This project, Pronunciation-assisted Subword Modeling, generates linguistically
meaningful subwords by analyzing a corpus and a dictionary. See ./run.sh for a
complete example. The method is described here,

@inproceedings{xu2019improving,
    title={Improving End-to-end Speech Recognition with Pronunciation-assisted Sub-word Modeling},
    author={Xu, Hainan and Ding, Shuoyang and Watanabe, Shinji},
    booktitle={ICASSP 2019-2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
    pages={7110--7114},
    year={2019},
    organization={IEEE}
}