SPES: Spectrogram Perturbation for Explainable Speech-to-Text Generation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866929760957890560 |
|---|---|
| author | Fucci, Dennis Gaido, Marco Savoldi, Beatrice Negri, Matteo Cettolo, Mauro Bentivogli, Luisa |
| author_facet | Fucci, Dennis Gaido, Marco Savoldi, Beatrice Negri, Matteo Cettolo, Mauro Bentivogli, Luisa |
| contents | Spurred by the demand for interpretable models, research on eXplainable AI for language technologies has experienced significant growth, with feature attribution methods emerging as a cornerstone of this progress. While prior work in NLP explored such methods for classification tasks and textual applications, explainability intersecting generation and speech is lagging, with existing techniques failing to account for the autoregressive nature of state-of-the-art models and to provide fine-grained, phonetically meaningful explanations. We address this gap by introducing Spectrogram Perturbation for Explainable Speech-to-text Generation (SPES), a feature attribution technique applicable to sequence generation tasks with autoregressive models. SPES provides explanations for each predicted token based on both the input spectrogram and the previously generated tokens. Extensive evaluation on speech recognition and translation demonstrates that SPES generates explanations that are faithful and plausible to humans. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2411_01710 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | SPES: Spectrogram Perturbation for Explainable Speech-to-Text Generation Fucci, Dennis Gaido, Marco Savoldi, Beatrice Negri, Matteo Cettolo, Mauro Bentivogli, Luisa Computation and Language Sound Audio and Speech Processing Spurred by the demand for interpretable models, research on eXplainable AI for language technologies has experienced significant growth, with feature attribution methods emerging as a cornerstone of this progress. While prior work in NLP explored such methods for classification tasks and textual applications, explainability intersecting generation and speech is lagging, with existing techniques failing to account for the autoregressive nature of state-of-the-art models and to provide fine-grained, phonetically meaningful explanations. We address this gap by introducing Spectrogram Perturbation for Explainable Speech-to-text Generation (SPES), a feature attribution technique applicable to sequence generation tasks with autoregressive models. SPES provides explanations for each predicted token based on both the input spectrogram and the previously generated tokens. Extensive evaluation on speech recognition and translation demonstrates that SPES generates explanations that are faithful and plausible to humans. |
| title | SPES: Spectrogram Perturbation for Explainable Speech-to-Text Generation |
| topic | Computation and Language Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2411.01710 |