Leveraging Linguistic Context in Dyadic Interactions to Improve Automatic Speech Recognition for Children.

Manoj Kumar,So Hyun Kim,Catherine Lord,Thomas D Lyon,Shrikanth Narayanan

Computer Speech & Language（2020）

引用 13|浏览52

暂无评分

摘要

•Participating adult’s speech during child-adult dyadic conversation helps improving child speech recognition.•Long-term semantic context from adult speech is captured by semantic response generation using seq2seq model.•Context-adapted language models result in 10.71% WER (absolute) and 60.55% perplexity (relative) improvement.

查看译文

关键词

Child speech,Automatic speech recognition,Autism spectrum disorder,Forensic interviews

AI 理解论文

溯源树

样例

生成溯源树，研究论文发展脉络

Chat Paper

正在生成论文摘要