{"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/multi-fidelity-gaussian-process-bandit","title":"Multi-fidelity Gaussian Process Bandit Optimisation","arxiv_id":"1603.06288","date":"2016-03-20","proceeding":null,"authors":["Kirthevasan Kandasamy","Gautam Dasarathy","Junier B. Oliva","Jeff Schneider","Barnabas Poczos"],"abstract":"In many scientific and engineering applications, we are tasked with the\nmaximisation of an expensive to evaluate black box function $f$. Traditional\nsettings for this problem assume just the availability of this single function.\nHowever, in many cases, cheap approximations to $f$ may be obtainable. For\nexample, the expensive real world behaviour of a robot can be approximated by a\ncheap computer simulation. We can use these approximations to eliminate low\nfunction value regions cheaply and use the expensive evaluations of $f$ in a\nsmall but promising region and speedily identify the optimum. We formalise this\ntask as a \\emph{multi-fidelity} bandit problem where the target function and\nits approximations are sampled from a Gaussian process. We develop MF-GP-UCB, a\nnovel method based on upper confidence bound techniques. In our theoretical\nanalysis we demonstrate that it exhibits precisely the above behaviour, and\nachieves better regret than strategies which ignore multi-fidelity information.\nEmpirically, MF-GP-UCB outperforms such naive strategies and other\nmulti-fidelity methods on several synthetic and real experiments.","url_abs":"http://arxiv.org/abs/1603.06288v4","url_pdf":"http://arxiv.org/pdf/1603.06288v4.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":"multi-fidelity-gaussian-process-bandit","repo_url":"https://github.com/kirthevasank/mf-gp-ucb","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1603.06288","atlas_url":"https://app.syntology.ai/?focus=1603.06288","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}