{"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/adversarial-constraint-learning-for","title":"Adversarial Constraint Learning for Structured Prediction","arxiv_id":"1805.10561","date":"2018-05-27","proceeding":null,"authors":["Hongyu Ren","Russell Stewart","Jiaming Song","Volodymyr Kuleshov","Stefano Ermon"],"abstract":"Constraint-based learning reduces the burden of collecting labels by having\nusers specify general properties of structured outputs, such as constraints\nimposed by physical laws. We propose a novel framework for simultaneously\nlearning these constraints and using them for supervision, bypassing the\ndifficulty of using domain expertise to manually specify constraints. Learning\nrequires a black-box simulator of structured outputs, which generates valid\nlabels, but need not model their corresponding inputs or the input-label\nrelationship. At training time, we constrain the model to produce outputs that\ncannot be distinguished from simulated labels by adversarial training.\nProviding our framework with a small number of labeled inputs gives rise to a\nnew semi-supervised structured prediction model; we evaluate this model on\nmultiple tasks --- tracking, pose estimation and time series prediction --- and\nfind that it achieves high accuracy with only a small number of labeled inputs.\nIn some cases, no labels are required at all.","url_abs":"http://arxiv.org/abs/1805.10561v2","url_pdf":"http://arxiv.org/pdf/1805.10561v2.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":"adversarial-constraint-learning-for","repo_url":"https://github.com/hyren/acl","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"pose-estimation","task_name":"Pose Estimation"},{"task_slug":"prediction","task_name":"Prediction"},{"task_slug":"structured-prediction","task_name":"Structured Prediction"},{"task_slug":"time-series-1","task_name":"Time Series"},{"task_slug":"time-series","task_name":"Time Series Analysis"},{"task_slug":"time-series-prediction","task_name":"Time Series Prediction"},{"task_slug":null,"task_name":"valid"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}