Demonstration of Adapt4Me: An Uncertainty-Aware Authoring Environment for Personalizing Automatic Speech Recognition to Non-normative Speech
This addresses the problem of labor-intensive data collection and technical complexity in ASR personalization for non-normative speakers, representing a novel approach but with incremental technical components.
The paper tackled the challenge of personalizing automatic speech recognition for non-normative speech by proposing Adapt4Me, a web-based environment that uses Bayesian active learning to enable end-to-end personalization without expert supervision, resulting in users being able to personalize robust ASR models effectively.
Personalizing Automatic Speech Recognition (ASR) for non-normative speech remains challenging because data collection is labor-intensive and model training is technically complex. To address these limitations, we propose Adapt4Me, a web-based decentralized environment that operationalizes Bayesian active learning to enable end-to-end personalization without expert supervision. The app exposes data selection, adaptation, and validation to lay users through a three-stage human-in-the-loop workflow: (1) rapid profiling via greedy phoneme sampling to capture speaker-specific acoustics; (2) backend personalization using Variational Inference Low-Rank Adaptation (VI-LoRA) to enable fast, incremental updates; and (3) continuous improvement, where users guide model refinement by resolving visualized model uncertainty via low-friction top-k corrections. By making epistemic uncertainty explicit, Adapt4Me reframes data efficiency as an interactive design feature rather than a purely algorithmic concern. We show how this enables users to personalize robust ASR models, transforming them from passive data sources into active authors of their own assistive technology.