Imaginez un syst\u00e8me de reconnaissance d’images qui identifie une photo de panda. Qu’est-ce qu’une Attaque adversariale ? C’est une modification subtile, souvent imperceptible \u00e0 l’\u0153il humain, apport\u00e9e \u00e0 cette image (le panda) qui trompe le syst\u00e8me et lui fait identifier autre chose, par exemple un gibbon.<\/p>\n
Les attaques adversariales exploitent les faiblesses des mod\u00e8les d’IA en introduisant de petites perturbations, appel\u00e9es \u00ab bruit adversarial \u00bb, dans les donn\u00e9es d’entr\u00e9e. Ce bruit, bien que minime, peut radicalement modifier l’interpr\u00e9tation du mod\u00e8le. Reprenons l’exemple du panda : quelques pixels modifi\u00e9s, invisibles pour nous, peuvent suffire \u00e0 le faire classer comme un gibbon par le syst\u00e8me. C’est comme ajouter un grain de sable, invisible \u00e0 l’\u0153il nu, dans les rouages d’une horloge pour la d\u00e9r\u00e9gler compl\u00e8tement.<\/p>\n
Les attaques adversariales repr\u00e9sentent une menace importante pour la s\u00e9curit\u00e9 et la fiabilit\u00e9 des syst\u00e8mes d’IA. Imaginez les cons\u00e9quences pour des voitures autonomes, o\u00f9 une perturbation sur un panneau de signalisation pourrait amener le v\u00e9hicule \u00e0 mal interpr\u00e9ter les instructions. En prompt engineering, ces attaques peuvent \u00eatre utilis\u00e9es pour manipuler la g\u00e9n\u00e9ration de texte, en introduisant des mots ou des phrases qui, bien que subtils, orientent le mod\u00e8le vers des r\u00e9sultats ind\u00e9sirables. Par exemple, un prompt l\u00e9g\u00e8rement modifi\u00e9 pourrait amener un chatbot \u00e0 formuler des r\u00e9ponses biais\u00e9es ou inappropri\u00e9es.<\/p>\n
Imaginez un syst\u00e8me de reconnaissance d’images qui identifie une photo de panda. Qu’est-ce qu’une Attaque adversariale ? C’est une modification subtile, souvent imperceptible \u00e0 l’\u0153il humain, apport\u00e9e \u00e0 cette image (le panda) qui trompe le syst\u00e8me et lui fait identifier autre chose, par exemple un gibbon. Comment fonctionnent les Attaques adversariales ? Les attaques adversariales […]<\/p>\n","protected":false},"author":0,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_uag_custom_page_level_css":"","site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"default","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"categories":[3],"tags":[434,44,75,12,155,72],"class_list":["post-824","post","type-post","status-publish","format-standard","hentry","category-a","tag-adversarial-examples","tag-apprentissage-automatique","tag-attaques-adversariales","tag-prompt-engineering","tag-robustesse","tag-securite-de-lia"],"uagb_featured_image_src":{"full":false,"thumbnail":false,"medium":false,"medium_large":false,"large":false,"1536x1536":false,"2048x2048":false},"uagb_author_info":{"display_name":"","author_link":"https:\/\/happynumeric.com\/lexique-intelligence-artificielle\/author\/"},"uagb_comment_info":0,"uagb_excerpt":"Imaginez un syst\u00e8me de reconnaissance d’images qui identifie une photo de panda. Qu’est-ce qu’une Attaque adversariale ? C’est une modification subtile, souvent imperceptible \u00e0 l’\u0153il humain, apport\u00e9e \u00e0 cette image (le panda) qui trompe le syst\u00e8me et lui fait identifier autre chose, par exemple un gibbon. Comment fonctionnent les Attaques adversariales ? Les attaques adversariales\u2026","_links":{"self":[{"href":"https:\/\/happynumeric.com\/lexique-intelligence-artificielle\/wp-json\/wp\/v2\/posts\/824","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/happynumeric.com\/lexique-intelligence-artificielle\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/happynumeric.com\/lexique-intelligence-artificielle\/wp-json\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/happynumeric.com\/lexique-intelligence-artificielle\/wp-json\/wp\/v2\/comments?post=824"}],"version-history":[{"count":1,"href":"https:\/\/happynumeric.com\/lexique-intelligence-artificielle\/wp-json\/wp\/v2\/posts\/824\/revisions"}],"predecessor-version":[{"id":942,"href":"https:\/\/happynumeric.com\/lexique-intelligence-artificielle\/wp-json\/wp\/v2\/posts\/824\/revisions\/942"}],"wp:attachment":[{"href":"https:\/\/happynumeric.com\/lexique-intelligence-artificielle\/wp-json\/wp\/v2\/media?parent=824"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/happynumeric.com\/lexique-intelligence-artificielle\/wp-json\/wp\/v2\/categories?post=824"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/happynumeric.com\/lexique-intelligence-artificielle\/wp-json\/wp\/v2\/tags?post=824"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}