@inproceedings{chi-etal-2021-align,
    title = "Align-Refine: Non-Autoregressive Speech Recognition via Iterative Realignment",
    author = "Chi, Ethan A.  and
      Salazar, Julian  and
      Kirchhoff, Katrin",
    editor = "Toutanova, Kristina  and
      Rumshisky, Anna  and
      Zettlemoyer, Luke  and
      Hakkani-Tur, Dilek  and
      Beltagy, Iz  and
      Bethard, Steven  and
      Cotterell, Ryan  and
      Chakraborty, Tanmoy  and
      Zhou, Yichao",
    booktitle = "Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies",
    month = jun,
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/add-emnlp-2024-awards/2021.naacl-main.154/",
    doi = "10.18653/v1/2021.naacl-main.154",
    pages = "1920--1927",
    abstract = "Non-autoregressive encoder-decoder models greatly improve decoding speed over autoregressive models, at the expense of generation quality. To mitigate this, iterative decoding models repeatedly infill or refine the proposal of a non-autoregressive model. However, editing at the level of output sequences limits model flexibility. We instead propose *iterative realignment*, which by refining latent alignments allows more flexible edits in fewer steps. Our model, Align-Refine, is an end-to-end Transformer which iteratively realigns connectionist temporal classification (CTC) alignments. On the WSJ dataset, Align-Refine matches an autoregressive baseline with a 14x decoding speedup; on LibriSpeech, we reach an LM-free test-other WER of 9.0{\%} (19{\%} relative improvement on comparable work) in three iterations. We release our code at \url{https://github.com/amazon-research/align-refine}."
}