Aphasic Speech Recognition using a Mixture of Speech Intelligibility Experts
This work addresses the challenge of robust speech recognition for aphasic speech analysis, which is crucial for clinical applications, but it is incremental as it builds on existing mixture of experts methods.
The paper tackles the problem of poor performance of standard automatic speech recognition models on aphasic speech by proposing a mixture of experts acoustic model that incorporates severity-based experts and a speech intelligibility detector, resulting in significant reductions in phone error rates across all severity stages.
Robust speech recognition is a key prerequisite for semantic feature extraction in automatic aphasic speech analysis. However, standard one-size-fits-all automatic speech recognition models perform poorly when applied to aphasic speech. One reason for this is the wide range of speech intelligibility due to different levels of severity (i.e., higher severity lends itself to less intelligible speech). To address this, we propose a novel acoustic model based on a mixture of experts (MoE), which handles the varying intelligibility stages present in aphasic speech by explicitly defining severity-based experts. At test time, the contribution of each expert is decided by estimating speech intelligibility with a speech intelligibility detector (SID). We show that our proposed approach significantly reduces phone error rates across all severity stages in aphasic speech compared to a baseline approach that does not incorporate severity information into the modeling process.