{"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/personalized-federated-learning-with-gaussian","title":"Personalized Federated Learning with Gaussian Processes","arxiv_id":"2106.15482","date":"2021-06-29","proceeding":"NeurIPS 2021 12","authors":["Idan Achituve","Aviv Shamsian","Aviv Navon","Gal Chechik","Ethan Fetaya"],"abstract":"Federated learning aims to learn a global model that performs well on client devices with limited cross-client communication. Personalized federated learning (PFL) further extends this setup to handle data heterogeneity between clients by learning personalized models. A key challenge in this setting is to learn effectively across clients even though each client has unique data that is often limited in size. Here we present pFedGP, a solution to PFL that is based on Gaussian processes (GPs) with deep kernel learning. GPs are highly expressive models that work well in the low data regime due to their Bayesian nature. However, applying GPs to PFL raises multiple challenges. Mainly, GPs performance depends heavily on access to a good kernel function, and learning a kernel requires a large training set. Therefore, we propose learning a shared kernel function across all clients, parameterized by a neural network, with a personal GP classifier for each client. We further extend pFedGP to include inducing points using two novel methods, the first helps to improve generalization in the low data regime and the second reduces the computational cost. We derive a PAC-Bayes generalization bound on novel clients and empirically show that it gives non-vacuous guarantees. Extensive experiments on standard PFL benchmarks with CIFAR-10, CIFAR-100, and CINIC-10, and on a new setup of learning under input noise show that pFedGP achieves well-calibrated predictions while significantly outperforming baseline methods, reaching up to 21% in accuracy gain.","url_abs":"https://arxiv.org/abs/2106.15482v2","url_pdf":"https://arxiv.org/pdf/2106.15482v2.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":"personalized-federated-learning-with-gaussian","repo_url":"https://github.com/IdanAchituve/pFedGP","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"federated-learning","task_name":"Federated Learning"},{"task_slug":"gaussian-processes","task_name":"Gaussian Processes"},{"task_slug":"personalized-federated-learning","task_name":"Personalized Federated Learning"}],"methods":[{"method_slug":"gps","method_name":"GPS"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/personalized-federated-learning-on-cifar-10","task":"Personalized Federated Learning","dataset":"CIFAR-10","model":"pFedGP-IP-compute","rank_in_archive_order":5,"of":7,"metrics":{"ACC@1-100Clients":"88.8","ACC@1-500Clients":"86.8","ACC@1-50Clients":"89.9"},"uses_additional_data":false},{"leaderboard":"/sota/personalized-federated-learning-on-cifar-10","task":"Personalized Federated Learning","dataset":"CIFAR-10","model":"pFedGP","rank_in_archive_order":6,"of":7,"metrics":{"ACC@1-100Clients":"88.8","ACC@1-500Clients":"87.6","ACC@1-50Clients":"89.2"},"uses_additional_data":false},{"leaderboard":"/sota/personalized-federated-learning-on-cifar-10","task":"Personalized Federated Learning","dataset":"CIFAR-10","model":"pFedGP-IP-data","rank_in_archive_order":7,"of":7,"metrics":{"ACC@1-100Clients":"87.4","ACC@1-500Clients":"86.9","ACC@1-50Clients":"88.6"},"uses_additional_data":false},{"leaderboard":"/sota/personalized-federated-learning-on-cifar-100","task":"Personalized Federated Learning","dataset":"CIFAR-100","model":"pFedGP-IP-data","rank_in_archive_order":1,"of":7,"metrics":{"ACC@1-100Clients":"58.5","ACC@1-500":"55.7","ACC@1-50Clients":"60.2"},"uses_additional_data":false},{"leaderboard":"/sota/personalized-federated-learning-on-cifar-100","task":"Personalized Federated Learning","dataset":"CIFAR-100","model":"pFedGP","rank_in_archive_order":2,"of":7,"metrics":{"ACC@1-100Clients":"61.3","ACC@1-500":"50.6","ACC@1-50Clients":"63.3"},"uses_additional_data":false},{"leaderboard":"/sota/personalized-federated-learning-on-cifar-100","task":"Personalized Federated Learning","dataset":"CIFAR-100","model":"pFedGP-IP-compute","rank_in_archive_order":3,"of":7,"metrics":{"ACC@1-100Clients":"59.8","ACC@1-500":"49.2","ACC@1-50Clients":"61.2"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2106.15482","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}