{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,7]],"date-time":"2026-05-07T04:17:46Z","timestamp":1778127466337,"version":"3.51.4"},"reference-count":53,"publisher":"MIT Press","license":[{"start":{"date-parts":[[2023,7,13]],"date-time":"2023-07-13T00:00:00Z","timestamp":1689206400000},"content-version":"vor","delay-in-days":193,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["direct.mit.edu"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2023,7,12]]},"abstract":"<jats:title>Abstract<\/jats:title><jats:p>Controllable summarization allows users to generate customized summaries with specified attributes. However, due to the lack of designated annotations of controlled summaries, existing work has to craft pseudo datasets by adapting generic summarization benchmarks. Furthermore, most research focuses on controlling single attributes individually (e.g., a short summary or a highly abstractive summary) rather than controlling a mix of attributes together (e.g., a short and highly abstractive summary). In this paper, we propose MACSum, the first human-annotated summarization dataset for controlling mixed attributes. It contains source texts from two domains, news articles and dialogues, with human-annotated summaries controlled by five designed attributes (Length, Extractiveness, Specificity, Topic, and Speaker). We propose two simple and effective parameter-efficient approaches for the new task of mixed controllable summarization based on hard prompt tuning and soft prefix tuning. Results and analysis demonstrate that hard prompt models yield the best performance on most metrics and human evaluations. However, mixed-attribute control is still challenging for summarization tasks. Our dataset and code are available at https:\/\/github.com\/psunlpgroup\/MACSum.<\/jats:p>","DOI":"10.1162\/tacl_a_00575","type":"journal-article","created":{"date-parts":[[2023,7,13]],"date-time":"2023-07-13T16:42:12Z","timestamp":1689266532000},"page":"787-803","update-policy":"https:\/\/doi.org\/10.1162\/mitpressjournals.corrections.policy","source":"Crossref","is-referenced-by-count":14,"title":["<scp>MACSum<\/scp>: Controllable Summarization with Mixed Attributes"],"prefix":"10.1162","volume":"11","author":[{"given":"Yusen","family":"Zhang","sequence":"first","affiliation":[{"name":"Penn State University, USA. yfz5488@psu.edu"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yang","family":"Liu","sequence":"additional","affiliation":[{"name":"Microsoft Research, USA. yaliu10@microsoft.com"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Ziyi","family":"Yang","sequence":"additional","affiliation":[{"name":"Microsoft Research, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yuwei","family":"Fang","sequence":"additional","affiliation":[{"name":"Microsoft Research, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yulong","family":"Chen","sequence":"additional","affiliation":[{"name":"Westlake University, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Dragomir","family":"Radev","sequence":"additional","affiliation":[{"name":"Yale University, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Chenguang","family":"Zhu","sequence":"additional","affiliation":[{"name":"Microsoft Research, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Michael","family":"Zeng","sequence":"additional","affiliation":[{"name":"Microsoft Research, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Rui","family":"Zhang","sequence":"additional","affiliation":[{"name":"Penn State University, USA. rmz5227@psu.edu"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"281","published-online":{"date-parts":[[2023,7,12]]},"reference":[{"key":"2023071316415385300_bib1","doi-asserted-by":"publisher","first-page":"6578","DOI":"10.18653\/v1\/2021.emnlp-main.528","article-title":"Aspect-controllable opinion summarization","volume-title":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing","author":"Amplayo","year":"2021"},{"key":"2023071316415385300_bib2","doi-asserted-by":"publisher","first-page":"3110","DOI":"10.18653\/v1\/D19-1307","article-title":"Better rewards yield better summaries: Learning to summarise without references","volume-title":"Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)","author":"B\u00f6hm","year":"2019"},{"key":"2023071316415385300_bib3","article-title":"Interactive document summarization","author":"Bornstein","year":"1999"},{"key":"2023071316415385300_bib4","first-page":"69","article-title":"pke: An open source python-based keyphrase extraction toolkit","volume-title":"Proceedings of COLING 2016, the 26th International Conference on Computational Linguistics: System Demonstrations","author":"Boudin","year":"2016"},{"key":"2023071316415385300_bib5","article-title":"Language models are few-shot learners","volume-title":"Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems 2020, NeurIPS 2020, December 6\u201312, 2020, virtual","author":"Brown","year":"2020"},{"key":"2023071316415385300_bib6","doi-asserted-by":"publisher","first-page":"5942","DOI":"10.18653\/v1\/2021.naacl-main.476","article-title":"Inference time style control for summarization","volume-title":"Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies","author":"Cao","year":"2021"},{"key":"2023071316415385300_bib7","doi-asserted-by":"publisher","first-page":"1213","DOI":"10.1162\/tacl_a_00423","article-title":"Controllable summarization with constrained Markov decision process","volume":"9","author":"Chan","year":"2021","journal-title":"Transactions of the Association for Computational Linguistics"},{"key":"2023071316415385300_bib8","doi-asserted-by":"crossref","first-page":"6057","DOI":"10.18653\/v1\/2022.findings-emnlp.448","article-title":"AdaPrompt: Adaptive model training for prompt-based NLP","volume-title":"Findings of the Association for Computational Linguistics: EMNLP 2022","author":"Chen","year":"2022"},{"key":"2023071316415385300_bib9","doi-asserted-by":"publisher","first-page":"484","DOI":"10.18653\/v1\/P16-1046","article-title":"Neural summarization by extracting sentences and words","volume-title":"Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","author":"Cheng","year":"2016"},{"issue":"1","key":"2023071316415385300_bib10","doi-asserted-by":"publisher","first-page":"37","DOI":"10.1177\/001316446002000104","article-title":"A coefficient of agreement for nominal scales","volume":"20","author":"Cohen","year":"1960","journal-title":"Educational and Psychological Measurement"},{"key":"2023071316415385300_bib11","first-page":"1","article-title":"Overview of DUC 2005","volume-title":"Proceedings of the Document Understanding Conference","author":"Dang","year":"2005"},{"key":"2023071316415385300_bib12","first-page":"305","article-title":"Bayesian query-focused summarization","volume-title":"Proceedings of the 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics","author":"Daum\u00e9","year":"2006"},{"key":"2023071316415385300_bib13","doi-asserted-by":"publisher","first-page":"4830","DOI":"10.18653\/v1\/2021.naacl-main.384","article-title":"GSum: A general framework for guided neural abstractive summarization","volume-title":"Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies","author":"Dou","year":"2021"},{"key":"2023071316415385300_bib14","doi-asserted-by":"publisher","first-page":"457","DOI":"10.1613\/jair.1523","article-title":"Lexrank: Graph-based lexical centrality as salience in text summarization","volume":"22","author":"Erkan","year":"2004","journal-title":"Journal of Artificial Intelligence Research"},{"key":"2023071316415385300_bib15","doi-asserted-by":"publisher","first-page":"45","DOI":"10.18653\/v1\/W18-2706","article-title":"Controllable abstractive summarization","volume-title":"Proceedings of the 2nd Workshop on Neural Machine Translation and Generation","author":"Fan","year":"2018"},{"key":"2023071316415385300_bib16","article-title":"Query-focused summarization by supervised sentence ranking and skewed word distributions","volume-title":"Proceedings of the Document Understanding Conference, DUC-2006, New York, USA","author":"Fisher","year":"2006"},{"key":"2023071316415385300_bib17","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2209.12356","article-title":"News summarization and evaluation in the era of GPT-3","volume":"abs\/2209.12356","author":"Goyal","year":"2022","journal-title":"ArXiv preprint"},{"key":"2023071316415385300_bib18","doi-asserted-by":"crossref","first-page":"464","DOI":"10.18653\/v1\/2022.emnlp-main.30","article-title":"HydraSum: Disentangling style features in text summarization with multi-decoder models","volume-title":"Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing","author":"Goyal","year":"2022"},{"key":"2023071316415385300_bib19","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2012.04281","article-title":"Ctrlsum: Towards generic controllable text summarization","author":"He","year":"2020","journal-title":"ArXiv preprint"},{"key":"2023071316415385300_bib20","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2208.09770","article-title":"Z-code++: A pre-trained language model optimized for abstractive summarization","author":"He","year":"2022","journal-title":"ArXiv preprint"},{"key":"2023071316415385300_bib21","first-page":"1693","article-title":"Teaching machines to read and comprehend","volume-title":"Advances in Neural Information Processing Systems 28: Annual Conference on Neural Information Processing Systems 2015, December 7\u201312, 2015, Montreal, Quebec, Canada","author":"Hermann","year":"2015"},{"key":"2023071316415385300_bib22","article-title":"Extractive entity-centric summarization as sentence selection using bi-encoders","volume-title":"Proceedings of the 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing (Volume 2: Short Papers)","author":"Hofmann-Coyle","year":"2022"},{"key":"2023071316415385300_bib23","doi-asserted-by":"publisher","first-page":"3045","DOI":"10.18653\/v1\/2021.emnlp-main.243","article-title":"The power of scale for parameter-efficient prompt tuning","volume-title":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing","author":"Lester","year":"2021"},{"key":"2023071316415385300_bib24","doi-asserted-by":"publisher","first-page":"125","DOI":"10.3115\/1075178.1075197","article-title":"iNeATS: Interactive multi-document summarization","volume-title":"The Companion Volume to the Proceedings of 41st Annual Meeting of the Association for Computational Linguistics","author":"Leuski","year":"2003"},{"key":"2023071316415385300_bib25","doi-asserted-by":"publisher","first-page":"7871","DOI":"10.18653\/v1\/2020.acl-main.703","article-title":"BART: Denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension","volume-title":"Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics","author":"Lewis","year":"2020"},{"key":"2023071316415385300_bib26","doi-asserted-by":"publisher","first-page":"4582","DOI":"10.18653\/v1\/2021.acl-long.353","article-title":"Prefix-tuning: Optimizing continuous prompts for generation","volume-title":"Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)","author":"Li","year":"2021"},{"key":"2023071316415385300_bib27","first-page":"74","article-title":"ROUGE: A package for automatic evaluation of summaries","volume-title":"Text Summarization Branches Out","author":"Lin","year":"2004"},{"key":"2023071316415385300_bib28","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2103.10385","article-title":"GPT understands, too","author":"Liu","year":"2021","journal-title":"ArXiv preprint"},{"key":"2023071316415385300_bib29","doi-asserted-by":"publisher","first-page":"6885","DOI":"10.18653\/v1\/2022.acl-long.474","article-title":"Length control in abstractive summarization by pretraining information selection","volume-title":"Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","author":"Liu","year":"2022"},{"key":"2023071316415385300_bib30","doi-asserted-by":"publisher","first-page":"4110","DOI":"10.18653\/v1\/D18-1444","article-title":"Controlling length in abstractive summarization using a convolutional neural network","volume-title":"Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing","author":"Liu","year":"2018"},{"key":"2023071316415385300_bib31","first-page":"34","article-title":"Text specificity and impact on quality of news summaries","volume-title":"Proceedings of the Workshop on Monolingual Text-To-Text Generation","author":"Louis","year":"2011"},{"key":"2023071316415385300_bib32","doi-asserted-by":"publisher","first-page":"3355","DOI":"10.18653\/v1\/2022.acl-long.237","article-title":"EntSUM: A data set for entity-centric extractive summarization","volume-title":"Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","author":"Maddela","year":"2022"},{"key":"2023071316415385300_bib33","doi-asserted-by":"publisher","first-page":"1039","DOI":"10.18653\/v1\/P19-1099","article-title":"Global optimization under length constraint for neural text summarization","volume-title":"Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics","author":"Makino","year":"2019"},{"key":"2023071316415385300_bib34","first-page":"74","article-title":"Generating summaries of multiple news articles","volume-title":"Proceedings of the 18th annual international ACM SIGIR conference on Research and development in information retrieval","author":"McKeown","year":"1995"},{"key":"2023071316415385300_bib35","doi-asserted-by":"publisher","first-page":"5316","DOI":"10.18653\/v1\/2022.acl-long.365","article-title":"Noisy channel language model prompting for few-shot text classification","volume-title":"Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","author":"Min","year":"2022"},{"key":"2023071316415385300_bib36","doi-asserted-by":"publisher","first-page":"1475","DOI":"10.1162\/tacl_a_00438","article-title":"Planning with learned entity prompts for abstractive summarization","volume":"9","author":"Narayan","year":"2021","journal-title":"Transactions of the Association for Computational Linguistics"},{"key":"2023071316415385300_bib37","article-title":"A deep reinforced model for abstractive summarization","volume-title":"6th International Conference on Learning Representations, ICLR 2018, Vancouver, BC, Canada, April 30 \u2013 May 3, 2018, Conference Track Proceedings","author":"Paulus","year":"2018"},{"key":"2023071316415385300_bib38","doi-asserted-by":"publisher","first-page":"5203","DOI":"10.18653\/v1\/2021.naacl-main.410","article-title":"Learning how to ask: Querying LMs with mixtures of soft prompts","volume-title":"Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies","author":"Qin","year":"2021"},{"issue":"140","key":"2023071316415385300_bib39","first-page":"1","article-title":"Exploring the limits of transfer learning with a unified text-to-text transformer.","volume":"21","author":"Raffel","year":"2020","journal-title":"Journal of Machine Learning Research"},{"key":"2023071316415385300_bib40","article-title":"Using information content to evaluate semantic similarity in a taxonomy","author":"Resnik","year":"1995","journal-title":"arXiv preprint cmp-lg\/9511007"},{"key":"2023071316415385300_bib41","doi-asserted-by":"publisher","first-page":"379","DOI":"10.18653\/v1\/D15-1044","article-title":"A neural attention model for abstractive sentence summarization","volume-title":"Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing","author":"Rush","year":"2015"},{"key":"2023071316415385300_bib42","doi-asserted-by":"publisher","first-page":"351","DOI":"10.18653\/v1\/2020.findings-emnlp.33","article-title":"Control, generate, augment: A scalable framework for multi-attribute text generation","volume-title":"Findings of the Association for Computational Linguistics: EMNLP 2020","author":"Russo","year":"2020"},{"key":"2023071316415385300_bib43","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2001.07331","article-title":"Length-controllable abstractive summarization by guiding with summary prototype","author":"Saito","year":"2020","journal-title":"ArXiv preprint"},{"key":"2023071316415385300_bib44","doi-asserted-by":"publisher","first-page":"2339","DOI":"10.18653\/v1\/2021.naacl-main.185","article-title":"It\u2019s not just size that matters: Small language models are also few-shot learners","volume-title":"Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies","author":"Schick","year":"2021"},{"key":"2023071316415385300_bib45","doi-asserted-by":"publisher","first-page":"1073","DOI":"10.18653\/v1\/P17-1099","article-title":"Get to the point: Summarization with pointer-generator networks","volume-title":"Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","author":"See","year":"2017"},{"key":"2023071316415385300_bib46","doi-asserted-by":"crossref","first-page":"4222","DOI":"10.18653\/v1\/2020.emnlp-main.346","article-title":"Eliciting knowledge from language models using automatically generated prompts","volume-title":"Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)","author":"Shin","year":"2020"},{"key":"2023071316415385300_bib47","first-page":"8902","article-title":"Controlling the amount of verbatim copying in abstractive summarization","volume-title":"The Thirty-Fourth AAAI Conference on Artificial Intelligence, AAAI 2020, The Thirty-Second Innovative Applications of Artificial Intelligence Conference, IAAI 2020, The Tenth AAAI Symposium on Educational Advances in Artificial Intelligence, EAAI 2020, New York, NY, USA, February 7\u201312, 2020","author":"Song","year":"2020"},{"key":"2023071316415385300_bib48","article-title":"Learning to summarize with human feedback","volume-title":"Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems 2020, NeurIPS 2020, December 6\u201312, 2020, virtual","author":"Stiennon","year":"2020"},{"key":"2023071316415385300_bib49","doi-asserted-by":"publisher","first-page":"6301","DOI":"10.18653\/v1\/2020.emnlp-main.510","article-title":"Summarizing text on any aspects: A knowledge-informed weakly-supervised approach","volume-title":"Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)","author":"Tan","year":"2020"},{"key":"2023071316415385300_bib50","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2020.emnlp-demos.6","article-title":"HuggingFace\u2019s transformers: State-of-the-art natural language processing","author":"Wolf","year":"2019","journal-title":"ArXiv preprint"},{"key":"2023071316415385300_bib51","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2212.10819","article-title":"Attend to the right context: A plug-and-play module for content-controllable summarization","author":"Xiao","year":"2022"},{"key":"2023071316415385300_bib52","doi-asserted-by":"crossref","DOI":"10.18653\/v1\/2022.findings-emnlp.366","article-title":"Unsupervised summarization with customized granularities","author":"Zhong","year":"2022","journal-title":"Findings of EMNLP 2022"},{"key":"2023071316415385300_bib53","doi-asserted-by":"publisher","first-page":"5905","DOI":"10.18653\/v1\/2021.naacl-main.472","article-title":"QMSum: A new benchmark for query-based multi-domain meeting summarization","volume-title":"Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies","author":"Zhong","year":"2021"}],"container-title":["Transactions of the Association for Computational Linguistics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/direct.mit.edu\/tacl\/article-pdf\/doi\/10.1162\/tacl_a_00575\/2143288\/tacl_a_00575.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/direct.mit.edu\/tacl\/article-pdf\/doi\/10.1162\/tacl_a_00575\/2143288\/tacl_a_00575.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,12,17]],"date-time":"2023-12-17T01:16:33Z","timestamp":1702775793000},"score":1,"resource":{"primary":{"URL":"https:\/\/direct.mit.edu\/tacl\/article\/doi\/10.1162\/tacl_a_00575\/116714\/MACSum-Controllable-Summarization-with-Mixed"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023]]},"references-count":53,"URL":"https:\/\/doi.org\/10.1162\/tacl_a_00575","relation":{},"ISSN":["2307-387X"],"issn-type":[{"value":"2307-387X","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2023]]},"published":{"date-parts":[[2023]]}}}