SpidR-Adapt: A Universal Speech Representation Model for Few-Shot Adaptation

TL;DR

SpidR-Adapt is a meta-learning based speech representation model that enables rapid, data-efficient adaptation to new languages, significantly outperforming traditional methods in low-resource scenarios.

Contribution

The paper introduces a novel meta-learning framework with a scalable bi-level optimization approach for fast speech unit adaptation in low-resource settings.

Findings

Achieves rapid phonemic discriminability improvements with less than 1 hour of data.

Surpasses in-domain toplines in downstream language modeling tasks.

Provides 100x greater data efficiency than standard multi-task training.

Abstract

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.