{"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/domain-adaptation-with-asymmetrically-relaxed","title":"Domain Adaptation with Asymmetrically-Relaxed Distribution Alignment","arxiv_id":"1903.01689","date":"2019-03-05","proceeding":"ICLR Workshop LLD 2019","authors":["Yifan Wu","Ezra Winston","Divyansh Kaushik","Zachary Lipton"],"abstract":"Domain adaptation addresses the common problem when the target distribution\ngenerating our test data drifts from the source (training) distribution. While\nabsent assumptions, domain adaptation is impossible, strict conditions, e.g.\ncovariate or label shift, enable principled algorithms. Recently-proposed\ndomain-adversarial approaches consist of aligning source and target encodings,\noften motivating this approach as minimizing two (of three) terms in a\ntheoretical bound on target error. Unfortunately, this minimization can cause\narbitrary increases in the third term, e.g. they can break down under shifting\nlabel distributions. We propose asymmetrically-relaxed distribution alignment,\na new approach that overcomes some limitations of standard domain-adversarial\nalgorithms. Moreover, we characterize precise assumptions under which our\nalgorithm is theoretically principled and demonstrate empirical benefits on\nboth synthetic and real datasets.","url_abs":"http://arxiv.org/abs/1903.01689v2","url_pdf":"http://arxiv.org/pdf/1903.01689v2.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":"domain-adaptation-with-asymmetrically-relaxed","repo_url":"https://github.com/timgaripov/asa","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=1903.01689","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}