{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/query-focused-abstractive-summarization","title":"Query Focused Abstractive Summarization: Incorporating Query Relevance, Multi-Document Coverage, and Summary Length Constraints into seq2seq Models","arxiv_id":"1801.07704","date":"2018-01-23","proceeding":null,"authors":["Tal Baumel","Matan Eyal","Michael Elhadad"],"abstract":"Query Focused Summarization (QFS) has been addressed mostly using extractive\nmethods. Such methods, however, produce text which suffers from low coherence.\nWe investigate how abstractive methods can be applied to QFS, to overcome such\nlimitations. Recent developments in neural-attention based sequence-to-sequence\nmodels have led to state-of-the-art results on the task of abstractive generic\nsingle document summarization. Such models are trained in an end to end method\non large amounts of training data. We address three aspects to make abstractive\nsummarization applicable to QFS: (a)since there is no training data, we\nincorporate query relevance into a pre-trained abstractive model; (b) since\nexisting abstractive models are trained in a single-document setting, we design\nan iterated method to embed abstractive models within the multi-document\nrequirement of QFS; (c) the abstractive models we adapt are trained to generate\ntext of specific length (about 100 words), while we aim at generating output of\na different size (about 250 words); we design a way to adapt the target size of\nthe generated summaries to a given size ratio. We compare our method (Relevance\nSensitive Attention for QFS) to extractive baselines and with various ways to\ncombine abstractive models on the DUC QFS datasets and demonstrate solid\nimprovements on ROUGE performance.","url_abs":"http://arxiv.org/abs/1801.07704v2","url_pdf":"http://arxiv.org/pdf/1801.07704v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[],"tasks":[{"task_slug":"abstractive-text-summarization","task_name":"Abstractive Text Summarization"},{"task_slug":"document-summarization","task_name":"Document Summarization"},{"task_slug":"query-based-extractive-summarization","task_name":"Query-Based Extractive Summarization"},{"task_slug":"query-focused-summarization","task_name":"Query-focused Summarization"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/query-based-extractive-summarization-on","task":"Query-Based Extractive Summarization","dataset":"Debatepedia","model":"RSA Word Count","rank_in_archive_order":1,"of":2,"metrics":{"ROUGE-1":"53.09"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1801.07704","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}