{"id":37425,"date":"2026-05-27T11:33:54","date_gmt":"2026-05-27T11:33:54","guid":{"rendered":"https:\/\/aisuperior.com\/?p=37425"},"modified":"2026-05-27T11:33:54","modified_gmt":"2026-05-27T11:33:54","slug":"machine-learning-in-social-cognition","status":"publish","type":"post","link":"https:\/\/aisuperior.com\/fr\/machine-learning-in-social-cognition\/","title":{"rendered":"Apprentissage automatique en cognition sociale : guide 2026"},"content":{"rendered":"<p><b>R\u00e9sum\u00e9 rapide\u00a0: <\/b><span style=\"font-weight: 400;\">L&#039;apprentissage automatique r\u00e9volutionne la recherche en cognition sociale en permettant l&#039;analyse de comportements interpersonnels complexes, la pr\u00e9diction des cons\u00e9quences sociales et la mise en \u00e9vidence de sch\u00e9mas d&#039;attribution des \u00e9tats mentaux humains. Les mod\u00e8les r\u00e9cents atteignent des scores AUC d&#039;environ 0,80 pour la pr\u00e9diction des comportements sociaux gr\u00e2ce \u00e0 l&#039;int\u00e9gration de la th\u00e9orie psychologique et d&#039;algorithmes avanc\u00e9s. Ces approches transforment la mani\u00e8re dont les scientifiques \u00e9tudient des ph\u00e9nom\u00e8nes aussi divers que l&#039;isolement social et le raisonnement de type th\u00e9orie de l&#039;esprit.<\/span><\/p>\n<p>&nbsp;<\/p>\n<p><span style=\"font-weight: 400;\">La cognition sociale \u2014 la fa\u00e7on dont les humains per\u00e7oivent, interpr\u00e8tent et r\u00e9agissent aux informations sociales \u2014 a traditionnellement \u00e9t\u00e9 \u00e9tudi\u00e9e au moyen d&#039;exp\u00e9riences contr\u00f4l\u00e9es et de mesures d&#039;auto-\u00e9valuation. Mais ces m\u00e9thodes ne permettent de saisir que des instantan\u00e9s du comportement.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">L&#039;apprentissage automatique change compl\u00e8tement la donne.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">En analysant simultan\u00e9ment des milliers de donn\u00e9es comportementales, les algorithmes peuvent d\u00e9tecter des sch\u00e9mas qui pourraient \u00e9chapper aux chercheurs humains. Leurs applications s&#039;\u00e9tendent de la psychologie clinique au d\u00e9veloppement de l&#039;intelligence artificielle.<\/span><\/p>\n<h2><span style=\"font-weight: 400;\">Pourquoi l&#039;apprentissage automatique est important pour la recherche en cognition sociale<\/span><\/h2>\n<p><span style=\"font-weight: 400;\">Les approches statistiques traditionnelles supposent des relations lin\u00e9aires et exigent que les chercheurs pr\u00e9cisent au pr\u00e9alable quelles variables sont pertinentes. La cognition sociale ne fonctionne pas ainsi.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Le comportement social humain r\u00e9sulte d&#039;interactions complexes entre les processus cognitifs, les \u00e9tats \u00e9motionnels, les contextes environnementaux et les histoires individuelles. L&#039;apprentissage automatique g\u00e8re naturellement cette complexit\u00e9.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">D&#039;apr\u00e8s une \u00e9tude publi\u00e9e dans Nature en ao\u00fbt 2025, l&#039;int\u00e9gration de la th\u00e9orie sociocognitive \u00e0 l&#039;apprentissage automatique a permis de cr\u00e9er des mod\u00e8les atteignant une aire sous la courbe (AUC) d&#039;environ 0,80 pour la pr\u00e9diction de comportements sociaux complexes. Le mod\u00e8le int\u00e9grait neuf pr\u00e9dicteurs, dont des mesures de d\u00e9tresse psychologique, l&#039;estime de soi, des facteurs d\u00e9mographiques et le contexte comportemental.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Voici ce qui rend ces approches si puissantes\u00a0: elles apprennent les sch\u00e9mas hi\u00e9rarchiques sans que les chercheurs aient \u00e0 sp\u00e9cifier manuellement chaque terme d\u2019interaction.<\/span><\/p>\n<p><img fetchpriority=\"high\" decoding=\"async\" class=\"alignnone size-full wp-image-35586\" src=\"https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/04\/Superior.webp\" alt=\"\" width=\"434\" height=\"116\" srcset=\"https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/04\/Superior.webp 434w, https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/04\/Superior-300x80.webp 300w, https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/04\/Superior-18x5.webp 18w\" sizes=\"(max-width: 434px) 100vw, 434px\" \/><\/p>\n<h2><span style=\"font-weight: 400;\">Explorez les donn\u00e9es de cognition sociale gr\u00e2ce \u00e0 l&#039;IA sup\u00e9rieure<\/span><\/h2>\n<p><span style=\"font-weight: 400;\">Les recherches en cognition sociale combinent souvent observations comportementales, analyses du langage, enregistrements exp\u00e9rimentaux et ensembles de donn\u00e9es statistiques. <\/span><a href=\"https:\/\/aisuperior.com\/fr\/\" target=\"_blank\" rel=\"noopener\"><span style=\"font-weight: 400;\">IA sup\u00e9rieure<\/span><\/a><span style=\"font-weight: 400;\"> peut aider les groupes de recherche et les organisations qui appliquent l&#039;apprentissage automatique pour mieux organiser, interpr\u00e9ter et analyser les informations cognitives et comportementales complexes.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">AI Superior peut vous aider avec\u00a0:<\/span><\/p>\n<ul>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><span style=\"font-weight: 400;\">Traitement des ensembles de donn\u00e9es comportementales structur\u00e9es et non structur\u00e9es<\/span><\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><span style=\"font-weight: 400;\">D\u00e9veloppement de mod\u00e8les de classification et de pr\u00e9diction<\/span><\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><span style=\"font-weight: 400;\">Application des m\u00e9thodes de TAL aux documents de recherche textuels<\/span><\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><span style=\"font-weight: 400;\">Construction de syst\u00e8mes de validation de concept analytiques<\/span><\/li>\n<li style=\"font-weight: 400;\" aria-level=\"1\"><span style=\"font-weight: 400;\">\u00c9valuation de la qualit\u00e9 du mod\u00e8le et des performances analytiques<\/span><\/li>\n<\/ul>\n<p><a href=\"https:\/\/aisuperior.com\/fr\/contact\/\" target=\"_blank\" rel=\"noopener\"><span style=\"font-weight: 400;\">Contactez AI Superior<\/span><\/a><span style=\"font-weight: 400;\"> discuter de la structure de la recherche et des ensembles de donn\u00e9es disponibles.<\/span><\/p>\n<p><img decoding=\"async\" class=\"alignnone wp-image-37427 size-full\" src=\"https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/05\/image1-7-14.avif\" alt=\"M\u00e9triques de performance d&#039;un mod\u00e8le d&#039;apprentissage automatique guid\u00e9 par la th\u00e9orie pr\u00e9disant les sch\u00e9mas de comportement social (Nature, 2025)\" width=\"1310\" height=\"704\" srcset=\"https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/05\/image1-7-14.avif 1310w, https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/05\/image1-7-14-300x161.avif 300w, https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/05\/image1-7-14-1024x550.avif 1024w, https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/05\/image1-7-14-768x413.avif 768w, https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/05\/image1-7-14-18x10.avif 18w\" sizes=\"(max-width: 1310px) 100vw, 1310px\" \/><\/p>\n<p>&nbsp;<\/p>\n<h2><span style=\"font-weight: 400;\">Pr\u00e9dire l&#039;isolement social et la solitude<\/span><\/h2>\n<p><span style=\"font-weight: 400;\">L&#039;isolement social a de graves cons\u00e9quences sur la sant\u00e9. Des recherches montrent qu&#039;il est li\u00e9 \u00e0 un d\u00e9r\u00e8glement immunitaire et \u00e0 un risque accru de mortalit\u00e9.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Mais comment pr\u00e9dire qui souffrira d&#039;isolement plut\u00f4t que de solitude\u00a0? Une \u00e9tude publi\u00e9e dans Nature en juillet 2024 a appliqu\u00e9 l&#039;apprentissage automatique \u00e0 cette question sur trois groupes\u00a0: des personnes atteintes de schizophr\u00e9nie, de trouble bipolaire et des \u00e9chantillons issus de la population g\u00e9n\u00e9rale.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Les r\u00e9sultats ont r\u00e9v\u00e9l\u00e9 quelque chose d&#039;inattendu.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">L&#039;anh\u00e9donie sociale \u2014 une diminution du plaisir tir\u00e9 des interactions sociales \u2014 pr\u00e9disait \u00e0 la fois l&#039;isolement et la solitude dans tous les groupes. Ce r\u00e9sultat est coh\u00e9rent. Cependant, la cognition non sociale n&#039;expliquait une variance unique de l&#039;isolement que chez les personnes atteintes de schizophr\u00e9nie.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Les mod\u00e8les d&#039;apprentissage automatique ont identifi\u00e9 l&#039;anh\u00e9donie sociale et la cognition non sociale comme des pr\u00e9dicteurs cl\u00e9s de l&#039;isolement dans les \u00e9chantillons de schizophr\u00e9nie, la solitude pr\u00e9sentant des sch\u00e9mas similaires dans tous les groupes.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Cela d\u00e9montre la capacit\u00e9 de l&#039;apprentissage automatique \u00e0 identifier des pr\u00e9dicteurs sp\u00e9cifiques \u00e0 une population par opposition aux pr\u00e9dicteurs universels \u2014 chose que les m\u00e9thodes traditionnelles peinent \u00e0 accomplir efficacement.<\/span><\/p>\n<h2><span style=\"font-weight: 400;\">Th\u00e9orie de l&#039;esprit et intelligence artificielle<\/span><\/h2>\n<p><span style=\"font-weight: 400;\">La th\u00e9orie de l&#039;esprit d\u00e9signe la compr\u00e9hension que les autres ont des \u00e9tats mentaux \u2014 croyances, d\u00e9sirs, intentions \u2014 diff\u00e9rents des n\u00f4tres. Elle est fondamentale pour l&#039;interaction sociale.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Les mod\u00e8les d&#039;apprentissage automatique peuvent-ils d\u00e9velopper des capacit\u00e9s de th\u00e9orie de l&#039;esprit\u00a0? Des travaux r\u00e9cents semblent le confirmer, avec toutefois des r\u00e9serves.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Les recherches sur les mod\u00e8les augment\u00e9s par la th\u00e9orie de l&#039;esprit montrent des am\u00e9liorations de performance par rapport aux mod\u00e8les de base, les am\u00e9liorations de score variant selon la taille du mod\u00e8le.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Le probl\u00e8me, c&#039;est que ces mod\u00e8les ne \u201c comprennent \u201d pas vraiment les \u00e9tats mentaux comme le font les humains. Ils effectuent une reconnaissance de formes \u00e0 une \u00e9chelle extraordinaire.<\/span><\/p>\n<h2><span style=\"font-weight: 400;\">Trajectoires cognitives apr\u00e8s un traumatisme cr\u00e2nien<\/span><\/h2>\n<p><span style=\"font-weight: 400;\">La pr\u00e9diction des sch\u00e9mas de r\u00e9cup\u00e9ration apr\u00e8s un traumatisme cr\u00e2nien demeure malheureusement impr\u00e9cise. Trop de variables interagissent de mani\u00e8re non lin\u00e9aire.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Une \u00e9tude publi\u00e9e dans Nature en janvier 2026 a analys\u00e9 les approches d&#039;apprentissage automatique dans 30 \u00e9tudes publi\u00e9es portant sur 2 364 participants, majoritairement des hommes, pr\u00e9sentant un m\u00e9lange de traumatismes cr\u00e2niens l\u00e9gers et mod\u00e9r\u00e9s \u00e0 graves.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Les chercheurs ont appliqu\u00e9 des mod\u00e8les de for\u00eat al\u00e9atoire, de gradient boosting et d&#039;extreme gradient boosting dans le cadre de PROGRESS-Plus pour l&#039;analyse des d\u00e9terminants sociaux. Ils ont pr\u00e9dit le taux d&#039;\u00e9volution cognitive, et non seulement l&#039;\u00e9tat initial.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Les principaux facteurs pr\u00e9dictifs identifi\u00e9s sont les intervalles de temps, les indicateurs structurels au niveau national, l&#039;\u00e2ge et la variabilit\u00e9 du niveau d&#039;\u00e9ducation. L&#039;analyse des explications additives de Shapley a permis de d\u00e9terminer les facteurs d\u00e9terminants des pr\u00e9dictions pour chaque cas individuel.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Cette approche comble une lacune importante. Les param\u00e8tres sociaux influen\u00e7ant l&#039;\u00e9volution des traumatismes cr\u00e2niens ont \u00e9t\u00e9 insuffisamment \u00e9tudi\u00e9s, ce qui engendre des lacunes dans la pratique clinique. L&#039;apprentissage automatique permet de quantifier ces influences auparavant difficiles \u00e0 cerner.<\/span><\/p>\n<h2><span style=\"font-weight: 400;\">Statut socio-\u00e9conomique et d\u00e9veloppement c\u00e9r\u00e9bral<\/span><\/h2>\n<p><span style=\"font-weight: 400;\">Le statut socio-\u00e9conomique laisse-t-il des traces neuronales\u00a0? Une \u00e9tude d\u2019octobre 2025 a appliqu\u00e9 des mod\u00e8les de r\u00e9seaux \u00e9lastiques \u00e0 des donn\u00e9es de neuroimagerie multimodales provenant d\u2019adolescents.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Les mod\u00e8les pr\u00e9disaient les revenus \u00e0 partir des seules donn\u00e9es d&#039;imagerie c\u00e9r\u00e9brale, sans aucune information d\u00e9mographique initiale. Les donn\u00e9es d&#039;imagerie par tenseur de diffusion, d&#039;IRM structurelle et de connectivit\u00e9 fonctionnelle au repos ont servi d&#039;entr\u00e9es.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Le mod\u00e8le multimodal le plus performant a atteint une AUC de 0,75 sur les donn\u00e9es de test sans informations d\u00e9mographiques et d&#039;environ 0,779 avec les donn\u00e9es d\u00e9mographiques.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Les mod\u00e8les distinguant les enfants des tranches de revenus extr\u00eames ont montr\u00e9 de solides performances, avec une AUC de 0,81 sans donn\u00e9es d\u00e9mographiques et de 0,863 avec donn\u00e9es d\u00e9mographiques.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">L&#039;imagerie par tenseur de diffusion s&#039;est av\u00e9r\u00e9e la plus discriminante, suivie de l&#039;IRM structurelle. Les caract\u00e9ristiques les plus pr\u00e9dictives \u00e9taient distribu\u00e9es globalement plut\u00f4t que localis\u00e9es, notamment dans les r\u00e9gions associ\u00e9es aux fonctions ex\u00e9cutives et au langage.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">L&#039;inclusion de donn\u00e9es d\u00e9mographiques a am\u00e9lior\u00e9 les performances du mod\u00e8le, avec des am\u00e9liorations plus importantes observ\u00e9es pour les donn\u00e9es de connectivit\u00e9 fonctionnelle au repos.<\/span><\/p>\n<table>\n<thead>\n<tr>\n<th><span style=\"font-weight: 400;\">Type de mod\u00e8le<\/span><\/th>\n<th><span style=\"font-weight: 400;\">AUC (Test, sans donn\u00e9es d\u00e9mographiques)<\/span><\/th>\n<th><span style=\"font-weight: 400;\">AUC (Test, avec donn\u00e9es d\u00e9mographiques)<\/span><\/th>\n<th><span style=\"font-weight: 400;\">Caract\u00e9ristiques les plus discriminantes<\/span><\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><span style=\"font-weight: 400;\">Revenu multimodal<\/span><\/td>\n<td><span style=\"font-weight: 400;\">0.75<\/span><\/td>\n<td><span style=\"font-weight: 400;\">0.779<\/span><\/td>\n<td><span style=\"font-weight: 400;\">Int\u00e9grit\u00e9 de la substance blanche, distribution globale<\/span><\/td>\n<\/tr>\n<tr>\n<td><span style=\"font-weight: 400;\">Extr\u00eames de revenus<\/span><\/td>\n<td><span style=\"font-weight: 400;\">0.81<\/span><\/td>\n<td><span style=\"font-weight: 400;\">0.863<\/span><\/td>\n<td><span style=\"font-weight: 400;\">Fonction ex\u00e9cutive, r\u00e9gions linguistiques<\/span><\/td>\n<\/tr>\n<tr>\n<td><span style=\"font-weight: 400;\">DTI uniquement<\/span><\/td>\n<td><span style=\"font-weight: 400;\">modalit\u00e9 unique la plus \u00e9lev\u00e9e<\/span><\/td>\n<td><span style=\"font-weight: 400;\">+2-4% avec donn\u00e9es d\u00e9mographiques<\/span><\/td>\n<td><span style=\"font-weight: 400;\">Organisation de la substance blanche<\/span><\/td>\n<\/tr>\n<tr>\n<td><span style=\"font-weight: 400;\">RSFC uniquement<\/span><\/td>\n<td><span style=\"font-weight: 400;\">modalit\u00e9 unique la plus basse<\/span><\/td>\n<td><span style=\"font-weight: 400;\">+10% avec donn\u00e9es d\u00e9mographiques<\/span><\/td>\n<td><span style=\"font-weight: 400;\">Mod\u00e8les de connectivit\u00e9 fonctionnelle<\/span><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h2><span style=\"font-weight: 400;\">\u00c9laboration de th\u00e9ories unifi\u00e9es de la cognition<\/span><\/h2>\n<p><span style=\"font-weight: 400;\">Les sciences cognitives sont confront\u00e9es \u00e0 un probl\u00e8me de fragmentation. Des th\u00e9ories existent pour des domaines sp\u00e9cifiques \u2014 le langage naturel \u00e0 des niveaux alg\u00e9briques, les algorithmes d&#039;apprentissage, les m\u00e9canismes de plasticit\u00e9 c\u00e9r\u00e9brale \u2014 mais les relier reste un d\u00e9fi.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">L\u2019apprentissage automatique pourrait-il servir de liant informatique\u00a0? Un article de Nature paru en mai 2026 examine cette possibilit\u00e9.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Deux approches informatiques se r\u00e9v\u00e8lent prometteuses\u00a0: les syst\u00e8mes symboliques qui manipulent des repr\u00e9sentations discr\u00e8tes et les r\u00e9seaux connexionnistes qui apprennent des mod\u00e8les distribu\u00e9s. Historiquement, ces deux camps ont rarement communiqu\u00e9.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">L&#039;apprentissage automatique, et plus particuli\u00e8rement l&#039;apprentissage profond, d\u00e9montre comment ces deux approches peuvent se compl\u00e9ter plut\u00f4t que s&#039;opposer. Les r\u00e9seaux de neurones apprennent des repr\u00e9sentations hi\u00e9rarchiques interpr\u00e9tables symboliquement. Les contraintes symboliques peuvent guider l&#039;architecture des r\u00e9seaux.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Cette synth\u00e8se pourrait permettre une int\u00e9gration \u00e0 diff\u00e9rents niveaux d&#039;analyse, depuis les th\u00e9ories computationnelles abstraites jusqu&#039;aux d\u00e9tails de l&#039;impl\u00e9mentation neuronale.<\/span><\/p>\n<h2><span style=\"font-weight: 400;\">Applications pratiques et orientations futures<\/span><\/h2>\n<p><span style=\"font-weight: 400;\">Quelles sont les applications concr\u00e8tes de cette recherche ?<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Les milieux cliniques en b\u00e9n\u00e9ficient imm\u00e9diatement. Les mod\u00e8les pr\u00e9dictifs de l&#039;isolement social permettent d&#039;identifier les personnes \u00e0 risque avant que la rupture ne s&#039;installe durablement. Les \u00e9valuations de la th\u00e9orie de l&#039;esprit pourraient \u00e9clairer les interventions aupr\u00e8s des personnes atteintes de troubles du spectre autistique.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Pour le d\u00e9veloppement de l&#039;IA, la recherche en cognition sociale fournit des mod\u00e8les. Si l&#039;objectif est de cr\u00e9er des machines qui collaborent naturellement avec les humains, il est essentiel de comprendre comment l&#039;intelligence biologique traite l&#039;information sociale.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Des chercheurs ont utilis\u00e9 l&#039;apprentissage automatique appliqu\u00e9 \u00e0 des donn\u00e9es EEG pour comprendre l&#039;attraction subjective, g\u00e9n\u00e9rant des portraits correspondant aux pr\u00e9f\u00e9rences individuelles avec une pr\u00e9cision sup\u00e9rieure \u00e0 80 % lors des tests. Ceci d\u00e9montre des applications au-del\u00e0 de la psychologie traditionnelle.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Des d\u00e9fis subsistent. Les mod\u00e8les d&#039;apprentissage automatique sont tr\u00e8s gourmands en donn\u00e9es. La cognition sociale implique des processus subtils et contextuels dont la mise \u00e0 l&#039;\u00e9chelle peut s&#039;av\u00e9rer complexe. Les consid\u00e9rations \u00e9thiques li\u00e9es \u00e0 la pr\u00e9diction des comportements sociaux exigent une attention particuli\u00e8re.<\/span><\/p>\n<h2><span style=\"font-weight: 400;\">FAQ<\/span><\/h2>\n<div class=\"schema-faq-code\">\n<div class=\"faq-question\">\n<h3 class=\"faq-q\">Qu\u2019est-ce que l\u2019apprentissage automatique en cognition sociale\u00a0?<\/h3>\n<div>\n<p class=\"faq-a\">L&#039;apprentissage automatique en cognition sociale utilise des algorithmes tels que les for\u00eats al\u00e9atoires, le gradient boosting et les r\u00e9seaux de neurones pour pr\u00e9dire et expliquer comment les individus per\u00e7oivent, interpr\u00e8tent et r\u00e9agissent aux informations sociales. Ces mod\u00e8les analysent les tendances dans les donn\u00e9es comportementales, de neuroimagerie et psychologiques afin de r\u00e9v\u00e9ler des relations que les statistiques traditionnelles pourraient ne pas mettre en \u00e9vidence.<\/p>\n<\/div>\n<\/div>\n<div class=\"faq-question\">\n<h3 class=\"faq-q\">Dans quelle mesure les mod\u00e8les d&#039;apprentissage automatique sont-ils pr\u00e9cis pour pr\u00e9dire le comportement social\u00a0?<\/h3>\n<div>\n<p class=\"faq-a\">Des \u00e9tudes r\u00e9centes d\u00e9montrent d&#039;excellentes performances. Les mod\u00e8les guid\u00e9s par la th\u00e9orie ont atteint une aire sous la courbe (AUC) de 0,80 pour la pr\u00e9diction des comportements sociaux, avec une sensibilit\u00e9 de 0,72 et une sp\u00e9cificit\u00e9 de 0,77 aux seuils optimaux. La pr\u00e9cision du mod\u00e8le d\u00e9pend fortement de la taille de l&#039;\u00e9chantillon, de la qualit\u00e9 des caract\u00e9ristiques et de la pertinence de la th\u00e9orie psychologique pour la s\u00e9lection des variables.<\/p>\n<\/div>\n<\/div>\n<div class=\"faq-question\">\n<h3 class=\"faq-q\">L&#039;IA peut-elle d\u00e9velopper une th\u00e9orie de l&#039;esprit ?<\/h3>\n<div>\n<p class=\"faq-a\">Les mod\u00e8les d&#039;IA peuvent apprendre \u00e0 simuler le raisonnement de la th\u00e9orie de l&#039;esprit. Les recherches montrent que les mod\u00e8les de langage enrichis par la th\u00e9orie de l&#039;esprit am\u00e9liorent les performances, avec des gains plus importants pour les petits mod\u00e8les et des am\u00e9liorations plus modestes pour les plus grands. Cependant, ces syst\u00e8mes effectuent une reconnaissance de formes plut\u00f4t que de comprendre v\u00e9ritablement les \u00e9tats mentaux comme le font les humains\u00a0; les m\u00e9canismes diff\u00e8rent fondamentalement.<\/p>\n<\/div>\n<\/div>\n<div class=\"faq-question\">\n<h3 class=\"faq-q\">Quels sont les facteurs pr\u00e9dictifs de l&#039;isolement social par rapport \u00e0 la solitude ?<\/h3>\n<div>\n<p class=\"faq-a\">Des \u00e9tudes d&#039;apprentissage automatique ont montr\u00e9 que l&#039;anh\u00e9donie sociale pr\u00e9dit l&#039;isolement et la solitude dans toutes les populations. Cependant, la cognition non sociale pr\u00e9dit sp\u00e9cifiquement l&#039;isolement chez les personnes atteintes de schizophr\u00e9nie. Cela sugg\u00e8re que des facteurs universels (diminution du plaisir social) et des m\u00e9canismes propres \u00e0 chaque population contribuent \u00e0 la d\u00e9connexion sociale.<\/p>\n<\/div>\n<\/div>\n<div class=\"faq-question\">\n<h3 class=\"faq-q\">Comment le statut socio-\u00e9conomique influence-t-il le d\u00e9veloppement c\u00e9r\u00e9bral\u00a0?<\/h3>\n<div>\n<p class=\"faq-a\">L&#039;imagerie c\u00e9r\u00e9brale multimodale combin\u00e9e \u00e0 l&#039;apprentissage automatique montre que le revenu pr\u00e9dit la structure et le fonctionnement du cerveau des adolescents avec une aire sous la courbe (AUC) comprise entre 0,75 et 0,81. L&#039;int\u00e9grit\u00e9 de la substance blanche et les caract\u00e9ristiques distribu\u00e9es globalement li\u00e9es aux fonctions ex\u00e9cutives et au langage sont les plus discriminantes. Les diff\u00e9rences sont les plus marqu\u00e9es entre les tranches de revenus extr\u00eames.<\/p>\n<\/div>\n<\/div>\n<div class=\"faq-question\">\n<h3 class=\"faq-q\">Quelles m\u00e9thodes d&#039;apprentissage automatique sont les plus performantes pour la cognition sociale\u00a0?<\/h3>\n<div>\n<p class=\"faq-a\">Les for\u00eats al\u00e9atoires, le gradient boosting et la r\u00e9gression Elastic Net sont fr\u00e9quemment utilis\u00e9s dans les \u00e9tudes concluantes. La m\u00e9thode optimale d\u00e9pend de la question pos\u00e9e\u00a0: les for\u00eats al\u00e9atoires g\u00e8rent efficacement les interactions non lin\u00e9aires, les Elastic Nets corrigent la multicolin\u00e9arit\u00e9 dans les donn\u00e9es c\u00e9r\u00e9brales et le gradient boosting atteint souvent d\u2019excellentes performances pr\u00e9dictives lorsqu\u2019il est correctement param\u00e9tr\u00e9.<\/p>\n<\/div>\n<\/div>\n<div class=\"faq-question\">\n<h3 class=\"faq-q\">Quels sont les enjeux \u00e9thiques ?<\/h3>\n<div>\n<p class=\"faq-a\">La pr\u00e9diction des comportements sociaux soul\u00e8ve des questions de respect de la vie priv\u00e9e, de risques de discrimination et de consentement. Les mod\u00e8les entra\u00een\u00e9s sur des donn\u00e9es biais\u00e9es peuvent perp\u00e9tuer les st\u00e9r\u00e9otypes. L&#039;utilisation de pr\u00e9dictions du statut socio-\u00e9conomique fond\u00e9es sur les neurosciences pourrait stigmatiser les groupes d\u00e9favoris\u00e9s. Les chercheurs doivent veiller \u00e0 ce que les mod\u00e8les am\u00e9liorent les conditions de vie sans permettre la surveillance ni renforcer les in\u00e9galit\u00e9s.<\/p>\n<h2><span style=\"font-weight: 400;\">Conclusion<\/span><\/h2>\n<p><span style=\"font-weight: 400;\">L\u2019apprentissage automatique transforme en profondeur la recherche en cognition sociale. Les mod\u00e8les pr\u00e9disent d\u00e9sormais des comportements sociaux complexes avec une pr\u00e9cision de discrimination de 80%, identifient les facteurs de risque d\u2019isolement propres \u00e0 certaines populations et r\u00e9v\u00e8lent les signatures neuronales du d\u00e9savantage social.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Ces avanc\u00e9es permettent non seulement de d\u00e9crire les ph\u00e9nom\u00e8nes, mais aussi de pr\u00e9dire les r\u00e9sultats et d&#039;expliquer les m\u00e9canismes. Les approches fond\u00e9es sur la th\u00e9orie, qui int\u00e8grent des cadres psychologiques \u00e0 la puissance des algorithmes, offrent de meilleures performances que chacune prise isol\u00e9ment.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">La convergence des sciences cognitives et de l&#039;apprentissage automatique ouvre la voie \u00e0 des th\u00e9ories unifi\u00e9es couvrant de multiples niveaux d&#039;analyse. \u00c0 mesure que la qualit\u00e9 des donn\u00e9es s&#039;am\u00e9liore et que les m\u00e9thodes progressent, il faut s&#039;attendre \u00e0 des avanc\u00e9es plus rapides dans la compr\u00e9hension des fondements computationnels de l&#039;intelligence sociale.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Pour les chercheurs, les cliniciens et les d\u00e9veloppeurs d&#039;IA, le message est clair\u00a0: l&#039;apprentissage automatique n&#039;est pas seulement un outil pour la recherche en cognition sociale, il devient une infrastructure essentielle pour la prochaine g\u00e9n\u00e9ration de d\u00e9couvertes.<\/span><\/p>\n<\/div>\n<\/div>\n<\/div>","protected":false},"excerpt":{"rendered":"<p>Quick Summary: Machine learning is revolutionizing social cognition research by enabling analysis of complex interpersonal behaviors, predicting social outcomes, and uncovering patterns in human mental state attribution. Recent models achieve AUC scores of approximately 0.80 in predicting social behaviors by integrating psychological theory with advanced algorithms. These approaches are transforming how scientists study everything from [&hellip;]<\/p>\n","protected":false},"author":7,"featured_media":37426,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"inline_featured_image":false,"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","ast-disable-related-posts":"","theme-transparent-header-meta":"default","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"set","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"categories":[1],"tags":[],"class_list":["post-37425","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-blog"],"acf":[],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v27.8 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>Machine Learning in Social Cognition: 2026 Guide<\/title>\n<meta name=\"description\" content=\"Discover how machine learning transforms social cognition research, from predicting behavior (AUC 0.80) to understanding Theory of Mind. Evidence-based insights.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/aisuperior.com\/fr\/machine-learning-in-social-cognition\/\" \/>\n<meta property=\"og:locale\" content=\"fr_FR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Machine Learning in Social Cognition: 2026 Guide\" \/>\n<meta property=\"og:description\" content=\"Discover how machine learning transforms social cognition research, from predicting behavior (AUC 0.80) to understanding Theory of Mind. Evidence-based insights.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/aisuperior.com\/fr\/machine-learning-in-social-cognition\/\" \/>\n<meta property=\"og:site_name\" content=\"aisuperior\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/aisuperior\" \/>\n<meta property=\"article:published_time\" content=\"2026-05-27T11:33:54+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/05\/unnamed-7-14.webp\" \/>\n\t<meta property=\"og:image:width\" content=\"1168\" \/>\n\t<meta property=\"og:image:height\" content=\"784\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/webp\" \/>\n<meta name=\"author\" content=\"kateryna\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@aisuperior\" \/>\n<meta name=\"twitter:site\" content=\"@aisuperior\" \/>\n<meta name=\"twitter:label1\" content=\"\u00c9crit par\" \/>\n\t<meta name=\"twitter:data1\" content=\"kateryna\" \/>\n\t<meta name=\"twitter:label2\" content=\"Dur\u00e9e de lecture estim\u00e9e\" \/>\n\t<meta name=\"twitter:data2\" content=\"8 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/aisuperior.com\\\/machine-learning-in-social-cognition\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/aisuperior.com\\\/machine-learning-in-social-cognition\\\/\"},\"author\":{\"name\":\"kateryna\",\"@id\":\"https:\\\/\\\/aisuperior.com\\\/#\\\/schema\\\/person\\\/14fcb7aaed4b2b617c4f75699394241c\"},\"headline\":\"Machine Learning in Social Cognition: 2026 Guide\",\"datePublished\":\"2026-05-27T11:33:54+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/aisuperior.com\\\/machine-learning-in-social-cognition\\\/\"},\"wordCount\":1574,\"publisher\":{\"@id\":\"https:\\\/\\\/aisuperior.com\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/aisuperior.com\\\/machine-learning-in-social-cognition\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/aisuperior.com\\\/wp-content\\\/uploads\\\/2026\\\/05\\\/unnamed-7-14.webp\",\"articleSection\":[\"Blog\"],\"inLanguage\":\"fr-FR\"},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/aisuperior.com\\\/machine-learning-in-social-cognition\\\/\",\"url\":\"https:\\\/\\\/aisuperior.com\\\/machine-learning-in-social-cognition\\\/\",\"name\":\"Machine Learning in Social Cognition: 2026 Guide\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/aisuperior.com\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/aisuperior.com\\\/machine-learning-in-social-cognition\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/aisuperior.com\\\/machine-learning-in-social-cognition\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/aisuperior.com\\\/wp-content\\\/uploads\\\/2026\\\/05\\\/unnamed-7-14.webp\",\"datePublished\":\"2026-05-27T11:33:54+00:00\",\"description\":\"Discover how machine learning transforms social cognition research, from predicting behavior (AUC 0.80) to understanding Theory of Mind. Evidence-based insights.\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/aisuperior.com\\\/machine-learning-in-social-cognition\\\/#breadcrumb\"},\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/aisuperior.com\\\/machine-learning-in-social-cognition\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\\\/\\\/aisuperior.com\\\/machine-learning-in-social-cognition\\\/#primaryimage\",\"url\":\"https:\\\/\\\/aisuperior.com\\\/wp-content\\\/uploads\\\/2026\\\/05\\\/unnamed-7-14.webp\",\"contentUrl\":\"https:\\\/\\\/aisuperior.com\\\/wp-content\\\/uploads\\\/2026\\\/05\\\/unnamed-7-14.webp\",\"width\":1168,\"height\":784},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/aisuperior.com\\\/machine-learning-in-social-cognition\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/aisuperior.com\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Machine Learning in Social Cognition: 2026 Guide\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/aisuperior.com\\\/#website\",\"url\":\"https:\\\/\\\/aisuperior.com\\\/\",\"name\":\"aisuperior\",\"description\":\"\",\"publisher\":{\"@id\":\"https:\\\/\\\/aisuperior.com\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/aisuperior.com\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"fr-FR\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/aisuperior.com\\\/#organization\",\"name\":\"aisuperior\",\"url\":\"https:\\\/\\\/aisuperior.com\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\\\/\\\/aisuperior.com\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/aisuperior.com\\\/wp-content\\\/uploads\\\/2026\\\/02\\\/logo-1.png.webp\",\"contentUrl\":\"https:\\\/\\\/aisuperior.com\\\/wp-content\\\/uploads\\\/2026\\\/02\\\/logo-1.png.webp\",\"width\":320,\"height\":59,\"caption\":\"aisuperior\"},\"image\":{\"@id\":\"https:\\\/\\\/aisuperior.com\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/www.facebook.com\\\/aisuperior\",\"https:\\\/\\\/x.com\\\/aisuperior\",\"https:\\\/\\\/www.linkedin.com\\\/company\\\/ai-superior\",\"https:\\\/\\\/www.instagram.com\\\/ai_superior\\\/\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/aisuperior.com\\\/#\\\/schema\\\/person\\\/14fcb7aaed4b2b617c4f75699394241c\",\"name\":\"kateryna\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\\\/\\\/aisuperior.com\\\/wp-content\\\/litespeed\\\/avatar\\\/6c451fec1b37608859459eb63b5a3380.jpg?ver=1781616670\",\"url\":\"https:\\\/\\\/aisuperior.com\\\/wp-content\\\/litespeed\\\/avatar\\\/6c451fec1b37608859459eb63b5a3380.jpg?ver=1781616670\",\"contentUrl\":\"https:\\\/\\\/aisuperior.com\\\/wp-content\\\/litespeed\\\/avatar\\\/6c451fec1b37608859459eb63b5a3380.jpg?ver=1781616670\",\"caption\":\"kateryna\"}}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Apprentissage automatique en cognition sociale : guide 2026","description":"D\u00e9couvrez comment l&#039;apprentissage automatique transforme la recherche en cognition sociale, de la pr\u00e9diction des comportements (AUC 0,80) \u00e0 la compr\u00e9hension de la th\u00e9orie de l&#039;esprit. Des perspectives fond\u00e9es sur des preuves.","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:\/\/aisuperior.com\/fr\/machine-learning-in-social-cognition\/","og_locale":"fr_FR","og_type":"article","og_title":"Machine Learning in Social Cognition: 2026 Guide","og_description":"Discover how machine learning transforms social cognition research, from predicting behavior (AUC 0.80) to understanding Theory of Mind. Evidence-based insights.","og_url":"https:\/\/aisuperior.com\/fr\/machine-learning-in-social-cognition\/","og_site_name":"aisuperior","article_publisher":"https:\/\/www.facebook.com\/aisuperior","article_published_time":"2026-05-27T11:33:54+00:00","og_image":[{"width":1168,"height":784,"url":"https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/05\/unnamed-7-14.webp","type":"image\/webp"}],"author":"kateryna","twitter_card":"summary_large_image","twitter_creator":"@aisuperior","twitter_site":"@aisuperior","twitter_misc":{"\u00c9crit par":"kateryna","Dur\u00e9e de lecture estim\u00e9e":"8 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/aisuperior.com\/machine-learning-in-social-cognition\/#article","isPartOf":{"@id":"https:\/\/aisuperior.com\/machine-learning-in-social-cognition\/"},"author":{"name":"kateryna","@id":"https:\/\/aisuperior.com\/#\/schema\/person\/14fcb7aaed4b2b617c4f75699394241c"},"headline":"Machine Learning in Social Cognition: 2026 Guide","datePublished":"2026-05-27T11:33:54+00:00","mainEntityOfPage":{"@id":"https:\/\/aisuperior.com\/machine-learning-in-social-cognition\/"},"wordCount":1574,"publisher":{"@id":"https:\/\/aisuperior.com\/#organization"},"image":{"@id":"https:\/\/aisuperior.com\/machine-learning-in-social-cognition\/#primaryimage"},"thumbnailUrl":"https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/05\/unnamed-7-14.webp","articleSection":["Blog"],"inLanguage":"fr-FR"},{"@type":"WebPage","@id":"https:\/\/aisuperior.com\/machine-learning-in-social-cognition\/","url":"https:\/\/aisuperior.com\/machine-learning-in-social-cognition\/","name":"Apprentissage automatique en cognition sociale : guide 2026","isPartOf":{"@id":"https:\/\/aisuperior.com\/#website"},"primaryImageOfPage":{"@id":"https:\/\/aisuperior.com\/machine-learning-in-social-cognition\/#primaryimage"},"image":{"@id":"https:\/\/aisuperior.com\/machine-learning-in-social-cognition\/#primaryimage"},"thumbnailUrl":"https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/05\/unnamed-7-14.webp","datePublished":"2026-05-27T11:33:54+00:00","description":"D\u00e9couvrez comment l&#039;apprentissage automatique transforme la recherche en cognition sociale, de la pr\u00e9diction des comportements (AUC 0,80) \u00e0 la compr\u00e9hension de la th\u00e9orie de l&#039;esprit. Des perspectives fond\u00e9es sur des preuves.","breadcrumb":{"@id":"https:\/\/aisuperior.com\/machine-learning-in-social-cognition\/#breadcrumb"},"inLanguage":"fr-FR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/aisuperior.com\/machine-learning-in-social-cognition\/"]}]},{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/aisuperior.com\/machine-learning-in-social-cognition\/#primaryimage","url":"https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/05\/unnamed-7-14.webp","contentUrl":"https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/05\/unnamed-7-14.webp","width":1168,"height":784},{"@type":"BreadcrumbList","@id":"https:\/\/aisuperior.com\/machine-learning-in-social-cognition\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/aisuperior.com\/"},{"@type":"ListItem","position":2,"name":"Machine Learning in Social Cognition: 2026 Guide"}]},{"@type":"WebSite","@id":"https:\/\/aisuperior.com\/#website","url":"https:\/\/aisuperior.com\/","name":"aisuperior","description":"","publisher":{"@id":"https:\/\/aisuperior.com\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/aisuperior.com\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"fr-FR"},{"@type":"Organization","@id":"https:\/\/aisuperior.com\/#organization","name":"aisuperior","url":"https:\/\/aisuperior.com\/","logo":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/aisuperior.com\/#\/schema\/logo\/image\/","url":"https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/02\/logo-1.png.webp","contentUrl":"https:\/\/aisuperior.com\/wp-content\/uploads\/2026\/02\/logo-1.png.webp","width":320,"height":59,"caption":"aisuperior"},"image":{"@id":"https:\/\/aisuperior.com\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/aisuperior","https:\/\/x.com\/aisuperior","https:\/\/www.linkedin.com\/company\/ai-superior","https:\/\/www.instagram.com\/ai_superior\/"]},{"@type":"Person","@id":"https:\/\/aisuperior.com\/#\/schema\/person\/14fcb7aaed4b2b617c4f75699394241c","name":"Katerina","image":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/aisuperior.com\/wp-content\/litespeed\/avatar\/6c451fec1b37608859459eb63b5a3380.jpg?ver=1781616670","url":"https:\/\/aisuperior.com\/wp-content\/litespeed\/avatar\/6c451fec1b37608859459eb63b5a3380.jpg?ver=1781616670","contentUrl":"https:\/\/aisuperior.com\/wp-content\/litespeed\/avatar\/6c451fec1b37608859459eb63b5a3380.jpg?ver=1781616670","caption":"kateryna"}}]}},"_links":{"self":[{"href":"https:\/\/aisuperior.com\/fr\/wp-json\/wp\/v2\/posts\/37425","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/aisuperior.com\/fr\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/aisuperior.com\/fr\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/aisuperior.com\/fr\/wp-json\/wp\/v2\/users\/7"}],"replies":[{"embeddable":true,"href":"https:\/\/aisuperior.com\/fr\/wp-json\/wp\/v2\/comments?post=37425"}],"version-history":[{"count":1,"href":"https:\/\/aisuperior.com\/fr\/wp-json\/wp\/v2\/posts\/37425\/revisions"}],"predecessor-version":[{"id":37428,"href":"https:\/\/aisuperior.com\/fr\/wp-json\/wp\/v2\/posts\/37425\/revisions\/37428"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/aisuperior.com\/fr\/wp-json\/wp\/v2\/media\/37426"}],"wp:attachment":[{"href":"https:\/\/aisuperior.com\/fr\/wp-json\/wp\/v2\/media?parent=37425"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/aisuperior.com\/fr\/wp-json\/wp\/v2\/categories?post=37425"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/aisuperior.com\/fr\/wp-json\/wp\/v2\/tags?post=37425"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}