{"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-unified-batch-online-learning-framework-for","title":"A Unified Batch Online Learning Framework for Click Prediction","arxiv_id":"1809.04673","date":"2018-09-12","proceeding":null,"authors":["Rishabh Iyer","Nimit Acharya","Tanuja Bompada","Denis Charles","Eren Manavoglu"],"abstract":"We present a unified framework for Batch Online Learning (OL) for Click\nPrediction in Search Advertisement. Machine Learning models once deployed, show\nnon-trivial accuracy and calibration degradation over time due to model\nstaleness. It is therefore necessary to regularly update models, and do so\nautomatically. This paper presents two paradigms of Batch Online Learning, one\nwhich incrementally updates the model parameters via an early stopping\nmechanism, and another which does so through a proximal regularization. We\nargue how both these schemes naturally trade-off between old and new data. We\nthen theoretically and empirically show that these two seemingly different\nschemes are closely related. Through extensive experiments, we demonstrate the\nutility of of our OL framework; how the two OL schemes relate to each other and\nhow they trade-off between the new and historical data. We then compare batch\nOL to full model retrains, and show how online learning is more robust to data\nissues. We also demonstrate the long term impact of Online Learning, the role\nof the initial Models in OL, the impact of delays in the update, and finally\nconclude with some implementation details and challenges in deploying a real\nworld online learning system in production. While this paper mostly focuses on\napplication of click prediction for search advertisement, we hope that the\nlessons learned here can be carried over to other problem domains.","url_abs":"http://arxiv.org/abs/1809.04673v1","url_pdf":"http://arxiv.org/pdf/1809.04673v1.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-unified-batch-online-learning-framework-for","repo_url":"https://github.com/rishabhk108/jensen-ol","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[],"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}