L’\u00e9quipe ARMEDIA organise un s\u00e9minaire le vendredi 6 juillet \u00e0 10h ouvert \u00e0 tous.
\nIl pourra \u00eatre suivi sur le site de TSP dans la salle A003 ou H218 (choix \u00e0 confirmer) et sur le site de Saclay dans la salle de visioconf\u00e9rence de NanoInnov .<\/p>\n
L’intervenant est le Pr. Faouzi Ghorbel (ENSI Tunis) et son expos\u00e9 porte sur :<\/p>\n
L’apprentissage Machine non lin\u00e9aire : Algorithmes non lin\u00e9aires et donn\u00e9es issues d’espaces non convexes<\/p>\n
Mots cl\u00e9s :<\/strong> Vari\u00e9t\u00e9 Riemannienne, Groupes de Lie, Classification, optimisation, invariant, Statistique intrins\u00e8que et extrins\u00e8que<\/p>\n R\u00e9sum\u00e9:<\/strong><\/p>\n Classiquement les algorithmes de classification sont appel\u00e9s non lin\u00e9aires quand les domaines relatifs aux diff\u00e9rentes classes sont en g\u00e9n\u00e9ral des parties de l’espace des observations s\u00e9par\u00e9es par des hyper surfaces non lin\u00e9aires. Toutefois, les donn\u00e9es \u00e0 classer peuvent appartenir \u00e0 des espaces non vectoriels (non lin\u00e9aires et m\u00eame non convexes) Il est souvent utile de conna\u00eetre leurs domaines de variations. Souvent, elles peuvent appartenir \u00e0 des vari\u00e9t\u00e9s diff\u00e9rentielles de type Riemmanniennes ou \u00e0 des groupes de Lie. La g\u00e9om\u00e9trie des , actuellement en vogue dans le domaine de la vision par ordinateur, est un des exemples des plus illustratifs Ainsi, nous essayerons de montrer ici sur des exemples concrets en reconnaissance de formes comment cette nouvelle g\u00e9n\u00e9ration d’algorithmes d’apprentissage tenant compte de la nature des supports des donn\u00e9es est en train de r\u00e9volutionner le domaine de la vision. D’autre part et \u00e0 partir de quelques exemples pr\u00eat\u00e9s au domaine de l\u2019apprentissage machine, nous tenterons de montrer comment les m\u00e9thodes d’optimisation s’am\u00e9liorent au sens pr\u00e9cision et cela en consid\u00e9rant que les donn\u00e9es proviennent d’espace de type vari\u00e9t\u00e9s.<\/p>\n Enfin, l’approche des invariants en analyse de formes, peut \u00eatre per\u00e7ue comme celle des statistiques extrins\u00e8ques alors que celle bas\u00e9e sur les espaces de Kendall compte dans la cat\u00e9gorie des statistiques intrins\u00e8ques. Nous les comparons dans les contextes de classification de formes (Euclidienne-contours plans, Affine-contours et enfin Euclidien-surfaces courb\u00e9es.<\/p>\n English summary ————————-<\/p>\n Nonlinear Machine Learning Algorithms and Data from Non-convex Spaces<\/p>\n Key words: Riemannian Manifolds, Lie groups, Classification, optimization, invariant, intrinsic and extrinsic statistics<\/p>\n Abstract: L’\u00e9quipe ARMEDIA organise un s\u00e9minaire le vendredi 6 juillet \u00e0 10h ouvert \u00e0 tous. Il pourra \u00eatre suivi sur le site de TSP dans la salle A003 ou H218 (choix \u00e0 confirmer) et sur le site de Saclay dans la salle de visioconf\u00e9rence de NanoInnov . L’intervenant est le Pr. Faouzi Ghorbel (ENSI Tunis) et […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","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":[169],"tags":[],"class_list":["post-1057","post","type-post","status-publish","format-standard","hentry","category-seminaires-armedia","entry"],"_links":{"self":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/1057","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\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/comments?post=1057"}],"version-history":[{"count":2,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/1057\/revisions"}],"predecessor-version":[{"id":5348,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/1057\/revisions\/5348"}],"wp:attachment":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/media?parent=1057"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/categories?post=1057"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/tags?post=1057"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
\n
\nClassically, classification algorithms are called nonlinear when the domains relating to the different classes are generally parts of the space of observations separated by nonlinear hyper-surfaces. However, the data to be classified can belong to non-vector spaces (non-linear and even non-convex). It is often useful to know their domains of variation. Often, they can belong to Riemannian type differential manifolds or to Lie groups. The geometry of, currently popular in the field of computer vision, is one of the most illustrative examples. Thus, we will try to show here on concrete examples in pattern recognition how this new generation of learning algorithms taking into account the nature of data carriers is revolutionizing the field of vision. On the other hand and from some examples in the field of machine learning, we will try to show how the optimization methods improve in the precision sense when considering that the data come from manifolds type space. Finally, the approach of invariants in Shape analysis, can be seen as that of extrinsic statistics while the one based on Kendall spaces counts in the category of intrinsic statistics. We compare them in the context of shape classification: Euclidean-planar contours, Affine-contours and finally Euclidean-curved surfaces.<\/p>\n","protected":false},"excerpt":{"rendered":"