pr\u00e9sentent<\/p>\n\n\n\n
Autoris\u00e9e \u00e0 pr\u00e9senter ses travaux en vue de l\u2019obtention du Doctorat de l’Institut Polytechnique de Paris, pr\u00e9par\u00e9 \u00e0 T\u00e9l\u00e9com SudParis en :<\/p>\n\n\n\n
le LUNDI 28 NOVEMBRE 2022 \u00e0 14h00<\/p>\n\n\n\n
Salle des Conseils – 7\u00e8me \u00e9tage
Universit\u00e9 Paris Cit\u00e9 – UFR Math\u00e9matique et Informatique – Laboratoire d’Informatique Paris Descartes 45 Rue des Saints-P\u00e8res, 75006 Paris<\/p>\n\n\n\n
Membres du jury :<\/strong><\/p>\n\n\n\n M. Hossam AFIFI<\/strong>, Professeur, T\u00e9l\u00e9com SudParis, FRANCE – Directeur de th\u00e8se La croissance massive des capteurs (temp\u00e9rature, humidit\u00e9, acc\u00e9l\u00e9rom\u00e8tre, capteur de position) et des appareils mobiles (smartphones, tablettes, smartwatch \u2026) fait que la quantit\u00e9 de donn\u00e9es g\u00e9n\u00e9r\u00e9es augmente de mani\u00e8re explosive. Cette immense quantit\u00e9 de donn\u00e9es peut \u00eatre collect\u00e9e et g\u00e9r\u00e9e. Le travail r\u00e9alis\u00e9 durant cette th\u00e8se vise \u00e0 proposer en un premier temps une approche qui traite un type de donn\u00e9es sp\u00e9cifique qui sont les s\u00e9ries temporelles. Pour ce faire nous avons utilis\u00e9 des m\u00e9thodes de classification bas\u00e9es sur des r\u00e9seaux de neurones convolutifs ainsi que des multi layer perceptron afin d\u2019extraire les informations pertinentes. Nous avons par la suite eu recours \u00e0 l\u2019utilisation des r\u00e9seaux de neurones r\u00e9currents pour r\u00e9aliser les pr\u00e9dictions. Les donn\u00e9es utilis\u00e9es provenaient de plusieurs sources : Donn\u00e9es m\u00e9t\u00e9orologiques, donn\u00e9es de consommation \u00e9nerg\u00e9tique, donn\u00e9es de production d\u2019\u00e9nergies renouvelables, donn\u00e9es cellulaires, donn\u00e9es de trace GPS de taxi. Nous avons \u00e9galement investigu\u00e9 plusieurs autres m\u00e9thodes telles que la compression s\u00e9mantique ainsi que le transfer learning. Les deux m\u00e9thodes d\u00e9crites pr\u00e9c\u00e9demment nous permettent pour la premi\u00e8re de ne transmettre que les poids des r\u00e9seaux de neurones ou en cas d’anomalie d\u00e9tect\u00e9e d’envoyer les donn\u00e9es la constituant. Le transfer learning nous permet quand \u00e0 lui de r\u00e9aliser de bonne pr\u00e9diction m\u00eame si les donn\u00e9es trait\u00e9es souffrent d’un manque ou d’un bruit. Ces traitements nous ont permis par la suite de mettre en place des m\u00e9canismes dynamique de d\u00e9tection d\u2019anomalie. L\u2019objectif du dernier volet de la th\u00e8se est le d\u00e9veloppement et l\u2019impl\u00e9mentation d\u2019une solution de management des ressources en ayant comme entr\u00e9e le r\u00e9sultat des phases pr\u00e9c\u00e9dentes. Pour mettre en place cette solution de gestion des ressources nous avons utilis\u00e9 plusieurs approches tel que l\u2019apprentissage par renforcement, la r\u00e9solution exacte ou encore des r\u00e9seaux de neurones r\u00e9currents. Une premi\u00e8re application est la mise en place d\u2019un syst\u00e8me de management de l\u2019\u00e9nergie et la seconde est la gestion du d\u00e9ploiement des drones pour assister les r\u00e9seaux cellulaires en cas d\u2019anomalies.<\/p>\n\n\n\n The massive growth of sensors (temperature, humidity, accelerometer, position sensor) and mobile devices (smartphones, tablets, smartwatch\u2026) increases the amount of data generated explosively. This immense amount of data can be collected and managed. The work carried out during this thesis aims to first propose an approach that deals with a specific type of data, which are time series. To do this, we used classification methods based on convolutional neural networks as well as multilayer perceptrons in order to extract the relevant information. We then used recurrent neural networks to make the predictions. The data used came from several sources: Weather data, energy consumption data, renewable energy production data, cellular data, taxi GPS track data. We also investigated several other methods such as semantic compression and transfer learning. The two methods described above allow us for the first to transmit only the weight of the neural networks or if an anomaly is detected, send the data. The transfer learning allows us to make good predictions even if the data is missing or noisy. These treatments, allowed us to set up dynamic anomaly detection mechanisms. The objective of the last part of the thesis is the development and implementation of a resource management solution having as input the result of the previous phases. To implement this resource management solution, we used several approaches such as reinforcement learning, exact resolution or even recurrent neural networks. A first application is the implementation of an energy management system and the second is the management of the deployment of drones to assist cellular networks when an anomaly occurs.<\/p>\n","protected":false},"excerpt":{"rendered":" L’Ecole doctorale : Ecole Doctorale de l’Institut Polytechnique de Pariset le Laboratoire de recherche SAMOVAR – Services r\u00e9partis, Architectures, MOd\u00e9lisation, Validation, Administration des R\u00e9seaux pr\u00e9sentent l\u2019AVIS DE SOUTENANCE de Madame Aicha DRIDI Autoris\u00e9e \u00e0 pr\u00e9senter ses travaux en vue de l\u2019obtention du Doctorat de l’Institut Polytechnique de Paris, pr\u00e9par\u00e9 \u00e0 T\u00e9l\u00e9com SudParis en : Informatique […]<\/p>\n","protected":false},"author":4,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ocean_post_layout":"","ocean_both_sidebars_style":"","ocean_both_sidebars_content_width":0,"ocean_both_sidebars_sidebars_width":0,"ocean_sidebar":"0","ocean_second_sidebar":"0","ocean_disable_margins":"enable","ocean_add_body_class":"","ocean_shortcode_before_top_bar":"","ocean_shortcode_after_top_bar":"","ocean_shortcode_before_header":"","ocean_shortcode_after_header":"","ocean_has_shortcode":"","ocean_shortcode_after_title":"","ocean_shortcode_before_footer_widgets":"","ocean_shortcode_after_footer_widgets":"","ocean_shortcode_before_footer_bottom":"","ocean_shortcode_after_footer_bottom":"","ocean_display_top_bar":"default","ocean_display_header":"default","ocean_header_style":"","ocean_center_header_left_menu":"0","ocean_custom_header_template":"0","ocean_custom_logo":0,"ocean_custom_retina_logo":0,"ocean_custom_logo_max_width":0,"ocean_custom_logo_tablet_max_width":0,"ocean_custom_logo_mobile_max_width":0,"ocean_custom_logo_max_height":0,"ocean_custom_logo_tablet_max_height":0,"ocean_custom_logo_mobile_max_height":0,"ocean_header_custom_menu":"0","ocean_menu_typo_font_family":"0","ocean_menu_typo_font_subset":"","ocean_menu_typo_font_size":0,"ocean_menu_typo_font_size_tablet":0,"ocean_menu_typo_font_size_mobile":0,"ocean_menu_typo_font_size_unit":"px","ocean_menu_typo_font_weight":"","ocean_menu_typo_font_weight_tablet":"","ocean_menu_typo_font_weight_mobile":"","ocean_menu_typo_transform":"","ocean_menu_typo_transform_tablet":"","ocean_menu_typo_transform_mobile":"","ocean_menu_typo_line_height":0,"ocean_menu_typo_line_height_tablet":0,"ocean_menu_typo_line_height_mobile":0,"ocean_menu_typo_line_height_unit":"","ocean_menu_typo_spacing":0,"ocean_menu_typo_spacing_tablet":0,"ocean_menu_typo_spacing_mobile":0,"ocean_menu_typo_spacing_unit":"","ocean_menu_link_color":"","ocean_menu_link_color_hover":"","ocean_menu_link_color_active":"","ocean_menu_link_background":"","ocean_menu_link_hover_background":"","ocean_menu_link_active_background":"","ocean_menu_social_links_bg":"","ocean_menu_social_hover_links_bg":"","ocean_menu_social_links_color":"","ocean_menu_social_hover_links_color":"","ocean_disable_title":"default","ocean_disable_heading":"default","ocean_post_title":"","ocean_post_subheading":"","ocean_post_title_style":"","ocean_post_title_background_color":"","ocean_post_title_background":0,"ocean_post_title_bg_image_position":"","ocean_post_title_bg_image_attachment":"","ocean_post_title_bg_image_repeat":"","ocean_post_title_bg_image_size":"","ocean_post_title_height":0,"ocean_post_title_bg_overlay":0.5,"ocean_post_title_bg_overlay_color":"","ocean_disable_breadcrumbs":"default","ocean_breadcrumbs_color":"","ocean_breadcrumbs_separator_color":"","ocean_breadcrumbs_links_color":"","ocean_breadcrumbs_links_hover_color":"","ocean_display_footer_widgets":"default","ocean_display_footer_bottom":"default","ocean_custom_footer_template":"0","ocean_post_oembed":"","ocean_post_self_hosted_media":"","ocean_post_video_embed":"","ocean_link_format":"","ocean_link_format_target":"self","ocean_quote_format":"","ocean_quote_format_link":"post","ocean_gallery_link_images":"off","ocean_gallery_id":[],"footnotes":""},"categories":[286,284],"tags":[],"class_list":["post-5044","post","type-post","status-publish","format-standard","hentry","category-fractualites-ennews-fr","category-seminaires-r3s-fr","entry"],"_links":{"self":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/5044","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/users\/4"}],"replies":[{"embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/comments?post=5044"}],"version-history":[{"count":2,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/5044\/revisions"}],"predecessor-version":[{"id":5046,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/5044\/revisions\/5046"}],"wp:attachment":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/media?parent=5044"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/categories?post=5044"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/tags?post=5044"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
M. Hassine MOUNGLA<\/strong>, Ma\u00eetre de conf\u00e9rences, Universit\u00e9 Paris Cit\u00e9, FRANCE – Co-encadrant de th\u00e8se
M. Enrico NATALIZIO<\/strong>, Professeur des universit\u00e9s, LORIA, FRANCE – Rapporteur
Mme Florence OSSART<\/strong>, Professeure des universit\u00e9s, Sorbonne Universit\u00e9, UPMC Universit\u00e9 Pierre et Mary Curie, FRANCE – Examinatrice
M. Yvon GOURHANT<\/strong>, Ing\u00e9nieur de recherche, Orange Labs, FRANCE – Examinateur
M. Ghislain AGOUA<\/strong>, Ing\u00e9nieur de recherche, EDF Labs, FRANCE – Examinateur
M. Marcelo DIAS DE AMORIM<\/strong>, Directeur de recherche, Universit\u00e9 Paris Sorbonne, FRANCE – Rapporteur<\/p>\n\n\n\n
R\u00e9sum\u00e9 :<\/strong><\/p>\n\n\n\n
Abstract : \u00ab\u00a0A novel efficient time series deep learning approach using classification,prediction and, reinforcement: Energy and Communication Use Cases\u00a0\u00bb<\/strong><\/p>\n\n\n\n