{"title":"提取与抽象:在单一编码器-解码器框架内统一提取与抽象摘要法","authors":"Yuping Wu, Hao Li, Hongbo Zhu, Goran Nenadic, Xiao-Jun Zeng","doi":"arxiv-2409.11827","DOIUrl":null,"url":null,"abstract":"Extract-then-Abstract is a naturally coherent paradigm to conduct abstractive\nsummarization with the help of salient information identified by the extractive\nmodel. Previous works that adopt this paradigm train the extractor and\nabstractor separately and introduce extra parameters to highlight the extracted\nsalients to the abstractor, which results in error accumulation and additional\ntraining costs. In this paper, we first introduce a parameter-free highlight\nmethod into the encoder-decoder framework: replacing the encoder attention mask\nwith a saliency mask in the cross-attention module to force the decoder to\nfocus only on salient parts of the input. A preliminary analysis compares\ndifferent highlight methods, demonstrating the effectiveness of our saliency\nmask. We further propose the novel extract-and-abstract paradigm, ExtAbs, which\njointly and seamlessly performs Extractive and Abstractive summarization tasks\nwithin single encoder-decoder model to reduce error accumulation. In ExtAbs,\nthe vanilla encoder is augmented to extract salients, and the vanilla decoder\nis modified with the proposed saliency mask to generate summaries. Built upon\nBART and PEGASUS, experiments on three datasets show that ExtAbs can achieve\nsuperior performance than baselines on the extractive task and performs\ncomparable, or even better than the vanilla models on the abstractive task.","PeriodicalId":501030,"journal":{"name":"arXiv - CS - Computation and Language","volume":"50 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Extract-and-Abstract: Unifying Extractive and Abstractive Summarization within Single Encoder-Decoder Framework\",\"authors\":\"Yuping Wu, Hao Li, Hongbo Zhu, Goran Nenadic, Xiao-Jun Zeng\",\"doi\":\"arxiv-2409.11827\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Extract-then-Abstract is a naturally coherent paradigm to conduct abstractive\\nsummarization with the help of salient information identified by the extractive\\nmodel. Previous works that adopt this paradigm train the extractor and\\nabstractor separately and introduce extra parameters to highlight the extracted\\nsalients to the abstractor, which results in error accumulation and additional\\ntraining costs. In this paper, we first introduce a parameter-free highlight\\nmethod into the encoder-decoder framework: replacing the encoder attention mask\\nwith a saliency mask in the cross-attention module to force the decoder to\\nfocus only on salient parts of the input. A preliminary analysis compares\\ndifferent highlight methods, demonstrating the effectiveness of our saliency\\nmask. We further propose the novel extract-and-abstract paradigm, ExtAbs, which\\njointly and seamlessly performs Extractive and Abstractive summarization tasks\\nwithin single encoder-decoder model to reduce error accumulation. In ExtAbs,\\nthe vanilla encoder is augmented to extract salients, and the vanilla decoder\\nis modified with the proposed saliency mask to generate summaries. Built upon\\nBART and PEGASUS, experiments on three datasets show that ExtAbs can achieve\\nsuperior performance than baselines on the extractive task and performs\\ncomparable, or even better than the vanilla models on the abstractive task.\",\"PeriodicalId\":501030,\"journal\":{\"name\":\"arXiv - CS - Computation and Language\",\"volume\":\"50 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-18\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - CS - Computation and Language\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2409.11827\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - CS - Computation and Language","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2409.11827","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Extract-and-Abstract: Unifying Extractive and Abstractive Summarization within Single Encoder-Decoder Framework
Extract-then-Abstract is a naturally coherent paradigm to conduct abstractive
summarization with the help of salient information identified by the extractive
model. Previous works that adopt this paradigm train the extractor and
abstractor separately and introduce extra parameters to highlight the extracted
salients to the abstractor, which results in error accumulation and additional
training costs. In this paper, we first introduce a parameter-free highlight
method into the encoder-decoder framework: replacing the encoder attention mask
with a saliency mask in the cross-attention module to force the decoder to
focus only on salient parts of the input. A preliminary analysis compares
different highlight methods, demonstrating the effectiveness of our saliency
mask. We further propose the novel extract-and-abstract paradigm, ExtAbs, which
jointly and seamlessly performs Extractive and Abstractive summarization tasks
within single encoder-decoder model to reduce error accumulation. In ExtAbs,
the vanilla encoder is augmented to extract salients, and the vanilla decoder
is modified with the proposed saliency mask to generate summaries. Built upon
BART and PEGASUS, experiments on three datasets show that ExtAbs can achieve
superior performance than baselines on the extractive task and performs
comparable, or even better than the vanilla models on the abstractive task.