SpidR-Adapt: A Universal Speech Representation Model for Few-Shot Adaptation
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908979327664128 |
|---|---|
| author | Luthra, Mahi Shen, Jiayi Poli, Maxime Ortiz, Angelo Higuchi, Yosuke Benchekroun, Youssef Gleize, Martin Saint-James, Charles-Eric Lin, Dongyan Rust, Phillip Villar, Angel Parimi, Surya Stark, Vanessa Moritz, Rashel Pino, Juan LeCun, Yann Dupoux, Emmanuel |
| author_facet | Luthra, Mahi Shen, Jiayi Poli, Maxime Ortiz, Angelo Higuchi, Yosuke Benchekroun, Youssef Gleize, Martin Saint-James, Charles-Eric Lin, Dongyan Rust, Phillip Villar, Angel Parimi, Surya Stark, Vanessa Moritz, Rashel Pino, Juan LeCun, Yann Dupoux, Emmanuel |
| contents | Human infants, with only a few hundred hours of speech exposure, acquire basic units of new languages, highlighting a striking efficiency gap compared to the data-hungry self-supervised speech models. To address this gap, this paper introduces SpidR-Adapt for rapid adaptation of speech units to new languages using minimal unlabeled data. We cast such low-resource speech representation learning as a meta-learning problem and construct a multi-task adaptive pre-training (MAdaPT) protocol which formulates the adaptation process as a bi-level optimization framework. To enable scalable meta-training under this framework, we propose a novel heuristic solution, first-order bi-level optimization (FOBLO), avoiding heavy computation costs. Finally, we stabilize meta-training by using a robust initialization through interleaved supervision which alternates self-supervised and supervised objectives. Empirically, SpidR-Adapt achieves rapid gains in phonemic discriminability (ABX) and downstream spoken language modeling scores (sWUGGY, sBLIMP, tSC), surpassing in-domain toplines after training on less than 1h of target-language audio and delivering $100\times$ greater data efficiency than standard multi-task training. These findings highlight a practical, architecture-agnostic path toward biologically inspired, data-efficient representations. We open-source the training code and model checkpoints at https://github.com/facebookresearch/spidr-adapt. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_21204 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | SpidR-Adapt: A Universal Speech Representation Model for Few-Shot Adaptation Luthra, Mahi Shen, Jiayi Poli, Maxime Ortiz, Angelo Higuchi, Yosuke Benchekroun, Youssef Gleize, Martin Saint-James, Charles-Eric Lin, Dongyan Rust, Phillip Villar, Angel Parimi, Surya Stark, Vanessa Moritz, Rashel Pino, Juan LeCun, Yann Dupoux, Emmanuel Computation and Language Artificial Intelligence Human infants, with only a few hundred hours of speech exposure, acquire basic units of new languages, highlighting a striking efficiency gap compared to the data-hungry self-supervised speech models. To address this gap, this paper introduces SpidR-Adapt for rapid adaptation of speech units to new languages using minimal unlabeled data. We cast such low-resource speech representation learning as a meta-learning problem and construct a multi-task adaptive pre-training (MAdaPT) protocol which formulates the adaptation process as a bi-level optimization framework. To enable scalable meta-training under this framework, we propose a novel heuristic solution, first-order bi-level optimization (FOBLO), avoiding heavy computation costs. Finally, we stabilize meta-training by using a robust initialization through interleaved supervision which alternates self-supervised and supervised objectives. Empirically, SpidR-Adapt achieves rapid gains in phonemic discriminability (ABX) and downstream spoken language modeling scores (sWUGGY, sBLIMP, tSC), surpassing in-domain toplines after training on less than 1h of target-language audio and delivering $100\times$ greater data efficiency than standard multi-task training. These findings highlight a practical, architecture-agnostic path toward biologically inspired, data-efficient representations. We open-source the training code and model checkpoints at https://github.com/facebookresearch/spidr-adapt. |
| title | SpidR-Adapt: A Universal Speech Representation Model for Few-Shot Adaptation |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2512.21204 |