{"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/sticking-the-landing-simple-lower-variance","title":"Sticking the Landing: Simple, Lower-Variance Gradient Estimators for Variational Inference","arxiv_id":"1703.09194","date":"2017-03-27","proceeding":"NeurIPS 2017 12","authors":["Geoffrey Roeder","Yuhuai Wu","David Duvenaud"],"abstract":"We propose a simple and general variant of the standard reparameterized\ngradient estimator for the variational evidence lower bound. Specifically, we\nremove a part of the total derivative with respect to the variational\nparameters that corresponds to the score function. Removing this term produces\nan unbiased gradient estimator whose variance approaches zero as the\napproximate posterior approaches the exact posterior. We analyze the behavior\nof this gradient estimator theoretically and empirically, and generalize it to\nmore complex variational distributions such as mixtures and importance-weighted\nposteriors.","url_abs":"http://arxiv.org/abs/1703.09194v3","url_pdf":"http://arxiv.org/pdf/1703.09194v3.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":"sticking-the-landing-simple-lower-variance","repo_url":"https://github.com/geoffroeder/iwae","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"variational-inference","task_name":"Variational Inference"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1703.09194","atlas_url":"https://app.syntology.ai/?focus=1703.09194","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}