Improving End-to-end Speech Recognition with Pronunciation-assisted Sub-word Modeling [article]

Hainan Xu, Shuoyang Ding, Shinji Watanabe
2019 arXiv   pre-print
Most end-to-end speech recognition systems model text directly as a sequence of characters or sub-words. Current approaches to sub-word extraction only consider character sequence frequencies, which at times produce inferior sub-word segmentation that might lead to erroneous speech recognition output. We propose pronunciation-assisted sub-word modeling (PASM), a sub-word extraction method that leverages the pronunciation information of a word. Experiments show that the proposed method can
more » ... y improve upon the character-based baseline, and also outperform commonly used byte-pair encoding methods.
arXiv:1811.04284v2 fatcat:s4s5lxwmcrfzxm5vt65u6ggg7a