Semantic Similarity for Detecting Recognition Errors in Automatic Speech Transcripts

12827

Knowledge Triples