{"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/gradient-sparsification-for-communication","title":"Gradient Sparsification for Communication-Efficient Distributed Optimization","arxiv_id":"1710.09854","date":"2017-10-26","proceeding":"NeurIPS 2018 12","authors":["Jianqiao Wangni","Jialei Wang","Ji Liu","Tong Zhang"],"abstract":"Modern large scale machine learning applications require stochastic\noptimization algorithms to be implemented on distributed computational\narchitectures. A key bottleneck is the communication overhead for exchanging\ninformation such as stochastic gradients among different workers. In this\npaper, to reduce the communication cost we propose a convex optimization\nformulation to minimize the coding length of stochastic gradients. To solve the\noptimal sparsification efficiently, several simple and fast algorithms are\nproposed for approximate solution, with theoretical guaranteed for sparseness.\nExperiments on $\\ell_2$ regularized logistic regression, support vector\nmachines, and convolutional neural networks validate our sparsification\napproaches.","url_abs":"http://arxiv.org/abs/1710.09854v1","url_pdf":"http://arxiv.org/pdf/1710.09854v1.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":"machine-learning","task_name":"BIG-bench Machine Learning"},{"task_slug":"distributed-optimization","task_name":"Distributed Optimization"},{"task_slug":"stochastic-optimization","task_name":"Stochastic Optimization"},{"task_slug":"regression-1","task_name":"regression"}],"methods":[{"method_slug":"gradient-sparsification","method_name":"Gradient Sparsification"}],"datasets_introduced":[],"methods_introduced":[{"slug":"gradient-sparsification","name":"Gradient Sparsification","full_name":"Gradient Sparsification"}],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1710.09854","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}