{"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/semi-generative-modelling-covariate-shift","title":"Semi-Generative Modelling: Covariate-Shift Adaptation with Cause and Effect Features","arxiv_id":"1807.07879","date":"2018-07-20","proceeding":null,"authors":["Julius von Kügelgen","Alexander Mey","Marco Loog"],"abstract":"Current methods for covariate-shift adaptation use unlabelled data to compute\nimportance weights or domain-invariant features, while the final model is\ntrained on labelled data only. Here, we consider a particular case of covariate\nshift which allows us also to learn from unlabelled data, that is, combining\nadaptation with semi-supervised learning. Using ideas from causality, we argue\nthat this requires learning with both causes, $X_C$, and effects, $X_E$, of a\ntarget variable, $Y$, and show how this setting leads to what we call a\nsemi-generative model, $P(Y,X_E|X_C,\\theta)$. Our approach is robust to domain\nshifts in the distribution of causal features and leverages unlabelled data by\nlearning a direct map from causes to effects. Experiments on synthetic data\ndemonstrate significant improvements in classification over purely-supervised\nand importance-weighting baselines.","url_abs":"http://arxiv.org/abs/1807.07879v2","url_pdf":"http://arxiv.org/pdf/1807.07879v2.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":"semi-generative-modelling-covariate-shift","repo_url":"https://github.com/Juliusvk/Semi-Generative-Modelling","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"domain-adaptation","task_name":"Domain Adaptation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1807.07879","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}