{"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/multinomial-adversarial-networks-for-multi","title":"Multinomial Adversarial Networks for Multi-Domain Text Classification","arxiv_id":"1802.05694","date":"2018-02-15","proceeding":"NAACL 2018 6","authors":["Xilun Chen","Claire Cardie"],"abstract":"Many text classification tasks are known to be highly domain-dependent.\nUnfortunately, the availability of training data can vary drastically across\ndomains. Worse still, for some domains there may not be any annotated data at\nall. In this work, we propose a multinomial adversarial network (MAN) to tackle\nthe text classification problem in this real-world multidomain setting (MDTC).\nWe provide theoretical justifications for the MAN framework, proving that\ndifferent instances of MANs are essentially minimizers of various f-divergence\nmetrics (Ali and Silvey, 1966) among multiple probability distributions. MANs\nare thus a theoretically sound generalization of traditional adversarial\nnetworks that discriminate over two distributions. More specifically, for the\nMDTC task, MAN learns features that are invariant across multiple domains by\nresorting to its ability to reduce the divergence among the feature\ndistributions of each domain. We present experimental results showing that MANs\nsignificantly outperform the prior art on the MDTC task. We also show that MANs\nachieve state-of-the-art performance for domains with no labeled data.","url_abs":"http://arxiv.org/abs/1802.05694v1","url_pdf":"http://arxiv.org/pdf/1802.05694v1.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":"multinomial-adversarial-networks-for-multi","repo_url":"https://github.com/ccsasuke/man","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"classification-1","task_name":"Classification"},{"task_slug":"cross-domain-text-classification","task_name":"Cross-Domain Text Classification"},{"task_slug":"domain-adaptation","task_name":"Domain Adaptation"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"text-classification","task_name":"Text Classification"},{"task_slug":"unsupervised-domain-adaptation","task_name":"Unsupervised Domain Adaptation"},{"task_slug":"text-classification-1","task_name":"text-classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1802.05694","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}