{"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/learning-functional-causal-models-with","title":"Learning Functional Causal Models with Generative Neural Networks","arxiv_id":"1709.05321","date":"2017-09-15","proceeding":null,"authors":["Olivier Goudet","Diviyan Kalainathan","Philippe Caillou","Isabelle Guyon","David Lopez-Paz","Michèle Sebag"],"abstract":"We introduce a new approach to functional causal modeling from observational\ndata, called Causal Generative Neural Networks (CGNN). CGNN leverages the power\nof neural networks to learn a generative model of the joint distribution of the\nobserved variables, by minimizing the Maximum Mean Discrepancy between\ngenerated and observed data. An approximate learning criterion is proposed to\nscale the computational cost of the approach to linear complexity in the number\nof observations. The performance of CGNN is studied throughout three\nexperiments. Firstly, CGNN is applied to cause-effect inference, where the task\nis to identify the best causal hypothesis out of $X\\rightarrow Y$ and\n$Y\\rightarrow X$. Secondly, CGNN is applied to the problem of identifying\nv-structures and conditional independences. Thirdly, CGNN is applied to\nmultivariate functional causal modeling: given a skeleton describing the direct\ndependences in a set of random variables $\\textbf{X} = [X_1, \\ldots, X_d]$,\nCGNN orients the edges in the skeleton to uncover the directed acyclic causal\ngraph describing the causal structure of the random variables. On all three\ntasks, CGNN is extensively assessed on both artificial and real-world data,\ncomparing favorably to the state-of-the-art. Finally, CGNN is extended to\nhandle the case of confounders, where latent variables are involved in the\noverall causal model.","url_abs":"http://arxiv.org/abs/1709.05321v3","url_pdf":"http://arxiv.org/pdf/1709.05321v3.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":"learning-functional-causal-models-with","repo_url":"https://github.com/GoudetOlivier/CGNN","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"learning-functional-causal-models-with","repo_url":"https://github.com/yrodill/internship-Angers","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}