Leveraging supplementary text data to kick-start automatic speech recognition system development with limited transcriptions

Open in new window