{"id":9072,"date":"2020-06-15T10:29:00","date_gmt":"2020-06-15T09:29:00","guid":{"rendered":"https:\/\/datascientest.com\/?p=9072"},"modified":"2023-11-09T14:14:06","modified_gmt":"2023-11-09T14:14:06","slug":"fonctionnement-des-reseaux-neurones","status":"publish","type":"post","link":"https:\/\/datascientest.com\/fonctionnement-des-reseaux-neurones","title":{"rendered":"R\u00e9seau de neurones : d\u00e9finition et fonctionnement"},"content":{"rendered":"\t\t<div data-elementor-type=\"wp-post\" data-elementor-id=\"9072\" class=\"elementor elementor-9072\" data-elementor-post-type=\"post\">\n\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-c6f16cb e-flex e-con-boxed e-con\" data-id=\"c6f16cb\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-0a8adae elementor-widget elementor-widget-heading\" data-id=\"0a8adae\" 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\">Quatri\u00e8me \u00e9pisode de notre s\u00e9rie consacr\u00e9e au\u00a0Deep Learning, n\u2019h\u00e9sitez pas \u00e0\u00a0lire ou relire les\u00a0premiers pour vous rafra\u00eechir la m\u00e9moire sur les\u00a0d\u00e9finitions de <a href=\"https:\/\/datascientest.com\/definition-deep-learning\/#introduction\">Deep Learning<\/a>, <a href=\"https:\/\/datascientest.com\/machine-learning-tout-savoir\">Machine Learning<\/a> et <a href=\"https:\/\/datascientest.com\/intelligence-artificielle-definition\">Intelligence Artificielle<\/a>\u00a0ainsi que sur les grandes lignes du\u00a0<a href=\"https:\/\/datascientest.com\/deep-learning-reseau-de-neurones-biologiques-ou-artificiels\">fonctionnement\u00a0et des\u00a0applications des r\u00e9seaux de neurones<\/a>. N'h\u00e9sitez pas \u00e9galement \u00e0 relire la partie d\u00e9di\u00e9e au CNN : Convolutional Neural Network . D\u00e9marrons sans plus tarder cette nouvelle partie !\u00a0<\/h2>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-408f7da elementor-widget elementor-widget-text-editor\" data-id=\"408f7da\" 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>Table des mati\u00e8res:<\/p><ol><li><a href=\"https:\/\/datascientest.com\/introduction-deep-learning\">Introduction au Deep Learning<\/a><\/li><li><a href=\"https:\/\/datascientest.com\/deep-learning-reseau-de-neurones-biologiques-ou-artificiels\">R\u00e9seaux de neurones : Biologique VS Artificiels<\/a><\/li><li><a href=\"https:\/\/datascientest.com\/convolutional-neural-network\">Convolutional Neural Network<\/a><\/li><li>Fonctionnement des r\u00e9seaux de neurones<ol><li><a href=\"#rappelreseauxneurones\">Rappel sur les r\u00e9seaux de neurones<\/a><\/li><li><a href=\"#descentedegradient\">Descente de gradient<\/a><\/li><\/ol><\/li><li><a href=\"https:\/\/datascientest.com\/transfer-learning\">Transfer Learning<\/a>\u00a0<\/li><\/ol>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-cc788cc elementor-widget elementor-widget-text-editor\" data-id=\"cc788cc\" 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;\">Vous allez maintenant d\u00e9couvrir en d\u00e9tails\u00a0comment les r\u00e9seaux de<\/span><b> neurones artificiels<\/b><span style=\"font-weight: 400;\"> apprennent et leur mani\u00e8re de <\/span><b>s\u2019articuler dans un r\u00e9seau<\/b><span style=\"font-weight: 400;\">.<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-0ab34e8 e-flex e-con-boxed e-con\" data-id=\"0ab34e8\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-db803d9 elementor-widget elementor-widget-menu-anchor\" data-id=\"db803d9\" data-element_type=\"widget\" data-widget_type=\"menu-anchor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.15.0 - 02-08-2023 *\/\nbody.elementor-page .elementor-widget-menu-anchor{margin-bottom:0}<\/style>\t\t<div id=\"rappelreseauxneurones\" class=\"elementor-menu-anchor\"><\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-2278fb1 elementor-widget elementor-widget-heading\" data-id=\"2278fb1\" 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\">Petit rappel sur les r\u00e9seaux de neurones<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-857feb1 elementor-widget elementor-widget-text-editor\" data-id=\"857feb1\" 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>Pour rappel, les <strong>multi-layers perceptron<\/strong> sont un empilement de perceptron qui sont mis sous forme de couche. Prenons l\u2019exemple suivant pour illustrer le concept :<\/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-a26f154 elementor-widget elementor-widget-image\" data-id=\"a26f154\" 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\t\t\t<img decoding=\"async\" width=\"512\" height=\"234\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/06\/DP_2.png\" class=\"attachment-full size-full wp-image-4624\" alt=\"Deep learning neuronnes\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/06\/DP_2.png 512w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/06\/DP_2-300x137.png 300w\" sizes=\"(max-width: 512px) 100vw, 512px\" \/>\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-898b193 elementor-widget elementor-widget-text-editor\" data-id=\"898b193\" 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;\">Le mod\u00e8le consiste \u00e0 classifier (en 10 classes) des images de chiffres manuscrits. Les carr\u00e9s verts sont la couche d\u2019entr\u00e9e de notre mod\u00e8le, c\u2019est-\u00e0-dire la valeur des pixels. La couche interm\u00e9diaire et la couche sortie sont des perceptrons repr\u00e9sent\u00e9s par des ronds gris. La derni\u00e8re couche poss\u00e8de 10 neurones puisqu\u2019il y a 10 classes et retourne la probabilit\u00e9 correspondante \u00e0 chaque classe.<\/span><\/p><p><span style=\"font-weight: 400;\">Nous pouvons tr\u00e8s facilement impl\u00e9menter le mod\u00e8le sur <a href=\"https:\/\/datascientest.com\/tensorflow\">TensorFlow<\/a> :<\/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-f92c2c9 elementor-widget elementor-widget-image\" data-id=\"f92c2c9\" 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=\"512\" height=\"230\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/06\/DP.png\" class=\"attachment-full size-full wp-image-4623\" alt=\"Tensorflow\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/06\/DP.png 512w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/06\/DP-300x135.png 300w\" sizes=\"(max-width: 512px) 100vw, 512px\" \/>\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\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-1e6553a e-flex e-con-boxed e-con\" data-id=\"1e6553a\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-ae513b7 elementor-widget elementor-widget-heading\" data-id=\"ae513b7\" 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\">Comment entra\u00eener le mod\u00e8le ?<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-21fe25c elementor-widget elementor-widget-text-editor\" data-id=\"21fe25c\" 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;\">Pour entra\u00eener notre mod\u00e8le, c\u2019est-\u00e0-dire trouver les poids adapt\u00e9s, il est n\u00e9cessaire de <\/span><b>d\u00e9finir une fonction qui va quantifier l\u2019erreur de notre mod\u00e8le <\/b><span style=\"font-weight: 400;\">. En g\u00e9n\u00e9ral, pour les probl\u00e8mes de classification, nous utilisons la fonction \u00ab\u00a0cross-entropy\u00a0\u00bb pour mesurer l\u2019erreur.\u00a0<\/span><\/p><p><span style=\"font-weight: 400;\">Notre probl\u00e8me d\u2019optimisation consiste \u00e0 trouver les poids de notre mod\u00e8le qui minimise notre fonction d\u2019erreur :\u00a0<\/span><\/p><pre style=\"text-align: center;\"><span style=\"font-weight: 400;\">J(w)<\/span><\/pre><p><span style=\"font-weight: 400;\">En Deep Learning, nous <\/span><b>sommes confront\u00e9s \u00e0 des probl\u00e8mes de minimisation complexe<\/b><span style=\"font-weight: 400;\"> et <\/span><b>sans solution explicite<\/b><span style=\"font-weight: 400;\">. C\u2019est pourquoi, des m\u00e9thodes d\u2019optimisation comme la <\/span><b>descente de gradient<\/b><span style=\"font-weight: 400;\"> sont utilis\u00e9e pour les <\/span><b>r\u00e9soudre.<\/b><\/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-6f01a33 elementor-align-center button-principale elementor-widget elementor-widget-button\" data-id=\"6f01a33\" 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\/formations-data\">\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\">Apprendre \u00e0 construire des r\u00e9seaux de neurones<\/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\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-2bf01b4 e-flex e-con-boxed e-con\" data-id=\"2bf01b4\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-64aba02 elementor-widget elementor-widget-menu-anchor\" data-id=\"64aba02\" data-element_type=\"widget\" data-widget_type=\"menu-anchor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t<div id=\"descentedegradient\" class=\"elementor-menu-anchor\"><\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-20b6ef1 elementor-widget elementor-widget-heading\" data-id=\"20b6ef1\" 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\">Descente de gradient : Pour descendre, il faut aller vers le bas<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-4096940 elementor-widget elementor-widget-text-editor\" data-id=\"4096940\" 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>Pour d\u00e9couvrir d&rsquo;une mani\u00e8re intuitive le concept de descente du gradient, regardez notre vid\u00e9o sur le sujet :\u00a0<\/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-5c28762 elementor-widget__width-initial elementor-widget elementor-widget-video\" data-id=\"5c28762\" data-element_type=\"widget\" data-settings=\"{&quot;youtube_url&quot;:&quot;https:\\\/\\\/youtu.be\\\/Ka4AnSXxF2g&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<div class=\"elementor-element elementor-element-0533438 elementor-widget elementor-widget-text-editor\" data-id=\"0533438\" 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>Dans notre cas, la <b>fonction de perte<\/b> (d\u2019erreur) peut \u00eatre vue comme une montagne o\u00f9 l\u2019on <b>recherche les coordonn\u00e9es <\/b>(poids de mod\u00e8le) du point d\u2019altitude minimale. Et, la valeur du gradient de la fonction de perte correspond \u00e0 l\u2019orientation et \u00e0 l\u2019inclinaison de la pente en un point donn\u00e9\u00a0:<\/p><ul><li><span style=\"font-weight: 400;\">Si la <\/span><b>norme du gradient <\/b><span style=\"font-weight: 400;\">est \u00e9lev\u00e9e, c\u2019est que la pente est tr\u00e8s raide.<\/span><\/li><li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">Si la norme du gradient est faible, c\u2019est que la pente est peu inclin\u00e9e.<\/span><\/li><li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">Si le gradient est nul, c\u2019est qu\u2019on se trouve sur un plat (minimum).<\/span><\/li><\/ul><p><span style=\"font-weight: 400;\">Par cons\u00e9quent, la descendre de gradient consiste initialement \u00e0 choisir les poids al\u00e9atoires, et de suivre l\u2019inverse du gradient \u00e0 chaque it\u00e9ration\u00a0:\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-c3ada68 elementor-widget elementor-widget-image\" data-id=\"c3ada68\" 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\t<a href=\"https:\/\/medium.com\/@acb28\/implement-linear-regression-using-gradient-descent-9209e45273ce\" target=\"_blank\" rel=\"nofollow\">\n\t\t\t\t\t\t\t<img decoding=\"async\" width=\"379\" height=\"284\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/06\/descente-de-gradient.gif\" class=\"attachment-full size-full wp-image-4685\" alt=\"descente de gradient\" \/>\t\t\t\t\t\t\t\t<\/a>\n\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-9e8d84a elementor-widget elementor-widget-image\" data-id=\"9e8d84a\" 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=\"193\" height=\"52\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/06\/DP_fonctionnement.png\" class=\"attachment-full size-full wp-image-4622\" alt=\"Deep learning\" \/>\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-06344fe elementor-widget elementor-widget-text-editor\" data-id=\"06344fe\" 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;\">Avec :<\/span><\/p><ul><li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">W<\/span><span style=\"font-weight: 400;\">t<\/span><span style=\"font-weight: 400;\">\u00a0<\/span><span style=\"font-weight: 400;\">: les param\u00e8tres de notre mod\u00e8le<\/span><\/li><li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">\u03b7\u00a0: pas du gradient, autrement dit, la vitesse d&rsquo;apprentissage<\/span><\/li><li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">J\u00a0<\/span><span style=\"font-weight: 400;\">: fonction de perte<\/span><\/li><\/ul><p><span style=\"font-weight: 400;\">Un <\/span><b>pas de gradient trop grand ou trop petit ne permettra pas \u00e0 l&rsquo;algorithme MLP de converger vers une solution satisfaisante<\/b><span style=\"font-weight: 400;\">. Trouver le bon pas de gradient est tout le challenge du Deep Learning.<\/span><\/p><p><span style=\"font-weight: 400;\">Pour r\u00e9sumer, le neurone artificiel autrement appel\u00e9 le perceptron se comporte comme un mod\u00e8le de classification lin\u00e9aire. Le perceptron multicouche quant \u00e0 lui lie et empile sous forme de couche les perceptrons pour apporter une plus grande\u00a0complexit\u00e9\u00a0\u00e0 l\u2019\u00e9tablissement des r\u00e8gles.<\/span><\/p><p><span style=\"font-weight: 400;\">Apr\u00e8s avoir d\u00e9fini les param\u00e8tres du mod\u00e8le, nous avons vu qu\u2019il \u00e9tait n\u00e9cessaire de quantifier l\u2019erreur de celui-ci. Pour ce faire, le Deep Learning utilise principalement la descente de gradient pour ajuster ses poids de fa\u00e7on \u00e0 minimiser la fonction de perte.<\/span><\/p><p>Pour explorer \u00a0les math\u00e9matiques qui se cachent derri\u00e8re cet algorithme de descente de gradient et l&rsquo;importance du taux d&rsquo;apprentissage d\u00e9couvrez <a href=\"https:\/\/datascientest.com\/descente-de-gradient\">cet<\/a><a href=\"https:\/\/datascientest.com\/descente-de-gradient\"> article<\/a>.<\/p><p>Dans la suite, nous allons vous pr\u00e9senter Le Transfer Learning, une cat\u00e9gorie d&rsquo;apprentissage \u00e0 succ\u00e8s utilis\u00e9e en Deep Learning et qui semble fortement s&rsquo;inspirer de l&rsquo;apprentissage humain !\u00a0<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-219013e e-flex e-con-boxed e-con\" data-id=\"219013e\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-358da7e elementor-widget elementor-widget-text-editor\" data-id=\"358da7e\" 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><strong>Article 1\u00a0:<\/strong> <a href=\"https:\/\/datascientest.com\/introduction-deep-learning\">Introduction au deep learning et aux r\u00e9seaux de neurones<\/a><\/p><p><strong>Article 2 :<\/strong>\u00a0<a href=\"https:\/\/datascientest.com\/convolutional-neural-network\">Convolutional Neural Network<\/a><\/p><p><strong>Article 3 :<\/strong> Fonctionnement des r\u00e9seaux de neurones<\/p><p><strong>Article 4 :<\/strong> <a href=\"https:\/\/datascientest.com\/transfer-learning\">Transfer Learning<\/a><\/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-d38bb97 elementor-align-center button-principale elementor-widget elementor-widget-button\" data-id=\"d38bb97\" 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\/formation-deep-learning\">\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\">D\u00e9couvrir notre formation en Deep Learning<\/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\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-60388ec e-flex e-con-boxed e-con\" data-id=\"60388ec\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-fb11db8 elementor-widget elementor-widget-html\" data-id=\"fb11db8\" data-element_type=\"widget\" data-widget_type=\"html.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<script type=\"application\/ld+json\">\r\n{\r\n  \"@context\": \"https:\/\/schema.org\",\r\n  \"@type\": \"FAQPage\",\r\n  \"mainEntity\": [{\r\n    \"@type\": \"Question\",\r\n    \"name\": \"Comment entra\u00eener le mod\u00e8le ?\",\r\n    \"acceptedAnswer\": {\r\n      \"@type\": \"Answer\",\r\n      \"text\": \"Pour entra\u00eener notre mod\u00e8le, c\u2019est-\u00e0-dire trouver les poids adapt\u00e9s, il est n\u00e9cessaire de d\u00e9finir une fonction qui va quantifier l\u2019erreur de notre mod\u00e8le . En g\u00e9n\u00e9ral, pour les probl\u00e8mes de classification, nous utilisons la fonction \u00ab cross-entropy \u00bb pour mesurer l\u2019erreur.\"\r\n    }\r\n  },{\r\n    \"@type\": \"Question\",\r\n    \"name\": \"Descente de gradient : Pour descendre, il faut aller vers le bas\",\r\n    \"acceptedAnswer\": {\r\n      \"@type\": \"Answer\",\r\n      \"text\": \"Dans notre cas, la fonction de perte (d\u2019erreur) peut \u00eatre vue comme une montagne o\u00f9 l\u2019on recherche les coordonn\u00e9es (poids de mod\u00e8le) du point d\u2019altitude minimale. Et, la valeur du gradient de la fonction de perte correspond \u00e0 l\u2019orientation et \u00e0 l\u2019inclinaison de la pente en un point donn\u00e9 : Si la norme du gradient est \u00e9lev\u00e9e, c\u2019est que la pente est tr\u00e8s raide.\"\r\n    }\r\n  }]\r\n}\r\n<\/script>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"<p>Quatri\u00e8me \u00e9pisode de notre s\u00e9rie consacr\u00e9e au&nbsp;Deep Learning, n\u2019h\u00e9sitez pas \u00e0&nbsp;lire ou relire les&nbsp;premiers pour vous rafra\u00eechir la m\u00e9moire sur les&nbsp;d\u00e9finitions de Deep Learning, Machine Learning et Intelligence Artificielle&nbsp;ainsi que sur les grandes lignes du&nbsp;fonctionnement&nbsp;et des&nbsp;applications des r\u00e9seaux de neurones. N&rsquo;h\u00e9sitez pas \u00e9galement \u00e0 relire la partie d\u00e9di\u00e9e au CNN : Convolutional Neural Network . [&hellip;]<\/p>\n","protected":false},"author":77,"featured_media":80135,"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>R\u00e9seau de neurones : d\u00e9finition et fonctionnement<\/title>\n<meta name=\"description\" content=\"D\u00e9couvrez comment les r\u00e9seaux de neurones fonctionnent et leurs importance pour la Data Science et le Deep Learning.\" \/>\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\/fonctionnement-des-reseaux-neurones\" \/>\n<meta property=\"og:locale\" content=\"fr_FR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"R\u00e9seau de neurones : d\u00e9finition et fonctionnement\" \/>\n<meta property=\"og:url\" content=\"https:\/\/datascientest.com\/fonctionnement-des-reseaux-neurones\" \/>\n<meta property=\"og:site_name\" content=\"Formation Data Science | DataScientest.com\" \/>\n<meta property=\"article:published_time\" content=\"2020-06-15T09:29:00+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2023-11-09T14:14:06+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/datascientest.com\/wp-content\/uploads\/2021\/06\/illu_reseaux_neurones-25.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\/fonctionnement-des-reseaux-neurones\",\"url\":\"https:\/\/datascientest.com\/fonctionnement-des-reseaux-neurones\",\"name\":\"R\u00e9seau de neurones : d\u00e9finition et fonctionnement\",\"isPartOf\":{\"@id\":\"https:\/\/datascientest.com\/#website\"},\"datePublished\":\"2020-06-15T09:29:00+00:00\",\"dateModified\":\"2023-11-09T14:14:06+00:00\",\"author\":{\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1\"},\"description\":\"D\u00e9couvrez comment les r\u00e9seaux de neurones fonctionnent et leurs importance pour la Data Science et le Deep Learning.\",\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/datascientest.com\/fonctionnement-des-reseaux-neurones\"]}]},{\"@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":"R\u00e9seau de neurones : d\u00e9finition et fonctionnement","description":"D\u00e9couvrez comment les r\u00e9seaux de neurones fonctionnent et leurs importance pour la Data Science et le Deep Learning.","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\/fonctionnement-des-reseaux-neurones","og_locale":"fr_FR","og_type":"article","og_title":"R\u00e9seau de neurones : d\u00e9finition et fonctionnement","og_url":"https:\/\/datascientest.com\/fonctionnement-des-reseaux-neurones","og_site_name":"Formation Data Science | DataScientest.com","article_published_time":"2020-06-15T09:29:00+00:00","article_modified_time":"2023-11-09T14:14:06+00:00","og_image":[{"width":1024,"height":562,"url":"https:\/\/datascientest.com\/wp-content\/uploads\/2021\/06\/illu_reseaux_neurones-25.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\/fonctionnement-des-reseaux-neurones","url":"https:\/\/datascientest.com\/fonctionnement-des-reseaux-neurones","name":"R\u00e9seau de neurones : d\u00e9finition et fonctionnement","isPartOf":{"@id":"https:\/\/datascientest.com\/#website"},"datePublished":"2020-06-15T09:29:00+00:00","dateModified":"2023-11-09T14:14:06+00:00","author":{"@id":"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1"},"description":"D\u00e9couvrez comment les r\u00e9seaux de neurones fonctionnent et leurs importance pour la Data Science et le Deep Learning.","inLanguage":"fr-FR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/datascientest.com\/fonctionnement-des-reseaux-neurones"]}]},{"@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\/9072"}],"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=9072"}],"version-history":[{"count":0,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts\/9072\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/media\/80135"}],"wp:attachment":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/media?parent=9072"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/categories?post=9072"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/tags?post=9072"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}