{"id":1148823,"date":"2025-12-09T11:48:07","date_gmt":"2025-12-09T19:48:07","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/"},"modified":"2025-12-09T12:04:40","modified_gmt":"2025-12-09T20:04:40","slug":"ml-robuste-adaptatif-et-modulaire-montreal","status":"publish","type":"msr-group","link":"https:\/\/www.microsoft.com\/en-us\/research\/theme\/ml-robuste-adaptatif-et-modulaire-montreal\/","title":{"rendered":"ML robuste, adaptatif et modulaire | Montr\u00e9al"},"content":{"rendered":"<section class=\"mb-3 moray-highlight\">\n\t<div class=\"card-img-overlay mx-lg-0\">\n\t\t<div class=\"card-background  has-background- card-background--full-bleed\">\n\t\t\t<img loading=\"lazy\" decoding=\"async\" width=\"3840\" height=\"1440\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2012\/12\/Theme-navy_RL_header_12_2019__1920x720.png\" class=\"attachment-full size-full\" alt=\"MSR theme: Reinforcement Learning Research\" style=\"\" srcset=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2012\/12\/Theme-navy_RL_header_12_2019__1920x720.png 3840w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2012\/12\/Theme-navy_RL_header_12_2019__1920x720-300x113.png 300w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2012\/12\/Theme-navy_RL_header_12_2019__1920x720-1024x384.png 1024w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2012\/12\/Theme-navy_RL_header_12_2019__1920x720-768x288.png 768w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2012\/12\/Theme-navy_RL_header_12_2019__1920x720-1536x576.png 1536w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2012\/12\/Theme-navy_RL_header_12_2019__1920x720-2048x768.png 2048w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2012\/12\/Theme-navy_RL_header_12_2019__1920x720-1920x720.png 1920w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2012\/12\/Theme-navy_RL_header_12_2019__1920x720-1600x600.png 1600w\" sizes=\"auto, (max-width: 3840px) 100vw, 3840px\" \/>\t\t<\/div>\n\t\t<!-- Foreground -->\n\t\t<div class=\"card-foreground d-flex mt-md-n5 my-lg-5 px-g px-lg-0\">\n\t\t\t<!-- Container -->\n\t\t\t<div class=\"container d-flex mt-md-n5 my-lg-5 \">\n\t\t\t\t<!-- Card wrapper -->\n\t\t\t\t<div class=\"w-100 w-lg-col-5\">\n\t\t\t\t\t<!-- Card -->\n\t\t\t\t\t<div class=\"card material-md-card py-5 px-md-5\">\n\t\t\t\t\t\t<div class=\"card-body \">\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/lab\/laboratoire-de-recherche-microsoft-montreal\/\" class=\"icon-link icon-link--reverse mb-2\" data-bi-cN=\"Retour au Laboratoire de recherche Microsoft \u2013 Montr\u00e9al\">\n\t\t\t\t\t\t\t\t\t<span class=\"c-glyph glyph-chevron-left\" aria-hidden=\"true\"><\/span>\n\t\t\t\t\t\t\t\t\tRetour au Laboratoire de recherche Microsoft \u2013 Montr\u00e9al\t\t\t\t\t\t\t\t<\/a>\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\n\n<h1 class=\"wp-block-heading\" id=\"ml-robuste-adaptatif-et-modulaire-montreal\">ML robuste, adaptatif et modulaire | Montr\u00e9al<\/h1>\n\n\n\n<p><\/p>\n\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t<\/div>\n\t\t<\/div>\n\t<\/div>\n<\/section>\n\n\n\n\n\n<p>Notre objectif est de comprendre les principes qui sous-tendent l&#8217;apprentissage et la g\u00e9n\u00e9ralisation, afin de cr\u00e9er des syst\u00e8mes d&#8217;IA fiables capables d&#8217;apprendre plus efficacement \u00e0 partir des donn\u00e9es disponibles, de collecter intelligemment des donn\u00e9es pertinentes suppl\u00e9mentaires, et de s&#8217;adapter et de raisonner rapidement sur des sc\u00e9narios inhabituels lorsqu&#8217;ils sont d\u00e9ploy\u00e9s dans la nature.<\/p>\n\n\n\n<p>Les syst\u00e8mes d\u2019IA d\u00e9ploy\u00e9s dans la nature sont souvent expos\u00e9s \u00e0 un flux d\u2019exemples et de sc\u00e9narios nouveaux qui peuvent diff\u00e9rer consid\u00e9rablement de ceux observ\u00e9s lors de la formation. Il est important qu\u2019un mod\u00e8le donne le plus de sens \u00e0 ces situations pour apporter des r\u00e9ponses robustes dans ces nouveaux contextes.<\/p>\n\n\n\n<p>Voici quelques-unes des pistes de recherche explor\u00e9es sous ce th\u00e8me :<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>D\u00e9finition d\u2019exemples pertinents pour le transfert de t\u00e2ches et l\u2019augmentation de la r\u00e9sistance aux fausses corr\u00e9lations<\/li>\n\n\n\n<li>Quantification de l\u2019incertitude et raisonnement dans des conditions incertaines, par exemple dans un contexte d\u2019apprentissage par renforcement hors ligne<\/li>\n\n\n\n<li>D\u00e9composition et r\u00e9utilisation du mod\u00e8le, o\u00f9 un mod\u00e8le repr\u00e9sente une combinaison de modules plus petits, chacun pouvant \u00eatre r\u00e9utilis\u00e9 pour une t\u00e2che diff\u00e9rente, ce qui facilite le transfert de connaissances<\/li>\n\n\n\n<li>Apprentissage de repr\u00e9sentations factoris\u00e9es et causales pour les images, le texte et les donn\u00e9es m\u00e9dicales<\/li>\n\n\n\n<li>Exemples de m\u00e9thodes d\u2019optimisation efficaces pour une adaptation rapide<\/li>\n<\/ul>\n\n\n\n<div style=\"height:25px\" aria-hidden=\"true\" class=\"wp-block-spacer\"><\/div>\n\n\n","protected":false},"excerpt":{"rendered":"<p>Notre objectif est de comprendre les principes qui sous-tendent l&#8217;apprentissage et la g\u00e9n\u00e9ralisation, afin de cr\u00e9er des syst\u00e8mes d&#8217;IA fiables capables d&#8217;apprendre plus efficacement \u00e0 partir des donn\u00e9es disponibles, de collecter intelligemment des donn\u00e9es pertinentes suppl\u00e9mentaires, et de s&#8217;adapter et de raisonner rapidement sur des sc\u00e9narios inhabituels lorsqu&#8217;ils sont d\u00e9ploy\u00e9s dans la nature. Les syst\u00e8mes [&hellip;]<\/p>\n","protected":false},"featured_media":627639,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":true,"_classifai_error":"","msr_group_start":"","footnotes":""},"research-area":[13561,13556,13546],"msr-group-type":[243688],"msr-locale":[268878],"msr-impact-theme":[],"class_list":["post-1148823","msr-group","type-msr-group","status-publish","has-post-thumbnail","hentry","msr-research-area-algorithms","msr-research-area-artificial-intelligence","msr-research-area-computational-sciences-mathematics","msr-group-type-theme","msr-locale-fr_ca"],"msr_group_start":"","msr_detailed_description":"","msr_further_details":"","msr_hero_images":[],"msr_research_lab":[1148609],"related-researchers":[{"type":"user_nicename","display_name":"Friederike Niedtner","user_id":39919,"people_section":"Section name 0","alias":"fniedtner"},{"type":"user_nicename","display_name":"Alessandro Sordoni","user_id":37230,"people_section":"Section name 0","alias":"alsordon"}],"related-publications":[757882,797209,797176,796861,796579,782530,771001,760240,758332,821704,757876,749932,747034,704842,694086,684684,677115,826714,1135961,1097337,1097325,852000,844159,843406,832471,831397,659424,826708,826702,826693,826684,826675,824389,821743,580795,600384,596701,595504,595489,584533,580996,580987,580978,610218,577473,556338,549297,487844,487835,487826,481131,455568,620541,620553,622881,627489,629289,629307,641892,643860,643866,656004,659082,659091,659097,659103,659109],"related-downloads":[],"related-videos":[885540],"related-projects":[852753,852783,615297],"related-events":[],"related-opportunities":[],"related-posts":[1136154],"tab-content":[],"msr_impact_theme":[],"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-group\/1148823","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-group"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-group"}],"version-history":[{"count":4,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-group\/1148823\/revisions"}],"predecessor-version":[{"id":1158035,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-group\/1148823\/revisions\/1158035"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media\/627639"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=1148823"}],"wp:term":[{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=1148823"},{"taxonomy":"msr-group-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-group-type?post=1148823"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=1148823"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=1148823"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}