@inproceedings{huang-etal-2024-decoding,
    title = "A Decoding Algorithm for Length-Control Summarization Based on Directed Acyclic Transformers",
    author = "Huang, Chenyang  and
      Zhou, Hao  and
      Jen, Cameron  and
      Zheng, Kangjie  and
      Zaiane, Osmar  and
      Mou, Lili",
    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/fix-sig-urls/2024.findings-emnlp.677/",
    doi = "10.18653/v1/2024.findings-emnlp.677",
    pages = "11572--11583",
    abstract = "Length-control summarization aims to condense long texts into a short one within a certain length limit. Previous approaches often use autoregressive (AR) models and treat the length requirement as a soft constraint, which may not always be satisfied. In this study, we propose a novel length-control decoding algorithm based on the directed acyclic Transformer (DAT). Our approach allows for multiple plausible sequence fragments and predicts a \textit{path} to connect them. In addition, we propose a Sequence Maximum a Posteriori (Seq-MAP) decoding algorithm that marginalizes different possible paths and finds the most probable summary satisfying the length budget. Our algorithm is based on beam search, which further facilitates a reranker for performance improvement. Experimental results on the Gigaword dataset demonstrate our state-of-the-art performance for length-control summarization."
}