Researchers enhance elderly ASR with LLM paraphrasing and speech synthesis

Researchers have developed a novel data augmentation technique to improve automatic speech recognition (ASR) for elderly individuals. This method utilizes large language models to paraphrase existing transcripts, generating elderly-contextual variations. These paraphrased texts are then converted into synthetic speech using text-to-speech synthesis with elderly reference speakers. Experiments demonstrated a significant reduction in word error rate, with up to a 58.2% improvement compared to baseline models. AI

IMPACT Enhances ASR performance for specific demographics, potentially improving accessibility of voice technologies for the elderly.

RANK_REASON Academic paper detailing a new method for data augmentation in ASR.

Read on arXiv cs.CL →

• LLM
• Whisper

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →