{"url":"/sota/extractive-document-summarization-on-cnn","task":{"name":"Extractive Text Summarization","url":"/task/extractive-document-summarization","note":null},"dataset":{"name":"CNN / Daily Mail","url":"/dataset/cnn-daily-mail-1"},"category":"Natural Language Processing","categories":["Natural Language Processing"],"category_note":null,"description":"Given a document, selecting a subset of the words or sentences which best represents a summary of the document.","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["ROUGE-1","ROUGE-2","ROUGE-L"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"ROUGE-1":"higher","ROUGE-2":"higher","ROUGE-L":"higher"}},"counts":{"rows":15,"rows_with_code":9,"rows_with_paper_page":14,"rows_dated":12,"rows_using_additional_data":0},"rows":[{"rank_in_archive_order":1,"model":"HAHSum","metrics":{"ROUGE-1":"44.68","ROUGE-2":"21.30","ROUGE-L":"40.75"},"uses_additional_data":false,"paper_date":null,"paper":"/paper/neural-extractive-summarization-with","paper_url":"https://aclanthology.org/2020.emnlp-main.295","paper_title":"Neural Extractive Summarization with Hierarchical Attentive Heterogeneous Graph Network","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":2,"model":"Scaled-MatchSum","metrics":{"ROUGE-1":"44.51","ROUGE-2":"20.07","ROUGE-L":"40.287"},"uses_additional_data":false,"paper_date":null,"paper":null,"paper_url":null,"paper_title":"","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":3,"model":"MatchSum","metrics":{"ROUGE-1":"44.41","ROUGE-2":"20.86","ROUGE-L":"40.55"},"uses_additional_data":false,"paper_date":"2020-04-19","paper":"/paper/extractive-summarization-as-text-matching","paper_url":"https://arxiv.org/abs/2004.08795v1","paper_title":"Extractive Summarization as Text Matching","code":"https://github.com/maszhongming/MatchSum","n_code_links":2,"syntology":null},{"rank_in_archive_order":4,"model":"A2Summ","metrics":{"ROUGE-1":"44.11","ROUGE-2":"20.31","ROUGE-L":"35.92"},"uses_additional_data":false,"paper_date":"2023-03-13","paper":"/paper/align-and-attend-multimodal-summarization","paper_url":"https://arxiv.org/abs/2303.07284v3","paper_title":"Align and Attend: Multimodal Summarization with Dual Contrastive Losses","code":"https://github.com/boheumd/A2Summ","n_code_links":2,"syntology":null},{"rank_in_archive_order":5,"model":"NeRoBERTa","metrics":{"ROUGE-1":"43.86","ROUGE-2":"20.64","ROUGE-L":"40.20"},"uses_additional_data":false,"paper_date":null,"paper":"/paper/considering-nested-tree-structure-in-sentence","paper_url":"https://aclanthology.org/2021.emnlp-main.330","paper_title":"Considering Nested Tree Structure in Sentence Extractive Summarization with Pre-trained Transformer","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":6,"model":"BERT-ext + RL","metrics":{"ROUGE-1":"42.76","ROUGE-2":"19.87","ROUGE-L":"39.11"},"uses_additional_data":false,"paper_date":"2019-09-19","paper":"/paper/summary-level-training-of-sentence-rewriting","paper_url":"https://arxiv.org/abs/1909.08752v3","paper_title":"Summary Level Training of Sentence Rewriting for Abstractive Summarization","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":7,"model":"PNBERT","metrics":{"ROUGE-1":"42.69","ROUGE-2":"19.60","ROUGE-L":"38.85"},"uses_additional_data":false,"paper_date":"2019-07-08","paper":"/paper/searching-for-effective-neural-extractive","paper_url":"https://arxiv.org/abs/1907.03491v1","paper_title":"Searching for Effective Neural Extractive Summarization: What Works and What's Next","code":"https://github.com/maszhongming/Effective_Extractive_Summarization","n_code_links":2,"syntology":null},{"rank_in_archive_order":8,"model":"HIBERT","metrics":{"ROUGE-1":"42.37","ROUGE-2":"19.95","ROUGE-L":"38.83"},"uses_additional_data":false,"paper_date":"2019-05-16","paper":"/paper/hibert-document-level-pre-training-of","paper_url":"https://arxiv.org/abs/1905.06566v1","paper_title":"HIBERT: Document Level Pre-training of Hierarchical Bidirectional Transformers for Document Summarization","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":9,"model":"HER","metrics":{"ROUGE-1":"42.3","ROUGE-2":"18.9","ROUGE-L":"37.9"},"uses_additional_data":false,"paper_date":"2019-11-01","paper":"/paper/reading-like-her-human-reading-inspired","paper_url":"https://aclanthology.org/D19-1300","paper_title":"Reading Like HER: Human Reading Inspired Extractive Summarization","code":"https://github.com/LLluoling/HER","n_code_links":1,"syntology":null},{"rank_in_archive_order":10,"model":"NeuSUM","metrics":{"ROUGE-1":"41.59","ROUGE-2":"19.01","ROUGE-L":"37.98"},"uses_additional_data":false,"paper_date":"2018-07-06","paper":"/paper/neural-document-summarization-by-jointly","paper_url":"http://arxiv.org/abs/1807.02305v1","paper_title":"Neural Document Summarization by Jointly Learning to Score and Select Sentences","code":"https://github.com/magic282/NeuSum","n_code_links":1,"syntology":null},{"rank_in_archive_order":11,"model":"BanditSum","metrics":{"ROUGE-1":"41.5","ROUGE-2":"18.7","ROUGE-L":"37.6"},"uses_additional_data":false,"paper_date":"2018-09-25","paper":"/paper/banditsum-extractive-summarization-as-a","paper_url":"https://arxiv.org/abs/1809.09672v3","paper_title":"BanditSum: Extractive Summarization as a Contextual Bandit","code":"https://github.com/yuedongP/BanditSum","n_code_links":1,"syntology":null},{"rank_in_archive_order":12,"model":"Latent","metrics":{"ROUGE-1":"41.05","ROUGE-2":"18.77","ROUGE-L":"37.54"},"uses_additional_data":false,"paper_date":"2018-08-22","paper":"/paper/neural-latent-extractive-document","paper_url":"http://arxiv.org/abs/1808.07187v2","paper_title":"Neural Latent Extractive Document Summarization","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":13,"model":"Lead-3 baseline","metrics":{"ROUGE-1":"40.34","ROUGE-2":"17.70","ROUGE-L":"36.57"},"uses_additional_data":false,"paper_date":"2017-04-14","paper":"/paper/get-to-the-point-summarization-with-pointer","paper_url":"http://arxiv.org/abs/1704.04368v2","paper_title":"Get To The Point: Summarization with Pointer-Generator Networks","code":"https://github.com/PaddlePaddle/PaddleNLP/tree/develop/examples/text_summarization/pointer_summarizer","n_code_links":39,"syntology":{"n_ran":30,"n_unverified":34,"n_samples":64,"n_pointer_only_licence":44}},{"rank_in_archive_order":14,"model":"REFRESH","metrics":{"ROUGE-1":"40.0","ROUGE-2":"18.2","ROUGE-L":"36.6"},"uses_additional_data":false,"paper_date":"2018-02-23","paper":"/paper/ranking-sentences-for-extractive","paper_url":"http://arxiv.org/abs/1802.08636v2","paper_title":"Ranking Sentences for Extractive Summarization with Reinforcement Learning","code":"https://github.com/shashiongithub/Refresh","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":1,"n_samples":1,"n_pointer_only_licence":0}},{"rank_in_archive_order":15,"model":"ITS","metrics":{"ROUGE-1":"30.80","ROUGE-2":"12.6"},"uses_additional_data":false,"paper_date":"2018-09-27","paper":"/paper/iterative-document-representation-learning","paper_url":"https://arxiv.org/abs/1809.10324v2","paper_title":"Iterative Document Representation Learning Towards Summarization with Polishing","code":"https://github.com/yingtaomj/Iterative-Document-Representation-Learning-Towards-Summarization-with-Polishing","n_code_links":1,"syntology":null}],"since_archive":{"claim":"Results that newer papers report for their own method, placed here by Syntology. A model pointed at the cell in the paper's own table; the number was read from that cell and checked against this leaderboard's metric, dataset, split and scale; an independent check that saw this leaderboard's other rows and every other leaderboard on the same dataset accepted it. Not reviewed by the paper's authors or by the archive's editors, and not ranked against the archive rows.","extraction_file_present":true,"measurement":{"test_papers":883,"papers_with_output":881,"judged_true":108,"judged":110,"wilson95_lower":0.9361,"measured_on":"2026-09-24","frozen_commit":"0e3de0df94"},"measurement_note":"blind adjudication of accepted entries on a held-out split of archive papers, rules frozen before the test","coverage":{"sentence":"Syntology has checked 6,885 of the 9,623 papers on this site that are newer than the archive; results from the others appear after they are checked.","complete":false,"papers_newer_than_archive":9623,"papers_checked":6885,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":2737},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-25T09:33:49+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":2,"rows_with_any_sample_ran":1,"distinct_papers_with_graph_line":2,"distinct_papers_with_any_sample_ran":1,"samples_over_distinct_papers":{"n_ran":30,"n_unverified":35,"n_samples":65,"n_pointer_only_licence":44,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":30,"n_unverified":35,"n_samples":65,"n_pointer_only_licence":44,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}