{"id":9415,"date":"2022-06-07T10:28:13","date_gmt":"2022-06-07T14:28:13","guid":{"rendered":"https:\/\/www.nuecho.com\/?p=9415"},"modified":"2022-06-07T12:37:29","modified_gmt":"2022-06-07T16:37:29","slug":"exploration-dapproches-pour-un-systeme-de-questions-reponses","status":"publish","type":"post","link":"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/","title":{"rendered":"Exploration d\u2019approches pour un syst\u00e8me de questions-r\u00e9ponses"},"content":{"rendered":"

D\u00e9finition du probl\u00e8me<\/span><\/h2>\n

La litt\u00e9rature scientifique pr\u00e9sente plusieurs fa\u00e7ons d\u2019aborder le probl\u00e8me, mais nous nous sommes int\u00e9ress\u00e9s plus sp\u00e9cifiquement \u00e0 la t\u00e2che de s\u00e9lection de r\u00e9ponse. Cette t\u00e2che vise \u00e0 pr\u00e9dire la r\u00e9ponse correcte parmi un ensemble de r\u00e9ponses candidates.<\/span><\/p>\n

\"\"<\/p>\n

Elle suppose qu\u2019il existe toujours une r\u00e9ponse correcte pour chaque question. Cependant, dans un syst\u00e8me r\u00e9el de questions-r\u00e9ponses, on souhaite parfois ne pas fournir de r\u00e9ponse, par exemple si un utilisateur pose une question hors domaine. La t\u00e2che de s\u00e9lection optionnelle de r\u00e9ponse (<\/span>answer triggering<\/span><\/i>) offre cette possibilit\u00e9.<\/span><\/p>\n

\"\"<\/p>\n

D\u00e9finition du syst\u00e8me<\/span><\/h2>\n

Pour accomplir la t\u00e2che de s\u00e9lection optionnelle de r\u00e9ponse pour une question donn\u00e9e, l\u2019impl\u00e9mentation choisie ex\u00e9cute deux sous-t\u00e2ches:<\/span><\/p>\n

    \n
  1. Un mod\u00e8le d\u2019apprentissage automatique (un classifieur) accepte en entr\u00e9e une repr\u00e9sentation vectorielle de la question et retourne les probabilit\u00e9s par classe. Chaque classe est associ\u00e9e \u00e0 une paire question-r\u00e9ponse.<\/span><\/li>\n
  2. La probabilit\u00e9 la plus \u00e9lev\u00e9e est compar\u00e9e \u00e0 un seuil pour d\u00e9terminer si la r\u00e9ponse sera retourn\u00e9e ou non.<\/span><\/li>\n<\/ol>\n

    \"\"<\/p>\n

    Des exp\u00e9riences ont ensuite \u00e9t\u00e9 r\u00e9alis\u00e9es pour trouver la meilleure combinaison de mod\u00e8le de vectorisation et de classifieur pour r\u00e9aliser la premi\u00e8re sous-t\u00e2che.<\/span><\/p>\n

     <\/p>\n

    Exp\u00e9riences<\/span><\/h2>\n

    Pour les exp\u00e9riences, l\u2019ensemble de donn\u00e9es bancaires <\/span>BANKING77<\/span><\/a>, cr\u00e9\u00e9 par l\u2019entreprise de solutions conversationnelles PolyAI, a \u00e9t\u00e9 utilis\u00e9. Les combinaisons de mod\u00e8le de vectorisation et de classifieur ont \u00e9t\u00e9 \u00e9valu\u00e9es sur l\u2019ensemble de test avec la m\u00e9trique d\u2019exactitude (<\/span>accuracy<\/span><\/i>), qui calcule le pourcentage de pr\u00e9dictions correctes.<\/span><\/p>\n

     <\/p>\n

    Mod\u00e8les de vectorisation<\/span><\/h3>\n

    Parmi les diff\u00e9rents mod\u00e8les de vectorisation, celui qui a le mieux perform\u00e9 est le Universal Sentence Encoder (USE) de Google. C\u2019est un r\u00e9seau de neurones pr\u00e9entra\u00een\u00e9 simultan\u00e9ment sur plusieurs t\u00e2ches de nature s\u00e9mantique qui accepte en entr\u00e9e un texte et produit en sortie un plongement de phrase (une repr\u00e9sentation vectorielle de la phrase). Le pr\u00e9entra\u00eenement est fait sur de tr\u00e8s grands corpus de texte comme Wikip\u00e9dia, ce qui lui permet de bien capturer la similarit\u00e9 s\u00e9mantique de phrases jamais vues au pr\u00e9alable, comme on peut voir dans l\u2019exemple ci-dessous.<\/span><\/p>\n

    \"\"<\/p>\n

    Similarit\u00e9 s\u00e9mantique de phrases tir\u00e9es de BANKING77 avec USE. <\/span><\/i>R\u00e9f\u00e9rence<\/span><\/i><\/a><\/p>\n

     <\/p>\n

    Classifieurs<\/span><\/h3>\n

    Parmi les classifieurs \u00e9valu\u00e9s, on compte notamment un classifieur des k plus proches voisins (KNN) et un r\u00e9seau de neurones. L\u2019avantage principal du KNN par rapport au r\u00e9seau de neurones est qu\u2019il n\u2019a pas besoin d\u2019\u00eatre entra\u00een\u00e9, il ne fait que m\u00e9moriser les donn\u00e9es d\u2019entra\u00eenement. Ajouter de nouvelles questions au mod\u00e8le ne n\u00e9cessite donc pas un r\u00e9entra\u00eenement. Un autre avantage est que ses pr\u00e9dictions sont interpr\u00e9tables. Pour pr\u00e9dire la classe d\u2019un exemple de test, le KNN trouve ses k plus proches voisins et retourne la classe majoritaire. Le nombre de voisins ainsi que la fonction de distance sont des hyperparam\u00e8tres configurables.<\/span><\/p>\n

    Pour illustrer le fonctionnement du KNN, un exemple simplifi\u00e9 est fourni ci-dessous pour un probl\u00e8me de classification binaire avec des donn\u00e9es en 2 dimensions.<\/span><\/p>\n

    \"\"<\/p>\n

    Exemple pour un KNN avec k=3 et distance euclidienne. Pour l\u2019exemple de test en gris, on
    \n<\/span><\/i>pr\u00e9dira la classe en bleu (classe majoritaire parmi les 3 plus proches voisins).<\/span><\/i><\/p>\n

     <\/p>\n

    Dans un premier temps, des exp\u00e9riences ont \u00e9t\u00e9 r\u00e9alis\u00e9es avec un KNN avec <\/span>distance cosinus<\/span><\/a>. Dans un deuxi\u00e8me temps, d\u2019autres exp\u00e9riences ont \u00e9t\u00e9 r\u00e9alis\u00e9es avec une fonction de distance apprise (<\/span>metric learning<\/span><\/i><\/a>) dans le but d\u2019am\u00e9liorer la performance. L\u2019algorithme utilis\u00e9 avait pour but de rapprocher les exemples d\u2019une m\u00eame classe et \u00e9loigner les exemples appartenant \u00e0 des classes diff\u00e9rentes. Dans les deux cas, l\u2019exactitude obtenue avec le r\u00e9seau de neurones \u00e9tait sup\u00e9rieure, ce qui fait en sorte que ce dernier a \u00e9t\u00e9 choisi comme classifieur.<\/span><\/p>\n

     <\/p>\n

    \u00c9valuation du syst\u00e8me<\/span><\/h2>\n

    Les exp\u00e9riences d\u00e9crites pr\u00e9c\u00e9demment ont permis d\u2019\u00e9tablir que le meilleur mod\u00e8le \u00e9tait celui combinant le Universal Sentence Encoder avec un MLP. Pour \u00e9valuer ce mod\u00e8le, une comparaison a \u00e9t\u00e9 effectu\u00e9e sur la t\u00e2che de s\u00e9lection optionnelle de r\u00e9ponse avec les mod\u00e8les de classification d\u2019intentions des engins NLU de Dialogflow ES et Rasa. Pour ce faire, une nouvelle intention \u201cout_of_scope\u201d (hors domaine) contenant <\/span>des questions sur la COVID-19<\/span><\/a> a \u00e9t\u00e9 ajout\u00e9e \u00e0 l\u2019ensemble de test de BANKING77. Pour tous les mod\u00e8les, rejeter des exemples hors domaine s\u2019est av\u00e9r\u00e9 plus difficile que bien classifier les exemples appartenant au domaine. Cependant, dans l\u2019ensemble, c\u2019est le mod\u00e8le USE combin\u00e9 \u00e0 un r\u00e9seau de neurones qui s\u2019est d\u00e9marqu\u00e9. Cette \u00e9valuation a donc d\u00e9montr\u00e9 que ce mod\u00e8le peut \u00eatre utilis\u00e9 pour d\u00e9velopper un syst\u00e8me de questions-r\u00e9ponses performant et efficace.<\/span><\/p>\n

    Pour plus de d\u00e9tails sur les mod\u00e8les utilis\u00e9s, la m\u00e9thodologie suivie et les r\u00e9sultats des exp\u00e9riences, nous vous invitons \u00e0 consulter cet article<\/a>.<\/span><\/p>\n","protected":false},"excerpt":{"rendered":"

    D\u00e9finition du probl\u00e8me La litt\u00e9rature scientifique pr\u00e9sente plusieurs fa\u00e7ons d\u2019aborder le probl\u00e8me, mais nous nous sommes int\u00e9ress\u00e9s plus sp\u00e9cifiquement \u00e0 la t\u00e2che de s\u00e9lection de r\u00e9ponse. Cette t\u00e2che vise \u00e0 pr\u00e9dire la r\u00e9ponse correcte parmi un ensemble de r\u00e9ponses candidates. Elle suppose qu\u2019il existe toujours une r\u00e9ponse correcte pour chaque question. Cependant, dans un syst\u00e8me […]<\/p>\n","protected":false},"author":20,"featured_media":9412,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_et_pb_use_builder":"","_et_pb_old_content":"","_et_gb_content_width":"","om_disable_all_campaigns":false,"_monsterinsights_skip_tracking":false,"_monsterinsights_sitenote_active":false,"_monsterinsights_sitenote_note":"","_monsterinsights_sitenote_category":0,"footnotes":""},"categories":[5],"tags":[],"class_list":["post-9415","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-blogue"],"aioseo_notices":[],"yoast_head":"\nExploration d\u2019approches pour un syst\u00e8me de questions-r\u00e9ponses · AI Virtual Voice Experts with Google Dialogflow CX - CCAI - Nu Echo<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/\" \/>\n<meta property=\"og:locale\" content=\"fr_FR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Exploration d\u2019approches pour un syst\u00e8me de questions-r\u00e9ponses · AI Virtual Voice Experts with Google Dialogflow CX - CCAI - Nu Echo\" \/>\n<meta property=\"og:description\" content=\"D\u00e9finition du probl\u00e8me La litt\u00e9rature scientifique pr\u00e9sente plusieurs fa\u00e7ons d\u2019aborder le probl\u00e8me, mais nous nous sommes int\u00e9ress\u00e9s plus sp\u00e9cifiquement \u00e0 la t\u00e2che de s\u00e9lection de r\u00e9ponse. Cette t\u00e2che vise \u00e0 pr\u00e9dire la r\u00e9ponse correcte parmi un ensemble de r\u00e9ponses candidates. Elle suppose qu\u2019il existe toujours une r\u00e9ponse correcte pour chaque question. Cependant, dans un syst\u00e8me […]\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/\" \/>\n<meta property=\"og:site_name\" content=\"AI Virtual Voice Experts with Google Dialogflow CX - CCAI - Nu Echo\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/nuechoinc\/\" \/>\n<meta property=\"article:published_time\" content=\"2022-06-07T14:28:13+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2022-06-07T16:37:29+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.nuecho.com\/wp-content\/uploads\/2022\/06\/Exploration-dapproches-pour-un-systeme-de-questions-reponses.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"1600\" \/>\n\t<meta property=\"og:image:height\" content=\"800\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"Laurence Dupont\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@nuecho\" \/>\n<meta name=\"twitter:site\" content=\"@nuecho\" \/>\n<meta name=\"twitter:label1\" content=\"\u00c9crit par\" \/>\n\t<meta name=\"twitter:data1\" content=\"Laurence Dupont\" \/>\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\":\"Article\",\"@id\":\"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/\"},\"author\":{\"name\":\"Laurence Dupont\",\"@id\":\"https:\/\/www.nuecho.com\/fr\/#\/schema\/person\/5f4b14f0aa400b3d4bcb9ce7b60d2708\"},\"headline\":\"Exploration d\u2019approches pour un syst\u00e8me de questions-r\u00e9ponses\",\"datePublished\":\"2022-06-07T14:28:13+00:00\",\"dateModified\":\"2022-06-07T16:37:29+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/\"},\"wordCount\":902,\"publisher\":{\"@id\":\"https:\/\/www.nuecho.com\/fr\/#organization\"},\"image\":{\"@id\":\"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/www.nuecho.com\/wp-content\/uploads\/2022\/06\/Exploration-dapproches-pour-un-systeme-de-questions-reponses.jpg\",\"articleSection\":[\"Blogue\"],\"inLanguage\":\"fr-FR\"},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/\",\"url\":\"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/\",\"name\":\"Exploration d\u2019approches pour un syst\u00e8me de questions-r\u00e9ponses · AI Virtual Voice Experts with Google Dialogflow CX - CCAI - Nu Echo\",\"isPartOf\":{\"@id\":\"https:\/\/www.nuecho.com\/fr\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/#primaryimage\"},\"image\":{\"@id\":\"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/www.nuecho.com\/wp-content\/uploads\/2022\/06\/Exploration-dapproches-pour-un-systeme-de-questions-reponses.jpg\",\"datePublished\":\"2022-06-07T14:28:13+00:00\",\"dateModified\":\"2022-06-07T16:37:29+00:00\",\"breadcrumb\":{\"@id\":\"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/#breadcrumb\"},\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/#primaryimage\",\"url\":\"https:\/\/www.nuecho.com\/wp-content\/uploads\/2022\/06\/Exploration-dapproches-pour-un-systeme-de-questions-reponses.jpg\",\"contentUrl\":\"https:\/\/www.nuecho.com\/wp-content\/uploads\/2022\/06\/Exploration-dapproches-pour-un-systeme-de-questions-reponses.jpg\",\"width\":1600,\"height\":800},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/www.nuecho.com\/fr\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Exploration d\u2019approches pour un syst\u00e8me de questions-r\u00e9ponses\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.nuecho.com\/fr\/#website\",\"url\":\"https:\/\/www.nuecho.com\/fr\/\",\"name\":\"AI Virtual Voice Experts with Google Dialogflow CX - CCAI - Nu Echo\",\"description\":\"Nu Echo\",\"publisher\":{\"@id\":\"https:\/\/www.nuecho.com\/fr\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.nuecho.com\/fr\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"fr-FR\"},{\"@type\":\"Organization\",\"@id\":\"https:\/\/www.nuecho.com\/fr\/#organization\",\"name\":\"Nu Echo - Conversation AI | IA Conversationnelle\",\"url\":\"https:\/\/www.nuecho.com\/fr\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\/\/www.nuecho.com\/fr\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/www.nuecho.com\/wp-content\/uploads\/2019\/03\/bottom-logo.png\",\"contentUrl\":\"https:\/\/www.nuecho.com\/wp-content\/uploads\/2019\/03\/bottom-logo.png\",\"width\":138,\"height\":27,\"caption\":\"Nu Echo - Conversation AI | IA Conversationnelle\"},\"image\":{\"@id\":\"https:\/\/www.nuecho.com\/fr\/#\/schema\/logo\/image\/\"},\"sameAs\":[\"https:\/\/www.facebook.com\/nuechoinc\/\",\"https:\/\/x.com\/nuecho\",\"https:\/\/ca.linkedin.com\/company\/nu-echo\"]},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.nuecho.com\/fr\/#\/schema\/person\/5f4b14f0aa400b3d4bcb9ce7b60d2708\",\"name\":\"Laurence Dupont\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\/\/www.nuecho.com\/fr\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/www.nuecho.com\/wp-content\/uploads\/2019\/07\/NuEcho_1704-214731_Web-150x150.jpg\",\"contentUrl\":\"https:\/\/www.nuecho.com\/wp-content\/uploads\/2019\/07\/NuEcho_1704-214731_Web-150x150.jpg\",\"caption\":\"Laurence Dupont\"},\"description\":\"Conceptrice de logiciel, passionn\u00e9e par l'apprentissage automatique et l'innovation conversationnelle. Software developer, passionate about machine learning and conversational innovation.\",\"url\":\"https:\/\/www.nuecho.com\/fr\/author\/ldupont\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Exploration d\u2019approches pour un syst\u00e8me de questions-r\u00e9ponses · AI Virtual Voice Experts with Google Dialogflow CX - CCAI - Nu Echo","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:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/","og_locale":"fr_FR","og_type":"article","og_title":"Exploration d\u2019approches pour un syst\u00e8me de questions-r\u00e9ponses · AI Virtual Voice Experts with Google Dialogflow CX - CCAI - Nu Echo","og_description":"D\u00e9finition du probl\u00e8me La litt\u00e9rature scientifique pr\u00e9sente plusieurs fa\u00e7ons d\u2019aborder le probl\u00e8me, mais nous nous sommes int\u00e9ress\u00e9s plus sp\u00e9cifiquement \u00e0 la t\u00e2che de s\u00e9lection de r\u00e9ponse. Cette t\u00e2che vise \u00e0 pr\u00e9dire la r\u00e9ponse correcte parmi un ensemble de r\u00e9ponses candidates. Elle suppose qu\u2019il existe toujours une r\u00e9ponse correcte pour chaque question. Cependant, dans un syst\u00e8me […]","og_url":"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/","og_site_name":"AI Virtual Voice Experts with Google Dialogflow CX - CCAI - Nu Echo","article_publisher":"https:\/\/www.facebook.com\/nuechoinc\/","article_published_time":"2022-06-07T14:28:13+00:00","article_modified_time":"2022-06-07T16:37:29+00:00","og_image":[{"width":1600,"height":800,"url":"https:\/\/www.nuecho.com\/wp-content\/uploads\/2022\/06\/Exploration-dapproches-pour-un-systeme-de-questions-reponses.jpg","type":"image\/jpeg"}],"author":"Laurence Dupont","twitter_card":"summary_large_image","twitter_creator":"@nuecho","twitter_site":"@nuecho","twitter_misc":{"\u00c9crit par":"Laurence Dupont","Dur\u00e9e de lecture estim\u00e9e":"5 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/#article","isPartOf":{"@id":"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/"},"author":{"name":"Laurence Dupont","@id":"https:\/\/www.nuecho.com\/fr\/#\/schema\/person\/5f4b14f0aa400b3d4bcb9ce7b60d2708"},"headline":"Exploration d\u2019approches pour un syst\u00e8me de questions-r\u00e9ponses","datePublished":"2022-06-07T14:28:13+00:00","dateModified":"2022-06-07T16:37:29+00:00","mainEntityOfPage":{"@id":"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/"},"wordCount":902,"publisher":{"@id":"https:\/\/www.nuecho.com\/fr\/#organization"},"image":{"@id":"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/#primaryimage"},"thumbnailUrl":"https:\/\/www.nuecho.com\/wp-content\/uploads\/2022\/06\/Exploration-dapproches-pour-un-systeme-de-questions-reponses.jpg","articleSection":["Blogue"],"inLanguage":"fr-FR"},{"@type":"WebPage","@id":"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/","url":"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/","name":"Exploration d\u2019approches pour un syst\u00e8me de questions-r\u00e9ponses · AI Virtual Voice Experts with Google Dialogflow CX - CCAI - Nu Echo","isPartOf":{"@id":"https:\/\/www.nuecho.com\/fr\/#website"},"primaryImageOfPage":{"@id":"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/#primaryimage"},"image":{"@id":"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/#primaryimage"},"thumbnailUrl":"https:\/\/www.nuecho.com\/wp-content\/uploads\/2022\/06\/Exploration-dapproches-pour-un-systeme-de-questions-reponses.jpg","datePublished":"2022-06-07T14:28:13+00:00","dateModified":"2022-06-07T16:37:29+00:00","breadcrumb":{"@id":"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/#breadcrumb"},"inLanguage":"fr-FR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/"]}]},{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/#primaryimage","url":"https:\/\/www.nuecho.com\/wp-content\/uploads\/2022\/06\/Exploration-dapproches-pour-un-systeme-de-questions-reponses.jpg","contentUrl":"https:\/\/www.nuecho.com\/wp-content\/uploads\/2022\/06\/Exploration-dapproches-pour-un-systeme-de-questions-reponses.jpg","width":1600,"height":800},{"@type":"BreadcrumbList","@id":"https:\/\/www.nuecho.com\/fr\/exploration-dapproches-pour-un-systeme-de-questions-reponses\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/www.nuecho.com\/fr\/"},{"@type":"ListItem","position":2,"name":"Exploration d\u2019approches pour un syst\u00e8me de questions-r\u00e9ponses"}]},{"@type":"WebSite","@id":"https:\/\/www.nuecho.com\/fr\/#website","url":"https:\/\/www.nuecho.com\/fr\/","name":"AI Virtual Voice Experts with Google Dialogflow CX - CCAI - Nu Echo","description":"Nu Echo","publisher":{"@id":"https:\/\/www.nuecho.com\/fr\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.nuecho.com\/fr\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"fr-FR"},{"@type":"Organization","@id":"https:\/\/www.nuecho.com\/fr\/#organization","name":"Nu Echo - Conversation AI | IA Conversationnelle","url":"https:\/\/www.nuecho.com\/fr\/","logo":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/www.nuecho.com\/fr\/#\/schema\/logo\/image\/","url":"https:\/\/www.nuecho.com\/wp-content\/uploads\/2019\/03\/bottom-logo.png","contentUrl":"https:\/\/www.nuecho.com\/wp-content\/uploads\/2019\/03\/bottom-logo.png","width":138,"height":27,"caption":"Nu Echo - Conversation AI | IA Conversationnelle"},"image":{"@id":"https:\/\/www.nuecho.com\/fr\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/nuechoinc\/","https:\/\/x.com\/nuecho","https:\/\/ca.linkedin.com\/company\/nu-echo"]},{"@type":"Person","@id":"https:\/\/www.nuecho.com\/fr\/#\/schema\/person\/5f4b14f0aa400b3d4bcb9ce7b60d2708","name":"Laurence Dupont","image":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/www.nuecho.com\/fr\/#\/schema\/person\/image\/","url":"https:\/\/www.nuecho.com\/wp-content\/uploads\/2019\/07\/NuEcho_1704-214731_Web-150x150.jpg","contentUrl":"https:\/\/www.nuecho.com\/wp-content\/uploads\/2019\/07\/NuEcho_1704-214731_Web-150x150.jpg","caption":"Laurence Dupont"},"description":"Conceptrice de logiciel, passionn\u00e9e par l'apprentissage automatique et l'innovation conversationnelle. Software developer, passionate about machine learning and conversational innovation.","url":"https:\/\/www.nuecho.com\/fr\/author\/ldupont\/"}]}},"_links":{"self":[{"href":"https:\/\/www.nuecho.com\/fr\/wp-json\/wp\/v2\/posts\/9415","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.nuecho.com\/fr\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.nuecho.com\/fr\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.nuecho.com\/fr\/wp-json\/wp\/v2\/users\/20"}],"replies":[{"embeddable":true,"href":"https:\/\/www.nuecho.com\/fr\/wp-json\/wp\/v2\/comments?post=9415"}],"version-history":[{"count":9,"href":"https:\/\/www.nuecho.com\/fr\/wp-json\/wp\/v2\/posts\/9415\/revisions"}],"predecessor-version":[{"id":9457,"href":"https:\/\/www.nuecho.com\/fr\/wp-json\/wp\/v2\/posts\/9415\/revisions\/9457"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.nuecho.com\/fr\/wp-json\/wp\/v2\/media\/9412"}],"wp:attachment":[{"href":"https:\/\/www.nuecho.com\/fr\/wp-json\/wp\/v2\/media?parent=9415"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.nuecho.com\/fr\/wp-json\/wp\/v2\/categories?post=9415"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.nuecho.com\/fr\/wp-json\/wp\/v2\/tags?post=9415"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}