{"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/a-unifying-bayesian-view-of-continual","title":"A Unifying Bayesian View of Continual Learning","arxiv_id":"1902.06494","date":"2019-02-18","proceeding":null,"authors":["Sebastian Farquhar","Yarin Gal"],"abstract":"Some machine learning applications require continual learning - where data\ncomes in a sequence of datasets, each is used for training and then permanently\ndiscarded. From a Bayesian perspective, continual learning seems\nstraightforward: Given the model posterior one would simply use this as the\nprior for the next task. However, exact posterior evaluation is intractable\nwith many models, especially with Bayesian neural networks (BNNs). Instead,\nposterior approximations are often sought. Unfortunately, when posterior\napproximations are used, prior-focused approaches do not succeed in evaluations\ndesigned to capture properties of realistic continual learning use cases. As an\nalternative to prior-focused methods, we introduce a new approximate Bayesian\nderivation of the continual learning loss. Our loss does not rely on the\nposterior from earlier tasks, and instead adapts the model itself by changing\nthe likelihood term. We call these approaches likelihood-focused. We then\ncombine prior- and likelihood-focused methods into one objective, tying the two\nviews together under a single unifying framework of approximate Bayesian\ncontinual learning.","url_abs":"http://arxiv.org/abs/1902.06494v1","url_pdf":"http://arxiv.org/pdf/1902.06494v1.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":"a-unifying-bayesian-view-of-continual","repo_url":"https://github.com/Saraharas/A-Unifying-Bayesian-View-of-Continual-Learning","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"a-unifying-bayesian-view-of-continual","repo_url":"https://github.com/Saraharas/Continual-learning","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"continual-learning","task_name":"Continual Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1902.06494","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}