{"id":797,"date":"2019-08-22T15:50:16","date_gmt":"2019-08-22T13:50:16","guid":{"rendered":"https:\/\/project.inria.fr\/epfl-Inria\/?page_id=797"},"modified":"2022-09-07T16:05:07","modified_gmt":"2022-09-07T14:05:07","slug":"performance-study-in-the-computation-of-distributed-generative-adversarial-networks","status":"publish","type":"page","link":"https:\/\/project.inria.fr\/epfl-Inria\/performance-study-in-the-computation-of-distributed-generative-adversarial-networks\/","title":{"rendered":"Performance study in the computation of distributed generative adversarial networks"},"content":{"rendered":"<h3><span style=\"color: #ff0000;\"><strong>Performance study in the computation of distributed generative<br \/>\nadversarial networks<\/strong><\/span><\/h3>\n<p><strong>Principal investigators<\/strong><br \/>\n<a href=\"https:\/\/aguirguis.netlify.app\/\">Guirguis Arsany<\/a>,Ph.D. student, <a href=\"https:\/\/dcl.epfl.ch\/site\/\">DCL lab<\/a>, Ecole Polytechnique F\u00e9d\u00e9rale de Lausanne<br \/>\n<a href=\"https:\/\/erwanlemerrer.github.io\">Erwan Le Merrer<\/a>, Doctor, <a href=\"https:\/\/team.inria.fr\/wide\/\">WIDE<\/a> research team, Inria<\/p>\n<p style=\"text-align: justify;\"><strong>Abstract<\/strong><br \/>\nThe objective of this collaboration is to investigate the efficient distributed computation of generative adversarial networks (GANs) over a set of client devices. In particular, we will study the implementation of GANs in the setup of federated learning, where a server is leveraged as a central point for model synchronization.<br \/>\nIn this model, the client data can remain on their devices, increasing privacy w.r.t. approaches that require the collection of all the data on a single location. Several angles are of interest in this collaboration:<br \/>\n(i) the fault tolerance aspect as not been studied in a constructive way, which leaves space for algorithms allowing robust learning.<br \/>\n(ii) The training of GANs is data intensive, as well as compute-intensive; the computation-communication trade-off that arise in this approach is yet to be understood, for an accurate comparison with central, or fully decentralized approaches (such as gossip-based ones for instance).<br \/>\n(iii) The intrinsic difference of GANs w.r.t. regular deep learning models, when it comes to distributed computation, is to be fully qualified for efficient proposals.<\/p>\n<p><strong>Website<\/strong>:\u00a0under construction<a><br \/>\n<span style=\"color: #000000;\"><span style=\"color: #333333;\"><strong>Keywords<\/strong><\/span>:<\/span> <span style=\"color: #808080;\">Machine learning, federated learning, generative adversarial networks, distributed computing, fault tolerance<\/span><\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Performance study in the computation of distributed generative adversarial networks Principal investigators Guirguis Arsany,Ph.D. student, DCL lab, Ecole Polytechnique F\u00e9d\u00e9rale de Lausanne Erwan Le Merrer, Doctor, WIDE research team, Inria[&#8230;]<\/p>\n<p> <a class=\"continue-reading-link\" href=\"https:\/\/project.inria.fr\/epfl-Inria\/performance-study-in-the-computation-of-distributed-generative-adversarial-networks\/\"><span>Continue reading<\/span><i class=\"crycon-right-dir\"><\/i><\/a> <\/p>\n","protected":false},"author":309,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"_mc_calendar":[],"footnotes":""},"class_list":["post-797","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/project.inria.fr\/epfl-Inria\/wp-json\/wp\/v2\/pages\/797","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/project.inria.fr\/epfl-Inria\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/project.inria.fr\/epfl-Inria\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/project.inria.fr\/epfl-Inria\/wp-json\/wp\/v2\/users\/309"}],"replies":[{"embeddable":true,"href":"https:\/\/project.inria.fr\/epfl-Inria\/wp-json\/wp\/v2\/comments?post=797"}],"version-history":[{"count":8,"href":"https:\/\/project.inria.fr\/epfl-Inria\/wp-json\/wp\/v2\/pages\/797\/revisions"}],"predecessor-version":[{"id":1046,"href":"https:\/\/project.inria.fr\/epfl-Inria\/wp-json\/wp\/v2\/pages\/797\/revisions\/1046"}],"wp:attachment":[{"href":"https:\/\/project.inria.fr\/epfl-Inria\/wp-json\/wp\/v2\/media?parent=797"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}