Hypernetworks for Personalizing ASR to Atypical Speech

TL;DR

This paper introduces a hypernetwork-based approach for personalized ASR that efficiently adapts to atypical speech without prior knowledge of speech disorders, achieving significant error reduction and better generalization.

Contribution

It presents a novel hypernetwork method for on-the-fly, utterance-level adaptation in ASR, reducing the need for disorder-specific models and minimal parameter tuning.

Findings

Halves Word Error Rate with only 0.03% of weights.

Achieves 75.2% relative WER reduction with 0.1% of parameters.

Generalizes better to out-of-distribution speakers.

Abstract

Parameter-efficient fine-tuning (PEFT) for personalizing automatic speech recognition (ASR) has recently shown promise for adapting general population models to atypical speech. However, these approaches assume a priori knowledge of the atypical speech disorder being adapted for -- the diagnosis of which requires expert knowledge that is not always available. Even given this knowledge, data scarcity and high inter/intra-speaker variability further limit the effectiveness of traditional fine-tuning. To circumvent these challenges, we first identify the minimal set of model parameters required for ASR adaptation. Our analysis of each individual parameter's effect on adaptation performance allows us to reduce Word Error Rate (WER) by half while adapting 0.03% of all weights. Alleviating the need for cohort-specific models, we next propose the novel use of a meta-learned hypernetwork to generate highly individualized, utterance-level adaptations on-the-fly for a diverse set of atypical speech characteristics. Evaluating adaptation at the global, cohort and individual-level, we show that hypernetworks generalize better to out-of-distribution speakers, while maintaining an overall relative WER reduction of 75.2% using 0.1% of the full parameter budget.