{"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/adversarial-feature-augmentation-for","title":"Adversarial Feature Augmentation for Unsupervised Domain Adaptation","arxiv_id":"1711.08561","date":"2017-11-23","proceeding":"CVPR 2018 6","authors":["Riccardo Volpi","Pietro Morerio","Silvio Savarese","Vittorio Murino"],"abstract":"Recent works showed that Generative Adversarial Networks (GANs) can be\nsuccessfully applied in unsupervised domain adaptation, where, given a labeled\nsource dataset and an unlabeled target dataset, the goal is to train powerful\nclassifiers for the target samples. In particular, it was shown that a GAN\nobjective function can be used to learn target features indistinguishable from\nthe source ones. In this work, we extend this framework by (i) forcing the\nlearned feature extractor to be domain-invariant, and (ii) training it through\ndata augmentation in the feature space, namely performing feature augmentation.\nWhile data augmentation in the image space is a well established technique in\ndeep learning, feature augmentation has not yet received the same level of\nattention. We accomplish it by means of a feature generator trained by playing\nthe GAN minimax game against source features. Results show that both enforcing\ndomain-invariance and performing feature augmentation lead to superior or\ncomparable performance to state-of-the-art results in several unsupervised\ndomain adaptation benchmarks.","url_abs":"http://arxiv.org/abs/1711.08561v2","url_pdf":"http://arxiv.org/pdf/1711.08561v2.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":"adversarial-feature-augmentation-for","repo_url":"https://github.com/ricvolpi/adversarial_feature_augmentation","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"adversarial-feature-augmentation-for","repo_url":"https://github.com/ricvolpi/adversarial-feature-augmentation","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"data-augmentation","task_name":"Data Augmentation"},{"task_slug":"domain-adaptation","task_name":"Domain Adaptation"},{"task_slug":"unsupervised-domain-adaptation","task_name":"Unsupervised Domain Adaptation"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1711.08561","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}