MNASR: A Free Speech Corpus For Mongolian Speech Recognition And Accompanied Baselines

MNASR: A Free Speech Corpus For Mongolian Speech Recognition And Accompanied Baselines

연구 분야: Artificial Intelligence

학회: 2022 25th Conference of the Oriental COCOSDA International Committee for the Co-ordination and Standardisation of Speech Databases and Assessment Techniques (O-COCOSDA)

Thanks to the development of deep learning and the emergence of open source data sets, automatic speech recognition (ASR) has made great strides in mainstream languages such as Chinese and English. However, the research of ASR in Mongolian and other minority languages lags far behind the mainstream, due to low attention and limited open source data sets. To promote the development of new models and new methods for Mongolian ASR, this paper releases the MnASR database which contains 345 hours of Mongolian speech signal and the corresponding transcription. MnASR is the largest publicly available and free Mongolian speech database so far. Speech recognition baselines are made public at the same time. Both the database and the accompanied baselines are free for research purpose.