{"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/proximity-variational-inference","title":"Proximity Variational Inference","arxiv_id":"1705.08931","date":"2017-05-24","proceeding":null,"authors":["Jaan Altosaar","Rajesh Ranganath","David M. Blei"],"abstract":"Variational inference is a powerful approach for approximate posterior\ninference. However, it is sensitive to initialization and can be subject to\npoor local optima. In this paper, we develop proximity variational inference\n(PVI). PVI is a new method for optimizing the variational objective that\nconstrains subsequent iterates of the variational parameters to robustify the\noptimization path. Consequently, PVI is less sensitive to initialization and\noptimization quirks and finds better local optima. We demonstrate our method on\nthree proximity statistics. We study PVI on a Bernoulli factor model and\nsigmoid belief network with both real and synthetic data and compare to\ndeterministic annealing (Katahira et al., 2008). We highlight the flexibility\nof PVI by designing a proximity statistic for Bayesian deep learning models\nsuch as the variational autoencoder (Kingma and Welling, 2014; Rezende et al.,\n2014). Empirically, we show that PVI consistently finds better local optima and\ngives better predictive performance.","url_abs":"http://arxiv.org/abs/1705.08931v1","url_pdf":"http://arxiv.org/pdf/1705.08931v1.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":"proximity-variational-inference","repo_url":"https://github.com/altosaar/proximity_vi","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"variational-inference","task_name":"Variational Inference"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1705.08931","atlas_url":"https://app.syntology.ai/?focus=1705.08931","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}