{"id":72014,"date":"2021-06-21T09:37:06","date_gmt":"2021-06-21T08:37:06","guid":{"rendered":"https:\/\/datascientest.com\/?p=72014"},"modified":"2023-11-09T14:07:01","modified_gmt":"2023-11-09T14:07:01","slug":"les-autoencoders-modeles-dapprentissage-non-supervise","status":"publish","type":"post","link":"https:\/\/datascientest.com\/les-autoencoders-modeles-dapprentissage-non-supervise","title":{"rendered":"Les Autoencoders &#8211; mod\u00e8les d\u2019apprentissage non supervis\u00e9"},"content":{"rendered":"\t\t<div data-elementor-type=\"wp-post\" data-elementor-id=\"72014\" class=\"elementor elementor-72014\" data-elementor-post-type=\"post\">\n\t\t\t\t\t\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-02c96ea elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"02c96ea\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-21d20a9\" data-id=\"21d20a9\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-5769d69 elementor-widget elementor-widget-heading\" data-id=\"5769d69\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.15.0 - 02-08-2023 *\/\n.elementor-heading-title{padding:0;margin:0;line-height:1}.elementor-widget-heading .elementor-heading-title[class*=elementor-size-]>a{color:inherit;font-size:inherit;line-height:inherit}.elementor-widget-heading .elementor-heading-title.elementor-size-small{font-size:15px}.elementor-widget-heading .elementor-heading-title.elementor-size-medium{font-size:19px}.elementor-widget-heading .elementor-heading-title.elementor-size-large{font-size:29px}.elementor-widget-heading .elementor-heading-title.elementor-size-xl{font-size:39px}.elementor-widget-heading .elementor-heading-title.elementor-size-xxl{font-size:59px}<\/style><h2 class=\"elementor-heading-title elementor-size-default\">Aujourd\u2019hui, nous nous int\u00e9ressons \u00e0 une famille de m\u00e9thodes d\u2019apprentissage non supervis\u00e9, les autoencoders.<\/h2>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-ef78d8b elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"ef78d8b\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-2cd378f\" data-id=\"2cd378f\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-2a26855 elementor-widget elementor-widget-video\" data-id=\"2a26855\" data-element_type=\"widget\" data-settings=\"{&quot;youtube_url&quot;:&quot;https:\\\/\\\/youtu.be\\\/XpiruMSG2FY&quot;,&quot;video_type&quot;:&quot;youtube&quot;,&quot;controls&quot;:&quot;yes&quot;}\" data-widget_type=\"video.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.15.0 - 02-08-2023 *\/\n.elementor-widget-video .elementor-widget-container{overflow:hidden;transform:translateZ(0)}.elementor-widget-video .elementor-wrapper{aspect-ratio:var(--video-aspect-ratio)}.elementor-widget-video .elementor-wrapper iframe,.elementor-widget-video .elementor-wrapper video{height:100%;width:100%;display:flex;border:none;background-color:#000}@supports not (aspect-ratio:1\/1){.elementor-widget-video .elementor-wrapper{position:relative;overflow:hidden;height:0;padding-bottom:calc(100% \/ var(--video-aspect-ratio))}.elementor-widget-video .elementor-wrapper iframe,.elementor-widget-video .elementor-wrapper video{position:absolute;top:0;right:0;bottom:0;left:0}}.elementor-widget-video .elementor-open-inline .elementor-custom-embed-image-overlay{position:absolute;top:0;right:0;bottom:0;left:0;background-size:cover;background-position:50%}.elementor-widget-video .elementor-custom-embed-image-overlay{cursor:pointer;text-align:center}.elementor-widget-video .elementor-custom-embed-image-overlay:hover .elementor-custom-embed-play i{opacity:1}.elementor-widget-video .elementor-custom-embed-image-overlay img{display:block;width:100%;aspect-ratio:var(--video-aspect-ratio);-o-object-fit:cover;object-fit:cover;-o-object-position:center center;object-position:center center}@supports not (aspect-ratio:1\/1){.elementor-widget-video .elementor-custom-embed-image-overlay{position:relative;overflow:hidden;height:0;padding-bottom:calc(100% \/ var(--video-aspect-ratio))}.elementor-widget-video .elementor-custom-embed-image-overlay img{position:absolute;top:0;right:0;bottom:0;left:0}}.elementor-widget-video .e-hosted-video .elementor-video{-o-object-fit:cover;object-fit:cover}.e-con-inner>.elementor-widget-video,.e-con>.elementor-widget-video{width:var(--container-widget-width);--flex-grow:var(--container-widget-flex-grow)}<\/style>\t\t<div class=\"elementor-wrapper elementor-open-inline\">\n\t\t\t<div class=\"elementor-video\"><\/div>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-927b91f elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"927b91f\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-f48660a\" data-id=\"f48660a\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-61e7c98 elementor-widget elementor-widget-heading\" data-id=\"61e7c98\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h3 class=\"elementor-heading-title elementor-size-default\">Autoencoder : D\u00e9finition<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-8cda63d elementor-widget elementor-widget-text-editor\" data-id=\"8cda63d\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.15.0 - 02-08-2023 *\/\n.elementor-widget-text-editor.elementor-drop-cap-view-stacked .elementor-drop-cap{background-color:#69727d;color:#fff}.elementor-widget-text-editor.elementor-drop-cap-view-framed .elementor-drop-cap{color:#69727d;border:3px solid;background-color:transparent}.elementor-widget-text-editor:not(.elementor-drop-cap-view-default) .elementor-drop-cap{margin-top:8px}.elementor-widget-text-editor:not(.elementor-drop-cap-view-default) .elementor-drop-cap-letter{width:1em;height:1em}.elementor-widget-text-editor .elementor-drop-cap{float:left;text-align:center;line-height:1;font-size:50px}.elementor-widget-text-editor .elementor-drop-cap-letter{display:inline-block}<\/style>\t\t\t\t<p><span style=\"font-weight: 400;\">Les auto encodeurs sont des r\u00e9seaux de neurones un peu particuliers qui poss\u00e8dent exactement le <\/span><b>m\u00eame nombre de neurones sur leur couche d\u2019entr\u00e9e et leur couche de sortie<\/b><span style=\"font-weight: 400;\">. Le but pour un auto encodeur est d\u2019avoir une sortie la plus proche de l\u2019entr\u00e9e !<\/span><\/p><p><span style=\"font-weight: 400;\">L\u2019apprentissage est donc \u00ab auto-supervis\u00e9 \u00bb car la loss \u00e0 minimiser est le co\u00fbt de reconstruction entre la sortie et l\u2019entr\u00e9e<\/span><span style=\"font-weight: 400;\">. Les donn\u00e9es n\u2019ont donc pas \u00e0 \u00eatre \u00e9tiquet\u00e9es, car elles sont leurs propres \u00e9tiquettes, ce qui fait donc de ce mod\u00e8le un mod\u00e8le non supervis\u00e9.<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-728db3b elementor-widget elementor-widget-heading\" data-id=\"728db3b\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h3 class=\"elementor-heading-title elementor-size-default\">Architecture d\u2019un autoencoder<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-023396b elementor-widget elementor-widget-text-editor\" data-id=\"023396b\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p><span style=\"font-weight: 400;\">Un autoencoder a une architecture tr\u00e8s sp\u00e9cifique, car les couches cach\u00e9es sont plus petites que les couches d\u2019entr\u00e9e. On appelle ce type d\u2019architecture une<\/span><b> architecture \u00ab bottleneck \u00bb<\/b><span style=\"font-weight: 400;\">. On peut d\u00e9composer un auto encodeur en deux parties \u00e0 gauche et \u00e0 droite de ce \u00ab bottleneck \u00bb.<\/span><\/p><p><span style=\"font-weight: 400;\">La partie gauche s\u2019appelle l\u2019<\/span><b>encodeur<\/b><span style=\"font-weight: 400;\">. <\/span><span style=\"font-weight: 400;\">L\u2019encodeur transforme l\u2019entr\u00e9e en une repr\u00e9sentation dans un espace de dimension plus faible appel\u00e9 espace latent.<\/span><span style=\"font-weight: 400;\"> L\u2019encodeur compresse donc l\u2019entr\u00e9e dans une repr\u00e9sentation moins co\u00fbteuse.<\/span><\/p><p><strong>La partie droite est appel\u00e9e d\u00e9codeur, car elle doit reconstruire \u00e0 l\u2019aide de la repr\u00e9sentation latente de l\u2019entr\u00e9e, une sortie la plus fid\u00e8le \u00e0 l\u2019entr\u00e9e.<\/strong><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-9f0472d elementor-widget elementor-widget-image\" data-id=\"9f0472d\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.15.0 - 02-08-2023 *\/\n.elementor-widget-image{text-align:center}.elementor-widget-image a{display:inline-block}.elementor-widget-image a img[src$=\".svg\"]{width:48px}.elementor-widget-image img{vertical-align:middle;display:inline-block}<\/style>\t\t\t\t\t\t\t\t\t<figure class=\"wp-caption\">\n\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"800\" height=\"439\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2021\/06\/illu_autoencoder_schema-19.png\" class=\"attachment-large size-large wp-image-77769\" alt=\"autoencoder\" loading=\"lazy\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2021\/06\/illu_autoencoder_schema-19.png 1024w, https:\/\/datascientest.com\/wp-content\/uploads\/2021\/06\/illu_autoencoder_schema-19-300x165.png 300w, https:\/\/datascientest.com\/wp-content\/uploads\/2021\/06\/illu_autoencoder_schema-19-768x422.png 768w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/>\t\t\t\t\t\t\t\t\t\t\t<figcaption class=\"widget-image-caption wp-caption-text\">Sch\u00e9ma inspir\u00e9 d\u2019un cours de Fabien Moutarde - Mines de Paris (2021)<\/figcaption>\n\t\t\t\t\t\t\t\t\t\t<\/figure>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-2dd3066 elementor-widget elementor-widget-text-editor\" data-id=\"2dd3066\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p><span style=\"font-weight: 400;\">Lors de l\u2019apprentissage, l\u2019auto encodeur va donc apprendre \u00e0 chercher \u00e0<\/span><b> garder le plus d\u2019information possible entre l\u2019entr\u00e9e et l\u2019espace latent,<\/b><span style=\"font-weight: 400;\"> afin que le d\u00e9codeur ait les informations les plus essentielles pour reconstruire l\u2019image. Ainsi, l\u2019encodeur apprend les composantes les plus importantes d\u2019une entr\u00e9e pour avoir la meilleure compression possible.<\/span><\/p><p><span style=\"font-weight: 400;\">\u00c0 premi\u00e8re vue, cela ressemble fortement \u00e0 <a href=\"https:\/\/datascientest.com\/acp\">l\u2019analyse en composantes principales<\/a> (ACP ou PCA en anglais)<\/span><span style=\"font-weight: 400;\">. <\/span><span style=\"font-weight: 400;\">Cependant, les auto encodeurs permettent d\u2019ajouter des non-lin\u00e9arit\u00e9s gr\u00e2ce aux fonctions d\u2019activation et \u00e0 leur structure de r\u00e9seau de neurones, ce qui distingue un auto encodeur simple d\u2019une ACP.<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-1985c11 elementor-widget elementor-widget-heading\" data-id=\"1985c11\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h3 class=\"elementor-heading-title elementor-size-default\"> G\u00e9n\u00e9ralisation : autoencodeurs d\u00e9bruiteurs<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-6e9cca9 elementor-widget elementor-widget-text-editor\" data-id=\"6e9cca9\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p><span style=\"font-weight: 400;\">Ce principe d\u2019auto encodeur a \u00e9t\u00e9 g\u00e9n\u00e9ralis\u00e9 \u00e0 des t\u00e2ches plus vari\u00e9es, en particulier dans les t\u00e2ches de d\u00e9bruitage (on parle de denoising AE en anglais). Apr\u00e8s avoir appris comment <\/span><b>repr\u00e9senter de mani\u00e8re optimale les entr\u00e9es<\/b><span style=\"font-weight: 400;\">, en fournissant \u00e0 l\u2019auto encodeur des entr\u00e9es bruit\u00e9es, il est capable de donner en sortie des donn\u00e9es d\u00e9bruit\u00e9es qui sont tr\u00e8s proches de l\u2019entr\u00e9e originale.<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-f37aacd elementor-widget elementor-widget-heading\" data-id=\"f37aacd\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h3 class=\"elementor-heading-title elementor-size-default\">Mod\u00e8les g\u00e9n\u00e9ratifs : l\u2019espace latent<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-c89fd01 elementor-widget elementor-widget-text-editor\" data-id=\"c89fd01\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p><span style=\"font-weight: 400;\">Les auto encodeurs sont aussi tr\u00e8s utilis\u00e9s pour la <\/span><b>cr\u00e9ation de nouvelles entr\u00e9es<\/b><span style=\"font-weight: 400;\">. <\/span><span style=\"font-weight: 400;\">Dans ce cas- l\u00e0 on utilise la partie d\u00e9codeur pour cr\u00e9er des sorties \u00e0 partir de vecteurs de l\u2019espace latent<\/span><span style=\"font-weight: 400;\">. En choisissant des vecteurs \u00ab proches \u00bb d\u2019entr\u00e9e, la sortie y ressemblera fortement. Par exemple, en moyennant des vecteurs correspondant \u00e0 des entr\u00e9es du m\u00eame type, la sortie sera de ce type-l\u00e0.\u00a0<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-fc86ecc elementor-widget elementor-widget-image\" data-id=\"fc86ecc\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t\t\t\t<figure class=\"wp-caption\">\n\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"800\" height=\"202\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2021\/05\/rqGLkXyxh2g0Jlq2W8OPsfdkeUX7bOfcPGEhVfVFavcQRnO2rHNG5dBtpnef-d1RZe79w2dIsQ1HfIyijE_G2chm2OKl4z9T8422I2d4gJA9-OeASaFwHd9H9FpgzRXjk-1024x259.jpg\" class=\"attachment-large size-large wp-image-72076\" alt=\"\" loading=\"lazy\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2021\/05\/rqGLkXyxh2g0Jlq2W8OPsfdkeUX7bOfcPGEhVfVFavcQRnO2rHNG5dBtpnef-d1RZe79w2dIsQ1HfIyijE_G2chm2OKl4z9T8422I2d4gJA9-OeASaFwHd9H9FpgzRXjk-1024x259.jpg 1024w, https:\/\/datascientest.com\/wp-content\/uploads\/2021\/05\/rqGLkXyxh2g0Jlq2W8OPsfdkeUX7bOfcPGEhVfVFavcQRnO2rHNG5dBtpnef-d1RZe79w2dIsQ1HfIyijE_G2chm2OKl4z9T8422I2d4gJA9-OeASaFwHd9H9FpgzRXjk-300x76.jpg 300w, https:\/\/datascientest.com\/wp-content\/uploads\/2021\/05\/rqGLkXyxh2g0Jlq2W8OPsfdkeUX7bOfcPGEhVfVFavcQRnO2rHNG5dBtpnef-d1RZe79w2dIsQ1HfIyijE_G2chm2OKl4z9T8422I2d4gJA9-OeASaFwHd9H9FpgzRXjk-768x194.jpg 768w, https:\/\/datascientest.com\/wp-content\/uploads\/2021\/05\/rqGLkXyxh2g0Jlq2W8OPsfdkeUX7bOfcPGEhVfVFavcQRnO2rHNG5dBtpnef-d1RZe79w2dIsQ1HfIyijE_G2chm2OKl4z9T8422I2d4gJA9-OeASaFwHd9H9FpgzRXjk-1536x388.jpg 1536w, https:\/\/datascientest.com\/wp-content\/uploads\/2021\/05\/rqGLkXyxh2g0Jlq2W8OPsfdkeUX7bOfcPGEhVfVFavcQRnO2rHNG5dBtpnef-d1RZe79w2dIsQ1HfIyijE_G2chm2OKl4z9T8422I2d4gJA9-OeASaFwHd9H9FpgzRXjk.jpg 1600w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/>\t\t\t\t\t\t\t\t\t\t\t<figcaption class=\"widget-image-caption wp-caption-text\">Distribution dans l\u2019espace latent<\/figcaption>\n\t\t\t\t\t\t\t\t\t\t<\/figure>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-19c5e5b elementor-widget elementor-widget-text-editor\" data-id=\"19c5e5b\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p><span style=\"font-weight: 400;\">Cette arithm\u00e9tique est une des caract\u00e9ristiques de l\u2019espace latent et l\u2019on a des r\u00e9sultats encore plus int\u00e9ressants. Comme on peut associer les vecteurs de l\u2019espace latent \u00e0 leur sortie du d\u00e9codeur, cette arithm\u00e9tique se transmet \u00e0 la sortie ! Par exemple, en additionnant un vecteur correspondant \u00e0 un homme avec lunettes dans l\u2019espace avec un correspondant \u00e0 une femme sans lunettes et en soustrayant le vecteur correspondant \u00e0 un homme sans lunettes, on obtient des images de femmes avec lunettes.\u00a0<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-9f6d8e9 elementor-widget elementor-widget-image\" data-id=\"9f6d8e9\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"605\" height=\"286\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2021\/05\/without-glasses.jpg\" class=\"attachment-large size-large wp-image-72084\" alt=\"\" loading=\"lazy\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2021\/05\/without-glasses.jpg 605w, https:\/\/datascientest.com\/wp-content\/uploads\/2021\/05\/without-glasses-300x142.jpg 300w\" sizes=\"(max-width: 605px) 100vw, 605px\" \/>\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-f25b6ce elementor-widget elementor-widget-text-editor\" data-id=\"f25b6ce\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p><span style=\"font-weight: 400;\">Il y a \u00e9galement une transformation continue entre l\u2019espace latent et la sortie. <\/span><span style=\"font-weight: 400;\">Si on fait varier contin\u00fbment un vecteur de l\u2019espace latent, on obtient une transformation continue dans l\u2019espace de sortie.<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-ba1d107 elementor-widget elementor-widget-image\" data-id=\"ba1d107\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"605\" height=\"269\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2021\/05\/Interpolating.jpg\" class=\"attachment-large size-large wp-image-72089\" alt=\"Interpolating\" loading=\"lazy\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2021\/05\/Interpolating.jpg 605w, https:\/\/datascientest.com\/wp-content\/uploads\/2021\/05\/Interpolating-300x133.jpg 300w\" sizes=\"(max-width: 605px) 100vw, 605px\" \/>\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-e4699a6 elementor-widget elementor-widget-text-editor\" data-id=\"e4699a6\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p><span style=\"font-weight: 400;\">L\u2019int\u00e9r\u00eat des autoencodeurs que nous avons d\u00e9couvert ici est que ces derniers <\/span><span style=\"font-weight: 400;\">ne n\u00e9cessitent pas de donn\u00e9es \u00e9tiquet\u00e9es<\/span><span style=\"font-weight: 400;\">. Ces derni\u00e8res sont tr\u00e8s souvent longues \u00e0 cr\u00e9er, car il faut qu\u2019une personne labellise des centaines voir des milliers de donn\u00e9es une \u00e0 une.<\/span><\/p><p><span style=\"font-weight: 400;\">De plus, comme on vient de le voir, la partie d\u00e9codeur peut servir de mod\u00e8le g\u00e9n\u00e9ratif, ce qui peut \u00eatre utile pour faire de la <\/span><b>data augmentation<\/b><span style=\"font-weight: 400;\"> pour entra\u00eener d\u2019autres mod\u00e8les.<\/span><\/p><p><span style=\"font-weight: 400;\">\u00a0<\/span><span style=\"font-weight: 400;\">En dehors de ces exemples d\u2019utilisation, le champ d\u2019applications des autoencoders est tr\u00e8s vaste et tr\u00e8s prometteur.<\/span><\/p><p><span style=\"font-weight: 400;\">Si cet article vous a int\u00e9ress\u00e9 et que vous souhaitez en d\u00e9couvrir plus sur les possibilit\u00e9s du Deep Learning, notre formation Data Scientist vous apportera tous les outils et connaissances n\u00e9cessaires pour vous former au m\u00e9tier de Data Scientist.<\/span><\/p><p><span style=\"font-weight: 400;\">Si vous vous interrogez sur nos formations, n&rsquo;h\u00e9sitez\u00a0 pas \u00e0 nous contacter et \u00e0 prendre rendez-vous avec notre \u00e9quipe.<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-69dd7c0 elementor-align-center button-principale elementor-widget elementor-widget-button\" data-id=\"69dd7c0\" data-element_type=\"widget\" data-widget_type=\"button.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t<div class=\"elementor-button-wrapper\">\n\t\t\t<a class=\"elementor-button elementor-button-link elementor-size-sm elementor-animation-grow\" href=\"https:\/\/datascientest.com\/rdv\">\n\t\t\t\t\t\t<span class=\"elementor-button-content-wrapper\">\n\t\t\t\t\t\t<span class=\"elementor-button-text\">Prendre RDV<\/span>\n\t\t<\/span>\n\t\t\t\t\t<\/a>\n\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"<p>Aujourd\u2019hui, nous nous int\u00e9ressons \u00e0 une famille de m\u00e9thodes d\u2019apprentissage non supervis\u00e9, les autoencoders. Autoencoder : D\u00e9finition Les auto encodeurs sont des r\u00e9seaux de neurones un peu particuliers qui poss\u00e8dent exactement le m\u00eame nombre de neurones sur leur couche d\u2019entr\u00e9e et leur couche de sortie. Le but pour un auto encodeur est d\u2019avoir une sortie [&hellip;]<\/p>\n","protected":false},"author":77,"featured_media":77781,"comment_status":"open","ping_status":"open","sticky":false,"template":"elementor_theme","format":"standard","meta":{"content-type":"","inline_featured_image":false},"categories":[2497],"tags":[],"yoast_head":"<!-- This site is optimized with the Yoast SEO Premium plugin v21.7 (Yoast SEO v21.7) - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>Les Autoencoders - mod\u00e8les d\u2019apprentissage non supervis\u00e9<\/title>\n<meta name=\"description\" content=\"D\u00e9couvrez les autoencoders, le r\u00e9seau de neurones g\u00e9n\u00e9ralement utilis\u00e9 pour l&#039;apprentissage non supervis\u00e9 et des ses mod\u00e8les\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/datascientest.com\/les-autoencoders-modeles-dapprentissage-non-supervise\" \/>\n<meta property=\"og:locale\" content=\"fr_FR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Les Autoencoders - mod\u00e8les d\u2019apprentissage non supervis\u00e9\" \/>\n<meta property=\"og:url\" content=\"https:\/\/datascientest.com\/les-autoencoders-modeles-dapprentissage-non-supervise\" \/>\n<meta property=\"og:site_name\" content=\"Formation Data Science | DataScientest.com\" \/>\n<meta property=\"article:published_time\" content=\"2021-06-21T08:37:06+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2023-11-09T14:07:01+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/datascientest.com\/wp-content\/uploads\/2021\/06\/illu_autoencoder-18.png\" \/>\n\t<meta property=\"og:image:width\" content=\"1024\" \/>\n\t<meta property=\"og:image:height\" content=\"562\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"Raphael Kassel\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u00c9crit par\" \/>\n\t<meta name=\"twitter:data1\" content=\"Raphael Kassel\" \/>\n\t<meta name=\"twitter:label2\" content=\"Dur\u00e9e de lecture estim\u00e9e\" \/>\n\t<meta name=\"twitter:data2\" content=\"5 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/datascientest.com\/les-autoencoders-modeles-dapprentissage-non-supervise\",\"url\":\"https:\/\/datascientest.com\/les-autoencoders-modeles-dapprentissage-non-supervise\",\"name\":\"Les Autoencoders - mod\u00e8les d\u2019apprentissage non supervis\u00e9\",\"isPartOf\":{\"@id\":\"https:\/\/datascientest.com\/#website\"},\"datePublished\":\"2021-06-21T08:37:06+00:00\",\"dateModified\":\"2023-11-09T14:07:01+00:00\",\"author\":{\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1\"},\"description\":\"D\u00e9couvrez les autoencoders, le r\u00e9seau de neurones g\u00e9n\u00e9ralement utilis\u00e9 pour l'apprentissage non supervis\u00e9 et des ses mod\u00e8les\",\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/datascientest.com\/les-autoencoders-modeles-dapprentissage-non-supervise\"]}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/datascientest.com\/#website\",\"url\":\"https:\/\/datascientest.com\/\",\"name\":\"Formation Data Science | DataScientest.com\",\"description\":\"Leader de la formation en Data Science\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/datascientest.com\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"fr-FR\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1\",\"name\":\"Raphael Kassel\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/11\/raph.jpeg\",\"contentUrl\":\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/11\/raph.jpeg\",\"caption\":\"Raphael Kassel\"},\"description\":\"Raphael Kassel est le Head of Data Programs chez DataScientest. Il supervise une \u00e9quipe Data, impl\u00e9mente de nouveaux parcours comme \\\"Chef de projet IA\\\" et \\\"Data Product Manager\\\", et contribue \u00e0 la modernisation des op\u00e9rations du groupe. En collaboration \u00e9troite avec l'\u00e9quipe de d\u00e9veloppement, il adapte l'accompagnement acad\u00e9mique aux \u00e9volutions de l'\u00e9quipe data et \u00e0 l'internationalisation.\",\"sameAs\":[\"https:\/\/www.linkedin.com\/in\/raphaelkassel\/\",\"Data Science\"],\"url\":\"https:\/\/datascientest.com\/author\/raphaelkassel\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"Les Autoencoders - mod\u00e8les d\u2019apprentissage non supervis\u00e9","description":"D\u00e9couvrez les autoencoders, le r\u00e9seau de neurones g\u00e9n\u00e9ralement utilis\u00e9 pour l'apprentissage non supervis\u00e9 et des ses mod\u00e8les","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:\/\/datascientest.com\/les-autoencoders-modeles-dapprentissage-non-supervise","og_locale":"fr_FR","og_type":"article","og_title":"Les Autoencoders - mod\u00e8les d\u2019apprentissage non supervis\u00e9","og_url":"https:\/\/datascientest.com\/les-autoencoders-modeles-dapprentissage-non-supervise","og_site_name":"Formation Data Science | DataScientest.com","article_published_time":"2021-06-21T08:37:06+00:00","article_modified_time":"2023-11-09T14:07:01+00:00","og_image":[{"width":1024,"height":562,"url":"https:\/\/datascientest.com\/wp-content\/uploads\/2021\/06\/illu_autoencoder-18.png","type":"image\/png"}],"author":"Raphael Kassel","twitter_card":"summary_large_image","twitter_misc":{"\u00c9crit par":"Raphael Kassel","Dur\u00e9e de lecture estim\u00e9e":"5 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/datascientest.com\/les-autoencoders-modeles-dapprentissage-non-supervise","url":"https:\/\/datascientest.com\/les-autoencoders-modeles-dapprentissage-non-supervise","name":"Les Autoencoders - mod\u00e8les d\u2019apprentissage non supervis\u00e9","isPartOf":{"@id":"https:\/\/datascientest.com\/#website"},"datePublished":"2021-06-21T08:37:06+00:00","dateModified":"2023-11-09T14:07:01+00:00","author":{"@id":"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1"},"description":"D\u00e9couvrez les autoencoders, le r\u00e9seau de neurones g\u00e9n\u00e9ralement utilis\u00e9 pour l'apprentissage non supervis\u00e9 et des ses mod\u00e8les","inLanguage":"fr-FR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/datascientest.com\/les-autoencoders-modeles-dapprentissage-non-supervise"]}]},{"@type":"WebSite","@id":"https:\/\/datascientest.com\/#website","url":"https:\/\/datascientest.com\/","name":"Formation Data Science | DataScientest.com","description":"Leader de la formation en Data Science","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/datascientest.com\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"fr-FR"},{"@type":"Person","@id":"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1","name":"Raphael Kassel","image":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/datascientest.com\/#\/schema\/person\/image\/","url":"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/11\/raph.jpeg","contentUrl":"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/11\/raph.jpeg","caption":"Raphael Kassel"},"description":"Raphael Kassel est le Head of Data Programs chez DataScientest. Il supervise une \u00e9quipe Data, impl\u00e9mente de nouveaux parcours comme \"Chef de projet IA\" et \"Data Product Manager\", et contribue \u00e0 la modernisation des op\u00e9rations du groupe. En collaboration \u00e9troite avec l'\u00e9quipe de d\u00e9veloppement, il adapte l'accompagnement acad\u00e9mique aux \u00e9volutions de l'\u00e9quipe data et \u00e0 l'internationalisation.","sameAs":["https:\/\/www.linkedin.com\/in\/raphaelkassel\/","Data Science"],"url":"https:\/\/datascientest.com\/author\/raphaelkassel"}]}},"_links":{"self":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts\/72014"}],"collection":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/users\/77"}],"replies":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/comments?post=72014"}],"version-history":[{"count":0,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts\/72014\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/media\/77781"}],"wp:attachment":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/media?parent=72014"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/categories?post=72014"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/tags?post=72014"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}