{"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/bottom-up-abstractive-summarization","title":"Bottom-Up Abstractive Summarization","arxiv_id":"1808.10792","date":"2018-08-31","proceeding":"EMNLP 2018 10","authors":["Sebastian Gehrmann","Yuntian Deng","Alexander M. Rush"],"abstract":"Neural network-based methods for abstractive summarization produce outputs\nthat are more fluent than other techniques, but which can be poor at content\nselection. This work proposes a simple technique for addressing this issue: use\na data-efficient content selector to over-determine phrases in a source\ndocument that should be part of the summary. We use this selector as a\nbottom-up attention step to constrain the model to likely phrases. We show that\nthis approach improves the ability to compress text, while still generating\nfluent summaries. This two-step process is both simpler and higher performing\nthan other end-to-end content selection models, leading to significant\nimprovements on ROUGE for both the CNN-DM and NYT corpus. Furthermore, the\ncontent selector can be trained with as little as 1,000 sentences, making it\neasy to transfer a trained summarizer to a new domain.","url_abs":"http://arxiv.org/abs/1808.10792v2","url_pdf":"http://arxiv.org/pdf/1808.10792v2.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":[{"paper_slug":"bottom-up-abstractive-summarization","repo_url":"https://github.com/sebastianGehrmann/bottom-up-summary","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"bottom-up-abstractive-summarization","repo_url":"https://github.com/W4ngatang/qags","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"bottom-up-abstractive-summarization","repo_url":"https://github.com/j40903272/bottom-up-summary","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"bottom-up-abstractive-summarization","repo_url":"https://github.com/mythicalhacker/Text-Summarization","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"bottom-up-abstractive-summarization","repo_url":"https://github.com/prashanth41/text-summarization","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"NOASSERTION"}}],"tasks":[{"task_slug":"abstractive-text-summarization","task_name":"Abstractive Text Summarization"},{"task_slug":"document-summarization","task_name":"Document Summarization"},{"task_slug":"multi-document-summarization","task_name":"Multi-Document Summarization"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/abstractive-text-summarization-on-cnn-daily","task":"Abstractive Text Summarization","dataset":"CNN / Daily Mail","model":"Bottom-Up Summarization","rank_in_archive_order":35,"of":53,"metrics":{"ROUGE-1":"41.22","ROUGE-2":"18.68","ROUGE-L":"38.34"},"uses_additional_data":false},{"leaderboard":"/sota/document-summarization-on-cnn-daily-mail","task":"Document Summarization","dataset":"CNN / Daily Mail","model":"Bottom-Up Sum","rank_in_archive_order":16,"of":26,"metrics":{"PPL":"32.75","ROUGE-1":"41.22","ROUGE-2":"18.68","ROUGE-L":"38.34"},"uses_additional_data":false},{"leaderboard":"/sota/multi-document-summarization-on-multi-news","task":"Multi-Document Summarization","dataset":"Multi-News","model":"CopyTransformer","rank_in_archive_order":4,"of":6,"metrics":{"ROUGE-1":"43.57","ROUGE-2":"14.03","ROUGE-SU4":"17.37"},"uses_additional_data":false},{"leaderboard":"/sota/multi-document-summarization-on-multi-news","task":"Multi-Document Summarization","dataset":"Multi-News","model":"PG-BRNN","rank_in_archive_order":6,"of":6,"metrics":{"ROUGE-1":"42.80","ROUGE-2":"14.19","ROUGE-SU4":"16.75"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1808.10792","atlas_url":"https://app.syntology.ai/?focus=1808.10792","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1808.10792"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/W4ngatang/qags","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/sebastianGehrmann/bottom-up-summary","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/prashanth41/text-summarization","reach":{"status":"ok","spdx":"NOASSERTION"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/mythicalhacker/Text-Summarization","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/j40903272/bottom-up-summary","reach":{"status":"ok"}}],"summary":{"ran_draft_wrong":3,"ran_fixture":2,"unverified":1},"by_repo_kind":{"official":{"samples":6,"ran":5,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"17692adbeaedd54d","entry":"compile_substring","repo":"sebastianGehrmann/bottom-up-summary","repo_kind":"official","path":"preprocess_copy.py","file_url":"https://github.com/sebastianGehrmann/bottom-up-summary/blob/HEAD/preprocess_copy.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":true,"licence":"BSD-3-Clause","inline_ok":true,"mcp_get_code":{"code_sha256":"17692adbeaedd54d"}},{"code_sha256_prefix":"1489df8b19518049","entry":"format_json","repo":"sebastianGehrmann/bottom-up-summary","repo_kind":"official","path":"preprocess_copy.py","file_url":"https://github.com/sebastianGehrmann/bottom-up-summary/blob/HEAD/preprocess_copy.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":true,"licence":"BSD-3-Clause","inline_ok":true,"mcp_get_code":{"code_sha256":"1489df8b19518049"}},{"code_sha256_prefix":"6b890c90cd5ecfa6","entry":"get_sents","repo":"sebastianGehrmann/bottom-up-summary","repo_kind":"official","path":"prediction_to_text.py","file_url":"https://github.com/sebastianGehrmann/bottom-up-summary/blob/HEAD/prediction_to_text.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"deterministic","behaviour_fingerprint":true,"licence":"BSD-3-Clause","inline_ok":true,"mcp_get_code":{"code_sha256":"6b890c90cd5ecfa6"}},{"code_sha256_prefix":"20cb63f02f26da92","entry":"get_three","repo":"sebastianGehrmann/bottom-up-summary","repo_kind":"official","path":"prediction_to_text.py","file_url":"https://github.com/sebastianGehrmann/bottom-up-summary/blob/HEAD/prediction_to_text.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"BSD-3-Clause","inline_ok":true,"mcp_get_code":{"code_sha256":"20cb63f02f26da92"}},{"code_sha256_prefix":"12d9e8622f8706e2","entry":"splits","repo":"sebastianGehrmann/bottom-up-summary","repo_kind":"official","path":"preprocess_copy.py","file_url":"https://github.com/sebastianGehrmann/bottom-up-summary/blob/HEAD/preprocess_copy.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":true,"licence":"BSD-3-Clause","inline_ok":true,"mcp_get_code":{"code_sha256":"12d9e8622f8706e2"}},{"code_sha256_prefix":"e378ee2b26dd8ebc","entry":"get_phrases","repo":"sebastianGehrmann/bottom-up-summary","repo_kind":"official","path":"prediction_to_text.py","file_url":"https://github.com/sebastianGehrmann/bottom-up-summary/blob/HEAD/prediction_to_text.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"BSD-3-Clause","inline_ok":true,"mcp_get_code":{"code_sha256":"e378ee2b26dd8ebc"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}