@inproceedings{liu-etal-2024-enable,
    title = "Enable Fast Sampling for {S}eq2{S}eq Text Diffusion",
    author = "Liu, Pan  and
      Tian, Xiaohua  and
      Lin, Zhouhan",
    editor = "Al-Onaizan, Yaser  and
      Bansal, Mohit  and
      Chen, Yun-Nung",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2024",
    month = nov,
    year = "2024",
    address = "Miami, Florida, USA",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/jlcl-multiple-ingestion/2024.findings-emnlp.497/",
    doi = "10.18653/v1/2024.findings-emnlp.497",
    pages = "8495--8505",
    abstract = "Diffusion models exhibit promising capacity for generating high-quality text. However, owing to the curved nature of generation path, they necessitate traversing numerous steps to guarantee the text quality. In this paper, we propose an efficient model FMSeq, which utilizes flow matching to straighten the generation path, thereby enabling fast sampling for diffusion-based seq2seq text generation. Specifically, we construct transport flow only on the target sequences to adapt the diffusion-based model with flow matching. Furthermore, we explore different settings and identify target-parameterization, self-conditioning and time-difference as three effective techniques to improve the generation quality under a few steps. Experiments on four popular tasks demonstrate that FMSeq generates texts of comparable quality to the SOTA diffusion-based DiffuSeq in just 10 steps, achieving a 200-fold speedup."
}