{"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-smoother-way-to-train-structured-prediction","title":"A Smoother Way to Train Structured Prediction Models","arxiv_id":"1902.03228","date":"2019-02-08","proceeding":"NeurIPS 2018 12","authors":["Krishna Pillutla","Vincent Roulet","Sham M. Kakade","Zaid Harchaoui"],"abstract":"We present a framework to train a structured prediction model by performing\nsmoothing on the inference algorithm it builds upon. Smoothing overcomes the\nnon-smoothness inherent to the maximum margin structured prediction objective,\nand paves the way for the use of fast primal gradient-based optimization\nalgorithms. We illustrate the proposed framework by developing a novel primal\nincremental optimization algorithm for the structural support vector machine.\nThe proposed algorithm blends an extrapolation scheme for acceleration and an\nadaptive smoothing scheme and builds upon the stochastic variance-reduced\ngradient algorithm. We establish its worst-case global complexity bound and\nstudy several practical variants, including extensions to deep structured\nprediction. We present experimental results on two real-world problems, namely\nnamed entity recognition and visual object localization. The experimental\nresults show that the proposed framework allows us to build upon efficient\ninference algorithms to develop large-scale optimization algorithms for\nstructured prediction which can achieve competitive performance on the two\nreal-world problems.","url_abs":"http://arxiv.org/abs/1902.03228v1","url_pdf":"http://arxiv.org/pdf/1902.03228v1.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-smoother-way-to-train-structured-prediction","repo_url":"https://github.com/krishnap25/casimir","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok","spdx":"GPL-3.0"}}],"tasks":[{"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":"object-localization","task_name":"Object Localization"},{"task_slug":"prediction","task_name":"Prediction"},{"task_slug":"structured-prediction","task_name":"Structured Prediction"},{"task_slug":"named-entity-recognition","task_name":"named-entity-recognition"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1902.03228","atlas_url":"https://app.syntology.ai/?focus=1902.03228","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}