{"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/kernel-based-just-in-time-learning-for","title":"Kernel-Based Just-In-Time Learning for Passing Expectation Propagation Messages","arxiv_id":"1503.02551","date":"2015-03-09","proceeding":null,"authors":["Wittawat Jitkrittum","Arthur Gretton","Nicolas Heess","S. M. Ali Eslami","Balaji Lakshminarayanan","Dino Sejdinovic","Zoltán Szabó"],"abstract":"We propose an efficient nonparametric strategy for learning a message\noperator in expectation propagation (EP), which takes as input the set of\nincoming messages to a factor node, and produces an outgoing message as output.\nThis learned operator replaces the multivariate integral required in classical\nEP, which may not have an analytic expression. We use kernel-based regression,\nwhich is trained on a set of probability distributions representing the\nincoming messages, and the associated outgoing messages. The kernel approach\nhas two main advantages: first, it is fast, as it is implemented using a novel\ntwo-layer random feature representation of the input message distributions;\nsecond, it has principled uncertainty estimates, and can be cheaply updated\nonline, meaning it can request and incorporate new training data when it\nencounters inputs on which it is uncertain. In experiments, our approach is\nable to solve learning problems where a single message operator is required for\nmultiple, substantially different data sets (logistic regression for a variety\nof classification problems), where it is essential to accurately assess\nuncertainty and to efficiently and robustly update the message operator.","url_abs":"http://arxiv.org/abs/1503.02551v2","url_pdf":"http://arxiv.org/pdf/1503.02551v2.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":"kernel-based-just-in-time-learning-for","repo_url":"https://github.com/wittawatj/kernel-ep","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"regression-1","task_name":"regression"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}