{"id":2680,"date":"2023-07-18T20:48:46","date_gmt":"2023-07-18T20:48:46","guid":{"rendered":"https:\/\/dailyai.com\/?p=2680"},"modified":"2023-07-18T21:23:28","modified_gmt":"2023-07-18T21:23:28","slug":"ai-image-generators-quality-suffers-when-trained-on-ai-generated-data","status":"publish","type":"post","link":"https:\/\/dailyai.com\/fr\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/","title":{"rendered":"La qualit\u00e9 des g\u00e9n\u00e9rateurs d'images d'IA diminue lorsqu'ils sont form\u00e9s sur des donn\u00e9es g\u00e9n\u00e9r\u00e9es par l'IA"},"content":{"rendered":"

Une nouvelle \u00e9tude montre que l'entra\u00eenement des g\u00e9n\u00e9rateurs d'images par l'IA avec des images g\u00e9n\u00e9r\u00e9es par l'IA conduit finalement \u00e0 une r\u00e9duction significative de la qualit\u00e9 des r\u00e9sultats.\u00a0<\/b><\/p>\n

Baraniuk et son \u00e9quipe ont d\u00e9montr\u00e9 comment cette boucle d'entra\u00eenement probl\u00e9matique affecte les IA g\u00e9n\u00e9ratives, notamment les mod\u00e8les StyleGAN et de diffusion. Ces derniers font partie des mod\u00e8les utilis\u00e9s pour les g\u00e9n\u00e9rateurs d'images d'IA tels que Stable Diffusion, DALL-E et MidJourney.\u00a0<\/span><\/p>\n

Dans leur exp\u00e9rience<\/a>L'\u00e9quipe a entra\u00een\u00e9 les IA sur des images g\u00e9n\u00e9r\u00e9es par l'IA ou sur des images r\u00e9elles. 70 000 visages humains r\u00e9els provenant de Flickr.<\/span><\/p>\n

Lorsque chaque IA a \u00e9t\u00e9 entra\u00een\u00e9e sur ses propres images, les sorties du g\u00e9n\u00e9rateur d'images StyleGAN ont commenc\u00e9 \u00e0 afficher des motifs visuels d\u00e9form\u00e9s et ondul\u00e9s, tandis que les sorties du g\u00e9n\u00e9rateur d'images de diffusion sont devenues plus floues.<\/span><\/p>\n

Dans les deux cas, la formation d'IA sur des images g\u00e9n\u00e9r\u00e9es par l'IA a entra\u00een\u00e9 une perte de qualit\u00e9.\u00a0<\/span><\/p>\n

L'un des <\/span>\u00e9tude<\/span> Richard Baraniuk, de l'universit\u00e9 Rice au Texas, met en garde les auteurs de l'\u00e9tude : \"L'utilisation de donn\u00e9es synth\u00e9tiques, volontairement ou involontairement, va devenir une pente glissante.\"<\/span><\/p>\n

\"G\u00e9n\u00e9rateur
Plus les mod\u00e8les d'IA sont entra\u00een\u00e9s sur des images synth\u00e9tiques, plus leurs r\u00e9sultats deviennent flous. Source : Universit\u00e9 de Rice<\/a>.<\/figcaption><\/figure>\n

Bien que la baisse de la qualit\u00e9 des images ait \u00e9t\u00e9 att\u00e9nu\u00e9e par la s\u00e9lection d'images de meilleure qualit\u00e9 g\u00e9n\u00e9r\u00e9es par l'IA pour l'entra\u00eenement, cela a entra\u00een\u00e9 une perte de diversit\u00e9 des images.\u00a0<\/span><\/p>\n

Les chercheurs ont \u00e9galement essay\u00e9 d'incorporer un ensemble fixe d'images r\u00e9elles dans des ensembles d'entra\u00eenement comprenant principalement des images g\u00e9n\u00e9r\u00e9es par l'IA, une m\u00e9thode parfois utilis\u00e9e pour compl\u00e9ter de petits ensembles d'entra\u00eenement.\u00a0<\/span><\/p>\n

Cependant, cela n'a fait que retarder le d\u00e9clin de la qualit\u00e9 des images - il semble in\u00e9vitable que plus les donn\u00e9es g\u00e9n\u00e9r\u00e9es par l'IA entrent dans les ensembles de donn\u00e9es d'entra\u00eenement, plus le r\u00e9sultat devient mauvais. Il s'agit simplement de savoir quand.<\/span><\/p>\n

Des r\u00e9sultats raisonnables ont \u00e9t\u00e9 obtenus lorsque chaque IA a \u00e9t\u00e9 entra\u00een\u00e9e sur un m\u00e9lange d'images g\u00e9n\u00e9r\u00e9es par l'IA et d'un ensemble d'images authentiques en constante \u00e9volution. Cela a permis de maintenir la qualit\u00e9 et la diversit\u00e9 des images.<\/span><\/p>\n

Il est difficile de trouver un \u00e9quilibre entre quantit\u00e9 et qualit\u00e9 - les images synth\u00e9tiques sont potentiellement illimit\u00e9es par rapport aux images r\u00e9elles, mais leur utilisation a un co\u00fbt.\u00a0<\/span><\/p>\n

Les IA sont \u00e0 court de donn\u00e9es<\/span><\/h2>\n

Les IA sont avides de donn\u00e9es, mais les donn\u00e9es authentiques et de qualit\u00e9 sont une ressource limit\u00e9e.\u00a0<\/span><\/p>\n

Les r\u00e9sultats de cette recherche font \u00e9cho <\/span>\u00e9tudes similaires pour la g\u00e9n\u00e9ration de textes<\/span><\/a>o\u00f9 les r\u00e9sultats de l'IA ont tendance \u00e0 souffrir lorsque les mod\u00e8les sont form\u00e9s sur du texte g\u00e9n\u00e9r\u00e9 par l'IA.\u00a0<\/span><\/p>\n

Les chercheurs soulignent que les petites organisations ayant une capacit\u00e9 limit\u00e9e \u00e0 collecter des donn\u00e9es authentiques sont confront\u00e9es aux plus grandes difficult\u00e9s pour filtrer les images g\u00e9n\u00e9r\u00e9es par l'IA de leurs ensembles de donn\u00e9es.\u00a0<\/span><\/p>\n

En outre, le probl\u00e8me est aggrav\u00e9 par le fait que l'internet est inond\u00e9 de contenus g\u00e9n\u00e9r\u00e9s par l'IA, ce qui rend extr\u00eamement difficile la d\u00e9termination du type de donn\u00e9es sur lesquelles les mod\u00e8les sont form\u00e9s.\u00a0<\/span><\/p>\n

Sina Alemohammad, de l'universit\u00e9 de Rice, sugg\u00e8re que le d\u00e9veloppement de filigranes permettant d'identifier les images g\u00e9n\u00e9r\u00e9es par l'IA pourrait \u00eatre utile, mais pr\u00e9vient que le fait de ne pas tenir compte des filigranes cach\u00e9s peut d\u00e9grader la qualit\u00e9 des images g\u00e9n\u00e9r\u00e9es par l'IA.<\/span><\/p>\n

Alemohammad conclut : \"Vous \u00eates condamn\u00e9 si vous le faites et condamn\u00e9 si vous ne le faites pas. Mais il vaut mieux filigraner l'image que de ne pas le faire\".<\/span><\/p>\n

Les cons\u00e9quences \u00e0 long terme de la consommation des r\u00e9sultats de l'IA font l'objet de vifs d\u00e9bats, mais pour l'heure, les d\u00e9veloppeurs d'IA doivent trouver des solutions pour garantir la qualit\u00e9 de leurs mod\u00e8les. <\/span><\/p>","protected":false},"excerpt":{"rendered":"

Une nouvelle \u00e9tude montre que l'entra\u00eenement des g\u00e9n\u00e9rateurs d'images d'IA avec des images g\u00e9n\u00e9r\u00e9es par l'IA conduit finalement \u00e0 une r\u00e9duction significative de la qualit\u00e9 des r\u00e9sultats. Baraniuk et son \u00e9quipe ont d\u00e9montr\u00e9 comment cette boucle d'entra\u00eenement probl\u00e9matique affecte les IA g\u00e9n\u00e9ratives, notamment les mod\u00e8les StyleGAN et de diffusion. Ces derniers font partie des mod\u00e8les utilis\u00e9s pour les g\u00e9n\u00e9rateurs d'images d'IA tels que Stable Diffusion, DALL-E et MidJourney. Dans son exp\u00e9rience, l'\u00e9quipe a entra\u00een\u00e9 les IA sur des images g\u00e9n\u00e9r\u00e9es par l'IA ou sur des images r\u00e9elles. 70 000 visages humains r\u00e9els provenant de Flickr. Lorsque chaque IA a \u00e9t\u00e9 entra\u00een\u00e9e sur ses propres images g\u00e9n\u00e9r\u00e9es par l'IA, les sorties du g\u00e9n\u00e9rateur d'images StyleGAN ont commenc\u00e9 \u00e0 afficher des motifs visuels d\u00e9form\u00e9s et ondul\u00e9s,<\/p>","protected":false},"author":2,"featured_media":2682,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[84],"tags":[233,234,174],"class_list":["post-2680","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-industry","tag-image-generator","tag-midjourney","tag-stable-diffusion"],"yoast_head":"\nAI image generators' quality suffers when trained on AI-generated data | DailyAI<\/title>\n<meta name=\"description\" content=\"A new study shows that training AI image generators with AI-generated images could lead to a self-consuming loop.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/dailyai.com\/fr\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/\" \/>\n<meta property=\"og:locale\" content=\"fr_FR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"AI image generators' quality suffers when trained on AI-generated data | DailyAI\" \/>\n<meta property=\"og:description\" content=\"A new study shows that training AI image generators with AI-generated images could lead to a self-consuming loop.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/dailyai.com\/fr\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/\" \/>\n<meta property=\"og:site_name\" content=\"DailyAI\" \/>\n<meta property=\"article:published_time\" content=\"2023-07-18T20:48:46+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2023-07-18T21:23:28+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/07\/shutterstock_2252933361.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"1000\" \/>\n\t<meta property=\"og:image:height\" content=\"655\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"Sam Jeans\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@DailyAIOfficial\" \/>\n<meta name=\"twitter:site\" content=\"@DailyAIOfficial\" \/>\n<meta name=\"twitter:label1\" content=\"\u00c9crit par\" \/>\n\t<meta name=\"twitter:data1\" content=\"Sam Jeans\" \/>\n\t<meta name=\"twitter:label2\" content=\"Dur\u00e9e de lecture estim\u00e9e\" \/>\n\t<meta name=\"twitter:data2\" content=\"3 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"NewsArticle\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/07\\\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/07\\\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\\\/\"},\"author\":{\"name\":\"Sam Jeans\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/#\\\/schema\\\/person\\\/711e81f945549438e8bbc579efdeb3c9\"},\"headline\":\"AI image generators’ quality suffers when trained on AI-generated data\",\"datePublished\":\"2023-07-18T20:48:46+00:00\",\"dateModified\":\"2023-07-18T21:23:28+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/07\\\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\\\/\"},\"wordCount\":498,\"publisher\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/07\\\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2023\\\/07\\\/shutterstock_2252933361.jpg\",\"keywords\":[\"Image generator\",\"MidJourney\",\"Stable Diffusion\"],\"articleSection\":[\"Industry\"],\"inLanguage\":\"fr-FR\"},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/07\\\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\\\/\",\"url\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/07\\\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\\\/\",\"name\":\"AI image generators' quality suffers when trained on AI-generated data | DailyAI\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/07\\\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/07\\\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2023\\\/07\\\/shutterstock_2252933361.jpg\",\"datePublished\":\"2023-07-18T20:48:46+00:00\",\"dateModified\":\"2023-07-18T21:23:28+00:00\",\"description\":\"A new study shows that training AI image generators with AI-generated images could lead to a self-consuming loop.\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/07\\\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\\\/#breadcrumb\"},\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/dailyai.com\\\/2023\\\/07\\\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/07\\\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\\\/#primaryimage\",\"url\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2023\\\/07\\\/shutterstock_2252933361.jpg\",\"contentUrl\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2023\\\/07\\\/shutterstock_2252933361.jpg\",\"width\":1000,\"height\":655,\"caption\":\"AI image generator\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/07\\\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/dailyai.com\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"AI image generators’ quality suffers when trained on AI-generated data\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/#website\",\"url\":\"https:\\\/\\\/dailyai.com\\\/\",\"name\":\"DailyAI\",\"description\":\"Your Daily Dose of AI News\",\"publisher\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/dailyai.com\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"fr-FR\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/#organization\",\"name\":\"DailyAI\",\"url\":\"https:\\\/\\\/dailyai.com\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2023\\\/06\\\/Daily-Ai_TL_colour.png\",\"contentUrl\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2023\\\/06\\\/Daily-Ai_TL_colour.png\",\"width\":4501,\"height\":934,\"caption\":\"DailyAI\"},\"image\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/x.com\\\/DailyAIOfficial\",\"https:\\\/\\\/www.linkedin.com\\\/company\\\/dailyaiofficial\\\/\",\"https:\\\/\\\/www.youtube.com\\\/@DailyAIOfficial\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/#\\\/schema\\\/person\\\/711e81f945549438e8bbc579efdeb3c9\",\"name\":\"Sam Jeans\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/a24a4a8f8e2a1a275b7491dc9c9f032c401eabf23c3206da4628dc84b6dac5c8?s=96&d=robohash&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/a24a4a8f8e2a1a275b7491dc9c9f032c401eabf23c3206da4628dc84b6dac5c8?s=96&d=robohash&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/a24a4a8f8e2a1a275b7491dc9c9f032c401eabf23c3206da4628dc84b6dac5c8?s=96&d=robohash&r=g\",\"caption\":\"Sam Jeans\"},\"description\":\"Sam is a science and technology writer who has worked in various AI startups. When he\u2019s not writing, he can be found reading medical journals or digging through boxes of vinyl records.\",\"sameAs\":[\"https:\\\/\\\/www.linkedin.com\\\/in\\\/sam-jeans-6746b9142\\\/\"],\"url\":\"https:\\\/\\\/dailyai.com\\\/fr\\\/author\\\/samjeans\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"La qualit\u00e9 des g\u00e9n\u00e9rateurs d'images d'IA diminue lorsqu'ils sont form\u00e9s sur des donn\u00e9es g\u00e9n\u00e9r\u00e9es par l'IA | DailyAI","description":"Une nouvelle \u00e9tude montre que l'entra\u00eenement des g\u00e9n\u00e9rateurs d'images de l'IA avec des images g\u00e9n\u00e9r\u00e9es par l'IA pourrait conduire \u00e0 une boucle d'autoconsommation.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/dailyai.com\/fr\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/","og_locale":"fr_FR","og_type":"article","og_title":"AI image generators' quality suffers when trained on AI-generated data | DailyAI","og_description":"A new study shows that training AI image generators with AI-generated images could lead to a self-consuming loop.","og_url":"https:\/\/dailyai.com\/fr\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/","og_site_name":"DailyAI","article_published_time":"2023-07-18T20:48:46+00:00","article_modified_time":"2023-07-18T21:23:28+00:00","og_image":[{"width":1000,"height":655,"url":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/07\/shutterstock_2252933361.jpg","type":"image\/jpeg"}],"author":"Sam Jeans","twitter_card":"summary_large_image","twitter_creator":"@DailyAIOfficial","twitter_site":"@DailyAIOfficial","twitter_misc":{"\u00c9crit par":"Sam Jeans","Dur\u00e9e de lecture estim\u00e9e":"3 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"NewsArticle","@id":"https:\/\/dailyai.com\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/#article","isPartOf":{"@id":"https:\/\/dailyai.com\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/"},"author":{"name":"Sam Jeans","@id":"https:\/\/dailyai.com\/#\/schema\/person\/711e81f945549438e8bbc579efdeb3c9"},"headline":"AI image generators’ quality suffers when trained on AI-generated data","datePublished":"2023-07-18T20:48:46+00:00","dateModified":"2023-07-18T21:23:28+00:00","mainEntityOfPage":{"@id":"https:\/\/dailyai.com\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/"},"wordCount":498,"publisher":{"@id":"https:\/\/dailyai.com\/#organization"},"image":{"@id":"https:\/\/dailyai.com\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/#primaryimage"},"thumbnailUrl":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/07\/shutterstock_2252933361.jpg","keywords":["Image generator","MidJourney","Stable Diffusion"],"articleSection":["Industry"],"inLanguage":"fr-FR"},{"@type":"WebPage","@id":"https:\/\/dailyai.com\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/","url":"https:\/\/dailyai.com\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/","name":"La qualit\u00e9 des g\u00e9n\u00e9rateurs d'images d'IA diminue lorsqu'ils sont form\u00e9s sur des donn\u00e9es g\u00e9n\u00e9r\u00e9es par l'IA | DailyAI","isPartOf":{"@id":"https:\/\/dailyai.com\/#website"},"primaryImageOfPage":{"@id":"https:\/\/dailyai.com\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/#primaryimage"},"image":{"@id":"https:\/\/dailyai.com\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/#primaryimage"},"thumbnailUrl":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/07\/shutterstock_2252933361.jpg","datePublished":"2023-07-18T20:48:46+00:00","dateModified":"2023-07-18T21:23:28+00:00","description":"Une nouvelle \u00e9tude montre que l'entra\u00eenement des g\u00e9n\u00e9rateurs d'images de l'IA avec des images g\u00e9n\u00e9r\u00e9es par l'IA pourrait conduire \u00e0 une boucle d'autoconsommation.","breadcrumb":{"@id":"https:\/\/dailyai.com\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/#breadcrumb"},"inLanguage":"fr-FR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/dailyai.com\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/"]}]},{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/dailyai.com\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/#primaryimage","url":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/07\/shutterstock_2252933361.jpg","contentUrl":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/07\/shutterstock_2252933361.jpg","width":1000,"height":655,"caption":"AI image generator"},{"@type":"BreadcrumbList","@id":"https:\/\/dailyai.com\/2023\/07\/ai-image-generators-quality-suffers-when-trained-on-ai-generated-data\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/dailyai.com\/"},{"@type":"ListItem","position":2,"name":"AI image generators’ quality suffers when trained on AI-generated data"}]},{"@type":"WebSite","@id":"https:\/\/dailyai.com\/#website","url":"https:\/\/dailyai.com\/","name":"DailyAI","description":"Votre dose quotidienne de nouvelles sur l'IA","publisher":{"@id":"https:\/\/dailyai.com\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/dailyai.com\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"fr-FR"},{"@type":"Organization","@id":"https:\/\/dailyai.com\/#organization","name":"DailyAI","url":"https:\/\/dailyai.com\/","logo":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/dailyai.com\/#\/schema\/logo\/image\/","url":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/06\/Daily-Ai_TL_colour.png","contentUrl":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/06\/Daily-Ai_TL_colour.png","width":4501,"height":934,"caption":"DailyAI"},"image":{"@id":"https:\/\/dailyai.com\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/x.com\/DailyAIOfficial","https:\/\/www.linkedin.com\/company\/dailyaiofficial\/","https:\/\/www.youtube.com\/@DailyAIOfficial"]},{"@type":"Person","@id":"https:\/\/dailyai.com\/#\/schema\/person\/711e81f945549438e8bbc579efdeb3c9","name":"Sam Jeans","image":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/secure.gravatar.com\/avatar\/a24a4a8f8e2a1a275b7491dc9c9f032c401eabf23c3206da4628dc84b6dac5c8?s=96&d=robohash&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/a24a4a8f8e2a1a275b7491dc9c9f032c401eabf23c3206da4628dc84b6dac5c8?s=96&d=robohash&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/a24a4a8f8e2a1a275b7491dc9c9f032c401eabf23c3206da4628dc84b6dac5c8?s=96&d=robohash&r=g","caption":"Sam Jeans"},"description":"Sam est un r\u00e9dacteur scientifique et technologique qui a travaill\u00e9 dans diverses start-ups sp\u00e9cialis\u00e9es dans l'IA. Lorsqu'il n'\u00e9crit pas, on peut le trouver en train de lire des revues m\u00e9dicales ou de fouiller dans des bo\u00eetes de disques vinyles.","sameAs":["https:\/\/www.linkedin.com\/in\/sam-jeans-6746b9142\/"],"url":"https:\/\/dailyai.com\/fr\/author\/samjeans\/"}]}},"_links":{"self":[{"href":"https:\/\/dailyai.com\/fr\/wp-json\/wp\/v2\/posts\/2680","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/dailyai.com\/fr\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/dailyai.com\/fr\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/dailyai.com\/fr\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/dailyai.com\/fr\/wp-json\/wp\/v2\/comments?post=2680"}],"version-history":[{"count":9,"href":"https:\/\/dailyai.com\/fr\/wp-json\/wp\/v2\/posts\/2680\/revisions"}],"predecessor-version":[{"id":2696,"href":"https:\/\/dailyai.com\/fr\/wp-json\/wp\/v2\/posts\/2680\/revisions\/2696"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/dailyai.com\/fr\/wp-json\/wp\/v2\/media\/2682"}],"wp:attachment":[{"href":"https:\/\/dailyai.com\/fr\/wp-json\/wp\/v2\/media?parent=2680"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/dailyai.com\/fr\/wp-json\/wp\/v2\/categories?post=2680"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/dailyai.com\/fr\/wp-json\/wp\/v2\/tags?post=2680"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}