{"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/open-set-domain-adaptation-by-backpropagation","title":"Open Set Domain Adaptation by Backpropagation","arxiv_id":"1804.10427","date":"2018-04-27","proceeding":"ECCV 2018 9","authors":["Kuniaki Saito","Shohei Yamamoto","Yoshitaka Ushiku","Tatsuya Harada"],"abstract":"Numerous algorithms have been proposed for transferring knowledge from a\nlabel-rich domain (source) to a label-scarce domain (target). Almost all of\nthem are proposed for a closed-set scenario, where the source and the target\ndomain completely share the class of their samples. We call the shared class\nthe \\doublequote{known class.} However, in practice, when samples in target\ndomain are not labeled, we cannot know whether the domains share the class. A\ntarget domain can contain samples of classes that are not shared by the source\ndomain. We call such classes the \\doublequote{unknown class} and algorithms\nthat work well in the open set situation are very practical. However, most\nexisting distribution matching methods for domain adaptation do not work well\nin this setting because unknown target samples should not be aligned with the\nsource.\n  In this paper, we propose a method for an open set domain adaptation scenario\nwhich utilizes adversarial training. A classifier is trained to make a boundary\nbetween the source and the target samples whereas a generator is trained to\nmake target samples far from the boundary. Thus, we assign two options to the\nfeature generator: aligning them with source known samples or rejecting them as\nunknown target samples. This approach allows extracting features that separate\nunknown target samples from known target samples. Our method was extensively\nevaluated in domain adaptation setting and outperformed other methods with a\nlarge margin in most settings.","url_abs":"http://arxiv.org/abs/1804.10427v2","url_pdf":"http://arxiv.org/pdf/1804.10427v2.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":"open-set-domain-adaptation-by-backpropagation","repo_url":"https://github.com/ChenJinBIT/OSDA","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"open-set-domain-adaptation-by-backpropagation","repo_url":"https://github.com/ksaito-ut/OPDA_BP","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"open-set-domain-adaptation-by-backpropagation","repo_url":"https://github.com/redhat12345/Domain-Adaptation-Papers-and-Codes","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"open-set-domain-adaptation-by-backpropagation","repo_url":"https://github.com/thuml/Transfer-Learning-Library","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"domain-adaptation","task_name":"Domain Adaptation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1804.10427","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}