Drax: Speech Recognition with Discrete Flow Matching
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866918154670702592 |
|---|---|
| author | Navon, Aviv Shamsian, Aviv Glazer, Neta Segal-Feldman, Yael Hetz, Gill Keshet, Joseph Fetaya, Ethan |
| author_facet | Navon, Aviv Shamsian, Aviv Glazer, Neta Segal-Feldman, Yael Hetz, Gill Keshet, Joseph Fetaya, Ethan |
| contents | Diffusion and flow-based non-autoregressive (NAR) models have shown strong promise in large language modeling, however, their potential for automatic speech recognition (ASR) remains largely unexplored. We propose Drax, a discrete flow matching framework for ASR that enables efficient parallel decoding. To better align training with inference, we construct an audio-conditioned probability path that guides the model through trajectories resembling likely intermediate inference errors, rather than direct random noise to target transitions. Our theoretical analysis links the generalization gap to divergences between training and inference occupancies, controlled by cumulative velocity errors, thereby motivating our design choice. Empirical evaluation demonstrates that our approach attains recognition accuracy on par with state-of-the-art speech models while offering improved accuracy-efficiency trade-offs, highlighting discrete flow matching as a promising direction for advancing NAR ASR. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_04162 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Drax: Speech Recognition with Discrete Flow Matching Navon, Aviv Shamsian, Aviv Glazer, Neta Segal-Feldman, Yael Hetz, Gill Keshet, Joseph Fetaya, Ethan Audio and Speech Processing Machine Learning Sound Diffusion and flow-based non-autoregressive (NAR) models have shown strong promise in large language modeling, however, their potential for automatic speech recognition (ASR) remains largely unexplored. We propose Drax, a discrete flow matching framework for ASR that enables efficient parallel decoding. To better align training with inference, we construct an audio-conditioned probability path that guides the model through trajectories resembling likely intermediate inference errors, rather than direct random noise to target transitions. Our theoretical analysis links the generalization gap to divergences between training and inference occupancies, controlled by cumulative velocity errors, thereby motivating our design choice. Empirical evaluation demonstrates that our approach attains recognition accuracy on par with state-of-the-art speech models while offering improved accuracy-efficiency trade-offs, highlighting discrete flow matching as a promising direction for advancing NAR ASR. |
| title | Drax: Speech Recognition with Discrete Flow Matching |
| topic | Audio and Speech Processing Machine Learning Sound |
| url | https://arxiv.org/abs/2510.04162 |