{"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/hydrasum-disentangling-stylistic-features-in-1","title":"HydraSum: Disentangling Stylistic Features in Text Summarization using Multi-Decoder Models","arxiv_id":"2110.04400","date":"2021-10-08","proceeding":null,"authors":["Tanya Goyal","Nazneen Fatema Rajani","Wenhao Liu","Wojciech Kryściński"],"abstract":"Summarization systems make numerous \"decisions\" about summary properties during inference, e.g. degree of copying, specificity and length of outputs, etc. However, these are implicitly encoded within model parameters and specific styles cannot be enforced. To address this, we introduce HydraSum, a new summarization architecture that extends the single decoder framework of current models to a mixture-of-experts version with multiple decoders. We show that HydraSum's multiple decoders automatically learn contrasting summary styles when trained under the standard training objective without any extra supervision. Through experiments on three summarization datasets (CNN, Newsroom and XSum), we show that HydraSum provides a simple mechanism to obtain stylistically-diverse summaries by sampling from either individual decoders or their mixtures, outperforming baseline models. Finally, we demonstrate that a small modification to the gating strategy during training can enforce an even stricter style partitioning, e.g. high- vs low-abstractiveness or high- vs low-specificity, allowing users to sample from a larger area in the generation space and vary summary styles along multiple dimensions.","url_abs":"https://arxiv.org/abs/2110.04400v4","url_pdf":"https://arxiv.org/pdf/2110.04400v4.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":"hydrasum-disentangling-stylistic-features-in-1","repo_url":"https://github.com/salesforce/hydra-sum","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"BSD-3-Clause"}}],"tasks":[{"task_slug":"abstractive-text-summarization","task_name":"Abstractive Text Summarization"},{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"mixture-of-experts","task_name":"Mixture-of-Experts"},{"task_slug":"specificity","task_name":"Specificity"},{"task_slug":"text-summarization","task_name":"Text Summarization"}],"methods":[{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"bart","method_name":"BART"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2110.04400","atlas_url":"https://app.syntology.ai/?focus=2110.04400","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2110.04400"}},"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/salesforce/hydra-sum","reach":{"status":"ok","spdx":"BSD-3-Clause"}}],"summary":{"ran":3,"unverified":6},"by_repo_kind":{"official":{"samples":9,"ran":3,"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":"609763ac0cecc7ee","entry":"compute_accuracy_score","repo":"salesforce/hydra-sum","repo_kind":"official","path":"modelling/train_seq2seq.py","file_url":"https://github.com/salesforce/hydra-sum/blob/HEAD/modelling/train_seq2seq.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"BSD-3-Clause","inline_ok":true,"mcp_get_code":{"code_sha256":"609763ac0cecc7ee"}},{"code_sha256_prefix":"397bbbe7fee1493e","entry":"convert_examples_to_features_pegasus","repo":"salesforce/hydra-sum","repo_kind":"official","path":"modelling/train_seq2seq_utils.py","file_url":"https://github.com/salesforce/hydra-sum/blob/HEAD/modelling/train_seq2seq_utils.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"BSD-3-Clause","inline_ok":true,"mcp_get_code":{"code_sha256":"397bbbe7fee1493e"}},{"code_sha256_prefix":"1b2cef5d2e4a09bb","entry":"get_examples","repo":"salesforce/hydra-sum","repo_kind":"official","path":"modelling/train_seq2seq_utils.py","file_url":"https://github.com/salesforce/hydra-sum/blob/HEAD/modelling/train_seq2seq_utils.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"BSD-3-Clause","inline_ok":true,"mcp_get_code":{"code_sha256":"1b2cef5d2e4a09bb"}},{"code_sha256_prefix":"1b019761ed7a2bd0","entry":"convert_examples_to_features","repo":"salesforce/hydra-sum","repo_kind":"official","path":"modelling/train_seq2seq_utils.py","file_url":"https://github.com/salesforce/hydra-sum/blob/HEAD/modelling/train_seq2seq_utils.py","link_basis":"harvester_set","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":"1b019761ed7a2bd0"}},{"code_sha256_prefix":"b61cdc748ec6ff41","entry":"get_extractive_coverage","repo":"salesforce/hydra-sum","repo_kind":"official","path":"postprocessing/extractive_fragments_utils.py","file_url":"https://github.com/salesforce/hydra-sum/blob/HEAD/postprocessing/extractive_fragments_utils.py","link_basis":"harvester_set","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":"b61cdc748ec6ff41"}},{"code_sha256_prefix":"94735fc7331c68ce","entry":"get_extractive_fragments","repo":"salesforce/hydra-sum","repo_kind":"official","path":"postprocessing/extractive_fragments_utils.py","file_url":"https://github.com/salesforce/hydra-sum/blob/HEAD/postprocessing/extractive_fragments_utils.py","link_basis":"harvester_set","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":"94735fc7331c68ce"}},{"code_sha256_prefix":"0c36571d6b1304c6","entry":"get_fragment_density","repo":"salesforce/hydra-sum","repo_kind":"official","path":"postprocessing/extractive_fragments_utils.py","file_url":"https://github.com/salesforce/hydra-sum/blob/HEAD/postprocessing/extractive_fragments_utils.py","link_basis":"harvester_set","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":"0c36571d6b1304c6"}},{"code_sha256_prefix":"be2b77546711701c","entry":"get_overlap","repo":"salesforce/hydra-sum","repo_kind":"official","path":"postprocessing/lexical_overlap.py","file_url":"https://github.com/salesforce/hydra-sum/blob/HEAD/postprocessing/lexical_overlap.py","link_basis":"harvester_set","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":"be2b77546711701c"}},{"code_sha256_prefix":"6ff04903ffe27357","entry":"get_overlap_file","repo":"salesforce/hydra-sum","repo_kind":"official","path":"postprocessing/lexical_overlap.py","file_url":"https://github.com/salesforce/hydra-sum/blob/HEAD/postprocessing/lexical_overlap.py","link_basis":"harvester_set","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":"6ff04903ffe27357"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}