Quantification of stylistic differences in human- and ASR-produced transcripts of African American English
Domain:
natural language processingRecord type:
paperdataset
Creator:
HeuKendelMcN
Host:
Common measures of accuracy used to assess the performance of automatic speech recognition (ASR) systems, as well as human transcribers, conflate multiple sources of error. Stylistic differences, such as verbatim vs non-verbatim, can play a significant role in ASR performance evaluation when differences exist between training and test datasets. The problem is compounded for speech from underrepresented varieties, where the speech to orthography mapping is not as standardized. We categorize the kinds of stylistic differences between 6 transcription versions, 4 human- and 2 ASR-produced, of 10 hours of African American English (AAE) speech. Focusing on verbatim features and AAE morphosyntactic features, we investigate the interactions of these categories with how well transcripts can be compared via word error rate (WER). The results, and overall analysis, help clarify how ASR outputs are a function of the decisions made by the training data's human transcribers. Published in Interspeech 2024 Proceedings, 5 pages excluding references, 5 figures
Similar
Some Similarities and Differences Between African American English and Southern White English in ChildrenAfrican American Evaluations of Black English and Standard American Englishghananlpcommunity/ghana-asr-transcriptsDifferences in patterns of symptom attribution in diagnosing schizophrenia between African American and non-African American clinicians.Dialect Awareness and Lexical Comprehension of Mainstream American English in African American English–Speaking ChildrenToward Responsible ASR for African American English Speakers: A Scoping Review of Bias and Equity in Speech Technology