{"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/classification-without-labels-learning-from","title":"Classification without labels: Learning from mixed samples in high energy physics","arxiv_id":"1708.02949","date":"2017-08-09","proceeding":null,"authors":["Eric M. Metodiev","Benjamin Nachman","Jesse Thaler"],"abstract":"Modern machine learning techniques can be used to construct powerful models\nfor difficult collider physics problems. In many applications, however, these\nmodels are trained on imperfect simulations due to a lack of truth-level\ninformation in the data, which risks the model learning artifacts of the\nsimulation. In this paper, we introduce the paradigm of classification without\nlabels (CWoLa) in which a classifier is trained to distinguish statistical\nmixtures of classes, which are common in collider physics. Crucially, neither\nindividual labels nor class proportions are required, yet we prove that the\noptimal classifier in the CWoLa paradigm is also the optimal classifier in the\ntraditional fully-supervised case where all label information is available.\nAfter demonstrating the power of this method in an analytical toy example, we\nconsider a realistic benchmark for collider physics: distinguishing quark-\nversus gluon-initiated jets using mixed quark/gluon training samples. More\ngenerally, CWoLa can be applied to any classification problem where labels or\nclass proportions are unknown or simulations are unreliable, but statistical\nmixtures of the classes are available.","url_abs":"http://arxiv.org/abs/1708.02949v3","url_pdf":"http://arxiv.org/pdf/1708.02949v3.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":"classification-without-labels-learning-from","repo_url":"https://github.com/hep-lbdl/gaiacwola","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"high","task_name":"Vocal Bursts Intensity Prediction"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1708.02949","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}