Two-Stage Adaptation for Non-Normative Speech Recognition: Revisiting Speaker-Independent Initialization for Personalization
This addresses the challenge of adapting ASR systems for individuals with speech disorders, though it appears incremental as it builds on existing fine-tuning methods.
The paper tackles personalizing automatic speech recognition for non-normative speech (e.g., dysarthric, aphasic) by proposing a two-stage adaptation framework that first fine-tunes on multi-speaker non-normative data and then on individual speakers, showing it consistently improves personalization over direct speaker-specific fine-tuning while maintaining manageable out-of-domain trade-offs.
Personalizing automatic speech recognition (ASR) systems for non-normative speech, such as dysarthric and aphasic speech, is challenging. While speaker-specific fine-tuning (SS-FT) is widely used, it is typically initialized directly from a generic pre-trained model. Whether speaker-independent adaptation provides a stronger initialization prior under such mismatch remains unclear. In this work, we propose a two-stage adaptation framework consisting of speaker-independent fine-tuning (SI-FT) on multi-speaker non-normative data followed by SS-FT, and evaluate it through a controlled comparison with direct SS-FT under identical per-speaker conditions. Experiments on AphasiaBank and UA-Speech with Whisper-Large-v3 and Qwen3-ASR, alongside evaluation on typical-speech datasets TED-LIUM v3 and FLEURS, show that two-stage adaptation consistently improves personalization while maintaining manageable out-of-domain (OOD) trade-offs.