{"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/fast-and-scalable-bayesian-deep-learning-by","title":"Fast and Scalable Bayesian Deep Learning by Weight-Perturbation in Adam","arxiv_id":"1806.04854","date":"2018-06-13","proceeding":"ICML 2018 7","authors":["Mohammad Emtiyaz Khan","Didrik Nielsen","Voot Tangkaratt","Wu Lin","Yarin Gal","Akash Srivastava"],"abstract":"Uncertainty computation in deep learning is essential to design robust and\nreliable systems. Variational inference (VI) is a promising approach for such\ncomputation, but requires more effort to implement and execute compared to\nmaximum-likelihood methods. In this paper, we propose new natural-gradient\nalgorithms to reduce such efforts for Gaussian mean-field VI. Our algorithms\ncan be implemented within the Adam optimizer by perturbing the network weights\nduring gradient evaluations, and uncertainty estimates can be cheaply obtained\nby using the vector that adapts the learning rate. This requires lower memory,\ncomputation, and implementation effort than existing VI methods, while\nobtaining uncertainty estimates of comparable quality. Our empirical results\nconfirm this and further suggest that the weight-perturbation in our algorithm\ncould be useful for exploration in reinforcement learning and stochastic\noptimization.","url_abs":"http://arxiv.org/abs/1806.04854v3","url_pdf":"http://arxiv.org/pdf/1806.04854v3.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":"fast-and-scalable-bayesian-deep-learning-by","repo_url":"https://github.com/emtiyaz/vadam","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"fast-and-scalable-bayesian-deep-learning-by","repo_url":"https://github.com/lamantinushka/StructuredCovariance","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"fast-and-scalable-bayesian-deep-learning-by","repo_url":"https://github.com/tr7200/Vadam","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"},{"task_slug":"stochastic-optimization","task_name":"Stochastic Optimization"},{"task_slug":"variational-inference","task_name":"Variational Inference"}],"methods":[{"method_slug":"adam","method_name":"Adam"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1806.04854","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}