{"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/gradient-based-hyperparameter-optimization","title":"Gradient-based Hyperparameter Optimization through Reversible Learning","arxiv_id":"1502.03492","date":"2015-02-11","proceeding":null,"authors":["Dougal Maclaurin","David Duvenaud","Ryan P. Adams"],"abstract":"Tuning hyperparameters of learning algorithms is hard because gradients are\nusually unavailable. We compute exact gradients of cross-validation performance\nwith respect to all hyperparameters by chaining derivatives backwards through\nthe entire training procedure. These gradients allow us to optimize thousands\nof hyperparameters, including step-size and momentum schedules, weight\ninitialization distributions, richly parameterized regularization schemes, and\nneural network architectures. We compute hyperparameter gradients by exactly\nreversing the dynamics of stochastic gradient descent with momentum.","url_abs":"http://arxiv.org/abs/1502.03492v3","url_pdf":"http://arxiv.org/pdf/1502.03492v3.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":"gradient-based-hyperparameter-optimization","repo_url":"https://github.com/HIPS/hypergrad","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"gradient-based-hyperparameter-optimization","repo_url":"https://github.com/Przemo23/Reversing_Gradient_Differentiation_NN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}}],"tasks":[{"task_slug":"hyperparameter-optimization","task_name":"Hyperparameter Optimization"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1502.03492","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}