{"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/bohb-robust-and-efficient-hyperparameter","title":"BOHB: Robust and Efficient Hyperparameter Optimization at Scale","arxiv_id":"1807.01774","date":"2018-07-04","proceeding":"ICML 2018 7","authors":["Stefan Falkner","Aaron Klein","Frank Hutter"],"abstract":"Modern deep learning methods are very sensitive to many hyperparameters, and,\ndue to the long training times of state-of-the-art models, vanilla Bayesian\nhyperparameter optimization is typically computationally infeasible. On the\nother hand, bandit-based configuration evaluation approaches based on random\nsearch lack guidance and do not converge to the best configurations as quickly.\nHere, we propose to combine the benefits of both Bayesian optimization and\nbandit-based methods, in order to achieve the best of both worlds: strong\nanytime performance and fast convergence to optimal configurations. We propose\na new practical state-of-the-art hyperparameter optimization method, which\nconsistently outperforms both Bayesian optimization and Hyperband on a wide\nrange of problem types, including high-dimensional toy functions, support\nvector machines, feed-forward neural networks, Bayesian neural networks, deep\nreinforcement learning, and convolutional neural networks. Our method is robust\nand versatile, while at the same time being conceptually simple and easy to\nimplement.","url_abs":"http://arxiv.org/abs/1807.01774v1","url_pdf":"http://arxiv.org/pdf/1807.01774v1.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":"bohb-robust-and-efficient-hyperparameter","repo_url":"https://github.com/automl/HpBandSter","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok","spdx":"BSD-3-Clause"}},{"paper_slug":"bohb-robust-and-efficient-hyperparameter","repo_url":"https://github.com/PurityFan/bohb_in_nni","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"bohb-robust-and-efficient-hyperparameter","repo_url":"https://github.com/baggepinnen/hyperopt.jl","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"bohb-robust-and-efficient-hyperparameter","repo_url":"https://github.com/goktug97/bohb-hpo","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"bayesian-optimization","task_name":"Bayesian Optimization"},{"task_slug":"deep-reinforcement-learning","task_name":"Deep Reinforcement Learning"},{"task_slug":"hyperparameter-optimization","task_name":"Hyperparameter Optimization"},{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1807.01774","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}