{"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/on-the-job-learning-with-bayesian-decision","title":"On-the-Job Learning with Bayesian Decision Theory","arxiv_id":"1506.03140","date":"2015-06-10","proceeding":"NeurIPS 2015 12","authors":["Keenon Werling","Arun Chaganty","Percy Liang","Chris Manning"],"abstract":"Our goal is to deploy a high-accuracy system starting with zero training\nexamples. We consider an \"on-the-job\" setting, where as inputs arrive, we use\nreal-time crowdsourcing to resolve uncertainty where needed and output our\nprediction when confident. As the model improves over time, the reliance on\ncrowdsourcing queries decreases. We cast our setting as a stochastic game based\non Bayesian decision theory, which allows us to balance latency, cost, and\naccuracy objectives in a principled way. Computing the optimal policy is\nintractable, so we develop an approximation based on Monte Carlo Tree Search.\nWe tested our approach on three datasets---named-entity recognition, sentiment\nclassification, and image classification. On the NER task we obtained more than\nan order of magnitude reduction in cost compared to full human annotation,\nwhile boosting performance relative to the expert provided labels. We also\nachieve a 8% F1 improvement over having a single human label the whole set, and\na 28% F1 improvement over online learning.","url_abs":"http://arxiv.org/abs/1506.03140v2","url_pdf":"http://arxiv.org/pdf/1506.03140v2.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":"on-the-job-learning-with-bayesian-decision","repo_url":"https://github.com/keenon/lense","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"on-the-job-learning-with-bayesian-decision","repo_url":"https://worksheets.codalab.org/worksheets/0x2ae89944846444539c2d08a0b7ff3f6f","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"cg","task_name":"NER"},{"task_slug":"named-entity-recognition-1","task_name":"Named Entity Recognition"},{"task_slug":"named-entity-recognition-ner","task_name":"Named Entity Recognition (NER)"},{"task_slug":"sentiment-analysis","task_name":"Sentiment Analysis"},{"task_slug":"sentiment-classification","task_name":"Sentiment Classification"},{"task_slug":"image-classification","task_name":"image-classification"},{"task_slug":"named-entity-recognition","task_name":"named-entity-recognition"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1506.03140","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}