Fine-tuning speech AI on a single user's voice dramatically improves recognition of impaired speech

Original: Adapting Foundation ASR Models to Dysarthric Speech: A Case Study

Source: arxiv.org ↗