A Decoding Algorithm for Length-Control Summarization Based on Directed Acyclic Transformers
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 \emph{path} to connect them. In addition, we propose a Sequence Maximum a Posteriori (SeqMAP) 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 and DUC2004 datasets demonstrate our state-of-the-art performance for length-control summarization.
Cite
@article{arxiv.2502.04535,
title = {A Decoding Algorithm for Length-Control Summarization Based on Directed Acyclic Transformers},
author = {Chenyang Huang and Hao Zhou and Cameron Jen and Kangjie Zheng and Osmar R. Zaïane and Lili Mou},
journal= {arXiv preprint arXiv:2502.04535},
year = {2025}
}
Comments
Findings of the Association for Computational Linguistics: EMNLP 2024