The Summary Loop: Learning to Write Abstractive Summaries Without Examples
37510Authors
nan
Abstract
AbstractThis work presents a new approach to unsupervised abstractive summarization based on maximizing a combination of coverage and fluency for a given length constraint. It introduces a novel method that encourages the inclusion of key terms from the original document into the summary: key terms are masked out of the original document and must be filled in by a coverage model using the current generated summary. A novel unsupervised training procedure leverages this coverage model along with a fluency model to generate and score summaries. When tested on popular news summarization datasets, the method outperforms previous unsupervised methods by more than 2 R-1 points, and approaches results of competitive supervised methods. Our model attains higher levels of abstraction with copied passages roughly two times shorter than prior work, and learns to compress and merge sentences without supervision.
Keywords
Full Content
Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 5135–5150 July 5 - 10, 2020. c⃝2020 Association for Computational Linguistics 5135 The Summary Loop: Learning to Write Abstractive Summaries Without Examples Philippe Laban UC Berkeley Andrew Hsi Bloomberg John Canny UC Berkeley Marti A. Hearst UC Berkeley∗ Abstract This work presents a new approach to unsu- pervised abstractive summarization based on maximizing a combination of coverage and flu- ency for a given length constraint. It intro- duces a novel method that encourages the in- clusion of key terms from the original docu- ment into the summary: key terms are masked out of the original document and must be filled in by a coverage model using the current gener- ated summary. A novel unsupervised training procedure leverages this coverage model along with a fluency model to generate and score summaries. When tested on popular news sum- marization datasets, the method outperforms previous unsupervised methods by more than 2 R-1 points, and approaches results of com- petitive supervised methods. Our model at- tains higher levels of abstraction with copied passages roughly two times shorter than prior work, and learns to compress and merge sen- tences without supervision. 1 Introduction Summarization, or the task of condensing a doc- ument’s main points into a shorter document, is important for many text domains, such as headlines for news and abstracts for research papers. This paper presents a novel unsupervised ab- stractive summarization method that generates sum- maries directly from source documents, without the aid of example summaries. This approach simul- taneously optimizes for the following important properties of a good summary: • coverage of the keywords of the document, • fluency of generated language, and • brevity of generated summaries. ∗Author emails: {phillab,canny,hearst}@berkeley.edu, ahsil@bloomberg.net Original Document: Chilean President announced Wednesday that his country, which has been paralyzed by protests over the last two weeks, will no longer host two major international summits. [...] The President has now canceled the hosting of the economic APEC fo- rum and COP25 environmental summit, which were both due to take place later this year. [...] Masked Document: that his country, which has beenannounced Wednesday by over the last two weeks, will no longer two major interna- tional . [...] The has now the of the and , which were both due to take place later this . [...] Summary Loop [10 word constraint]: Pinera can- celled the APEC summit at Santiago. Coverage Score: 0.22 Summary Loop [24 word constraint]: Pinera said Chileans have been canceled the hosting of the APEC summit, which was scheduled to take place in November. Coverage score: 0.33 Summary Loop [45 word constraint]: Sebastian Pin- era announced Wednesday that his country will not hold the APEC summit, which was scheduled to take place in Santiago. Pinera said that Chileans had been paralyzed by protests over the last two weeks. Coverage score: 0.39 Figure 1: Motivating example. A document from CNN.com (keywords generated by masking procedure are bolded), the masked version of …
Content truncated to 500 words
Labeled Text
AbstractThis work presents a new approach to unsupervised abstractive summarization based on maximizing a combination of coverage and fluency for a given length constraint. It introduces a novel method that encourages the inclusion of key terms from the original document into the summary: key terms are masked out of the original document and must be filled in by a coverage model using the current generated summary. A novel unsupervised training procedure leverages this coverage model along with a fluency model to generate and score summaries. When tested on popular news summarization datasets, the method outperforms previous unsupervised methods by more than 2 R-1 points, and approaches results of competitive supervised methods. Our model attains higher levels of abstraction with copied passages roughly two times shorter than prior work, and learns to compress and merge sentences without supervision.
Publication Details
- Paper ID:
- 37510
- Year:
- 2020
- Publication:
- TIPSTER
- Publication Type:
- conf
- DOI:
- nan