{"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/universum-prescription-regularization-using","title":"Universum Prescription: Regularization using Unlabeled Data","arxiv_id":"1511.03719","date":"2015-11-11","proceeding":null,"authors":["Xiang Zhang","Yann Lecun"],"abstract":"This paper shows that simply prescribing \"none of the above\" labels to\nunlabeled data has a beneficial regularization effect to supervised learning.\nWe call it universum prescription by the fact that the prescribed labels cannot\nbe one of the supervised labels. In spite of its simplicity, universum\nprescription obtained competitive results in training deep convolutional\nnetworks for CIFAR-10, CIFAR-100, STL-10 and ImageNet datasets. A qualitative\njustification of these approaches using Rademacher complexity is presented. The\neffect of a regularization parameter -- probability of sampling from unlabeled\ndata -- is also studied empirically.","url_abs":"http://arxiv.org/abs/1511.03719v7","url_pdf":"http://arxiv.org/pdf/1511.03719v7.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":[],"tasks":[{"task_slug":"image-classification","task_name":"Image Classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-classification-on-cifar-10","task":"Image Classification","dataset":"CIFAR-10","model":"Universum Prescription","rank_in_archive_order":176,"of":265,"metrics":{"Percentage correct":"93.3"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-cifar-100","task":"Image Classification","dataset":"CIFAR-100","model":"Universum Prescription","rank_in_archive_order":186,"of":211,"metrics":{"Percentage correct":"67.2"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1511.03719","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}