave To Explain Them All: A Unifying Perspective On Feature Attribution |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2410.01482 |