Descripteur
Documents disponibles dans cette catégorie (450)
Ajouter le résultat dans votre panier
Visionner les documents numériques
Affiner la recherche Interroger des sources externes
Etendre la recherche sur niveau(x) vers le bas
Towards expressive graph neural networks : Theory, algorithms, and applications / Georgios Dasoulas (2022)
Titre : Towards expressive graph neural networks : Theory, algorithms, and applications Type de document : Thèse/HDR Auteurs : Georgios Dasoulas, Auteur ; Michalis Vazirgiannis, Directeur de thèse ; Aladin Virmaux, Directeur de thèse Editeur : Paris : Institut Polytechnique de Paris Année de publication : 2022 Note générale : bibliographie
These de doctorat de l’Institut Polytechnique de Paris préparée à l’Ecole Polytechnique, spécialité InformatiqueLangues : Anglais (eng) Descripteur : [Vedettes matières IGN] Intelligence artificielle
[Termes IGN] apprentissage automatique
[Termes IGN] attention (apprentissage automatique)
[Termes IGN] entropie
[Termes IGN] graphe
[Termes IGN] isomorphisme
[Termes IGN] noeud
[Termes IGN] réseau neuronal de graphes
[Termes IGN] théorie des graphesIndex. décimale : THESE Thèses et HDR Résumé : (auteur) As the technological evolution of machine learning is accelerating nowadays, data plays a vital role in building intelligent models, being able to simulate phenomena, predict values and make decisions. In an increasing number of applications, data take the form of networks. The inherent graph structure of network data motivated the evolution of the graph representation learning field. Its scope includes generating meaningful representations for graphs and their components, i.e., the nodes and the edges. The research on graph representation learning was accelerated with the success of message passing frameworks applied on graphs, namely the Graph Neural Networks. Learning informative and expressive representations on graphs plays a critical role in a wide range of real-world applications, from telecommunication and social networks, urban design, chemistry, and biology. In this thesis, we study various aspects from which Graph Neural Networks can be more expressive, and we propose novel approaches to improve their performance in standard graph learning tasks. The main branches of the present thesis include: the universality of graph representations, the increase of the receptive field of graph neural networks, the design of stable deeper graph learning models, and alternatives to the standard message-passing framework. Performing both theoretical and experimental studies, we show how the proposed approaches can become valuable and efficient tools for designing more powerful graph learning models.In the first part of the thesis, we study the quality of graph representations as a function of their discrimination power, i.e., how easily we can differentiate graphs that are not isomorphic. Firstly, we show that standard message-passing schemes are not universal due to the inability of simple aggregators to separate nodes with ambiguities (similar attribute vectors and neighborhood structures). Based on the found limitations, we propose a simple coloring scheme that can provide universal representations with theoretical guarantees and experimental validations of the performance superiority. Secondly, moving beyond the standard message-passing paradigm, we propose an approach for treating a corpus of graphs as a whole instead of examining graph pairs. To do so, we learn a soft permutation matrix for each graph, and we project all graphs in a common vector space, achieving a solid performance on graph classification tasks.In the second part of the thesis, our primary focus is concentrated around the receptive field of the graph neural networks, i.e., how much information a node has in order to update its representation. To begin with, we study the spectral properties of operators that encode adjacency information. We propose a novel parametric family of operators that can adapt throughout training and provide a flexible framework for data-dependent neighborhood representations. We show that the incorporation of this approach has a substantial impact on both node classification and graph classification tasks. Next, we study how considering the k-hop neighborhood information for a node representation can output more powerful graph neural network models. The resulted models are proven capable of identifying structural properties, such as connectivity and triangle-freeness.In the third part of the thesis, we address the problem of long-range interactions, where nodes that lie in distant parts of the graph can affect each other. In this problem, we either need the design of deeper models or the reformulation of how proximity is defined in the graph. Firstly, we study the design of deeper attention models, focusing on graph attention. We calibrate the gradient flow of the model by introducing a novel normalization that enforces Lipschitz continuity. Next, we propose a data augmentation method for enriching the node attributes with information that encloses structural information based on local entropy measures. Note de contenu : 1. Introduction
2. Preliminaries
I- Discrimination power
3. Universal approximation on graphs
4. Learning soft permutations for graph representations
II- Receptive field
5. Learning graph shift operators
6. Increasing the receptive field with multiple hops
III- Beyond local interactions
7. Lipschitz continuity of graph attention
8. Structural symmetries in graphs
9. Conclusions and outlookNuméro de notice : 24076 Affiliation des auteurs : non IGN Thématique : INFORMATIQUE/MATHEMATIQUE Nature : Thèse française Note de thèse : Thèse de doctorat : Informatique : Palaiseau : 2022 DOI : sans En ligne : https://theses.hal.science/tel-03666690 Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=102200 A topology-based graph data model for indoor spatial-social networking / Mahdi Rahimi in International journal of geographical information science IJGIS, vol 35 n° 12 (December 2021)
[article]
Titre : A topology-based graph data model for indoor spatial-social networking Type de document : Article/Communication Auteurs : Mahdi Rahimi, Auteur ; Mohammad Reza Malek, Auteur ; Christophe Claramunt, Auteur ; Thierry Le Pors, Auteur Année de publication : 2021 Article en page(s) : pp 2517 - 2539 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Analyse spatiale
[Termes IGN] algorithme du simplexe
[Termes IGN] espace intérieur
[Termes IGN] graphe
[Termes IGN] modèle topologique de données
[Termes IGN] modélisation spatiale
[Termes IGN] représentation géométrique
[Termes IGN] représentation graphique
[Termes IGN] réseau social géodépendantRésumé : (auteur) This paper introduces a simplex-based enriched graph data model integrating a discrete and place-based indoor spatial model with a spatial-social network. The proposed model incorporates similarity and relevance measures, exhibited from Q-analysis of simplicial complexes, facilitating data manipulation and revealing latent relations in a spatial-social network. It also uses an indoor-specific metric representing the ease of access to process spatial-social queries in indoor environments. The proposed model’s experimental implementation shows the quantitative advantage of using graph-based representation and the qualitative superiority of simplex-based enrichment in processing spatial-social queries in indoor environments. Numéro de notice : A2021-875 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1080/13658816.2021.1912349 Date de publication en ligne : 14/04/2021 En ligne : https://doi.org/10.1080/13658816.2021.1912349 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=99138
in International journal of geographical information science IJGIS > vol 35 n° 12 (December 2021) . - pp 2517 - 2539[article]Binary space partitioning visibility tree for polygonal and environment light rendering / Hiroki Okuno in The Visual Computer, vol 37 n° 9 - 11 (September 2021)
[article]
Titre : Binary space partitioning visibility tree for polygonal and environment light rendering Type de document : Article/Communication Auteurs : Hiroki Okuno, Auteur ; Kei Iwasaki, Auteur Année de publication : 2021 Article en page(s) : pp 2499 - 2511 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique
[Termes IGN] arbre BSP
[Termes IGN] distribution du coefficient de réflexion bidirectionnelle BRDF
[Termes IGN] éclairage
[Termes IGN] éclairement lumineux
[Termes IGN] équation intégrale
[Termes IGN] intensité lumineuse
[Termes IGN] ombre
[Termes IGN] polygone
[Termes IGN] réflectance
[Termes IGN] visibilité (optique)Résumé : (auteur) In this paper, we present a geometric approach to render shadows for physically based materials under polygonal light sources. Direct illumination calculation from a polygonal light source involves the triple product integral of the lighting, the bidirectional reflectance distribution function (BRDF), and the visibility function over the polygonal domain, which is computation intensive. To achieve real-time performance, work on polygonal light shading exploits analytical solutions of boundary integrals along the edges of the polygonal light at the cost of lacking shadowing effects. We introduce a hierarchical representation for the precomputed visibility function to retain the merits of closed-form solutions for boundary integrals. Our method subdivides the polygonal light into a set of polygons visible from a point to be shaded. Experimental results show that our method can render complex shadows with a GGX microfacet BRDF from polygonal light sources at interactive frame rates. In addition, our visibility representation can be easily incorporated into environment lighting. Numéro de notice : A2021-644 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article DOI : 10.1007/s00371-021-02181-8 Date de publication en ligne : 14/06/2021 En ligne : https://doi.org/10.1007/s00371-021-02181-8 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98345
in The Visual Computer > vol 37 n° 9 - 11 (September 2021) . - pp 2499 - 2511[article]Constrained shortest path problems in bi-colored graphs: a label-setting approach / Amin AliAbdi in Geoinformatica, vol 25 n° 3 (July 2021)
[article]
Titre : Constrained shortest path problems in bi-colored graphs: a label-setting approach Type de document : Article/Communication Auteurs : Amin AliAbdi, Auteur ; Ali Mohades, Auteur ; Mansoor Davoodi, Auteur Année de publication : 2021 Article en page(s) : pp 513 - 531 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Géomatique
[Termes IGN] calcul d'itinéraire
[Termes IGN] chemin le plus court, algorithme du
[Termes IGN] données d'entrainement sans étiquette
[Termes IGN] graphe
[Termes IGN] programmation par contraintesRésumé : (auteur) Definition of an optimal path in the real-world routing problems is not necessarily the shortest one, because parameters such as travel time, safety, quality, and smoothness also played essential roles in the definition of optimality. In this paper, we use bi-colored graphs for modeling urban and heterogeneous environments and introduce variations of constraint routing problems. Bi-colored graphs are a kind of directed graphs whose vertices are divided into two subsets of white and gray. We consider two criteria, minimizing the length and minimizing the number of gray vertices and present two problems called gray vertices bounded shortest path problem and length bounded shortest path problem on bi-colored graphs. We propose an efficient time label-setting algorithm to solve these problems. Likewise, we simulate the algorithm and compare it with the related path planning methods on random graphs as well as real-world environments. The simulation results show the efficiency of the proposed algorithm. Numéro de notice : A2021-974 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE Nature : Article DOI : 10.1007/s10707-019-00385-8 Date de publication en ligne : 03/12/2019 En ligne : https://doi.org/10.1007/s10707-019-00385-8 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=100393
in Geoinformatica > vol 25 n° 3 (July 2021) . - pp 513 - 531[article]A Bayesian displacement field approach to accurate registration of SAR images / Mingtao Ding in Geocarto international, vol 36 n° 9 ([15/05/2021])
[article]
Titre : A Bayesian displacement field approach to accurate registration of SAR images Type de document : Article/Communication Auteurs : Mingtao Ding, Auteur ; Hongyan Wang, Auteur ; Lichun Sui, Auteur ; et al., Auteur Année de publication : 2021 Article en page(s) : pp 1007 - 1026 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image radar et applications
[Termes IGN] arc
[Termes IGN] enregistrement de données
[Termes IGN] estimation bayesienne
[Termes IGN] image radar moirée
[Termes IGN] implémentation (informatique)
[Termes IGN] inférence
[Termes IGN] méthode de Monte-Carlo par chaînes de Markov
[Termes IGN] méthode du maximum de vraisemblance (estimation)
[Termes IGN] processeur graphique
[Termes IGN] superposition d'images
[Termes IGN] transformationRésumé : (auteur) Precise registration of synthetic aperture radar (SAR) images is a nontrivial task since a change in radar-acquisition geometry generates image shifts. In existing system, either the transformation functions are oversimplified, or external measures such as digital elevation model and flight track are required to be precise. In this paper, we proposed a generative Bayesian approach to modelling the displacement vectors that map the position of each pixel in the image, thus avoiding degradation of the transformation function. Rather than providing a point estimate for the transformation function, the proposed method yields a full posterior density function of the transformation function. Especially, the Bayesian model learns all the parameters adaptively, and the procedure is fully automatic. The proposed model is comparable in accuracy to state-of-the-art optical flow methods on the challenging Sintel benchmarks, and outperforms currently published SAR image registration methods on some real SAR data with critical scenes. Numéro de notice : A2021-343 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1080/10106049.2019.1633418 Date de publication en ligne : 07/07/2019 En ligne : https://doi.org/10.1080/10106049.2019.1633418 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=97584
in Geocarto international > vol 36 n° 9 [15/05/2021] . - pp 1007 - 1026[article]Exemplaires(1)
Code-barres Cote Support Localisation Section Disponibilité 059-2021091 RAB Revue Centre de documentation En réserve L003 Disponible A new small area estimation algorithm to balance between statistical precision and scale / Cédric Vega in International journal of applied Earth observation and geoinformation, vol 97 (May 2021)PermalinkA BiLSTM-CNN model for predicting users’ next locations based on geotagged social media / Yi Bao in International journal of geographical information science IJGIS, vol 35 n° 4 (April 2021)PermalinkGraph convolutional networks by architecture search for PolSAR image classification / Hongying Liu in Remote sensing, vol 13 n° 7 (April-1 2021)PermalinkGraph convolutional autoencoder model for the shape coding and cognition of buildings in maps / Xiongfeng Yan in International journal of geographical information science IJGIS, vol 35 n° 3 (March 2021)PermalinkAn anchor-based graph method for detecting and classifying indoor objects from cluttered 3D point clouds / Fei Su in ISPRS Journal of photogrammetry and remote sensing, vol 172 (February 2021)PermalinkA spatiotemporal structural graph for characterizing land cover changes / Bin Wu in International journal of geographical information science IJGIS, vol 35 n° 2 (February 2021)Permalink3D urban scene understanding by analysis of LiDAR, color and hyperspectral data / David Duque-Arias (2021)PermalinkDétection et reconstruction 3D d’arbres urbains par segmentation de nuages de points : apport de l’apprentissage profond / Victor Alteirac (2021)PermalinkFinding the most navigable path in road networks / Ramneek Kaur in Geoinformatica, vol 25 n° 1 (January 2021)PermalinkHyperspectral and multispectral image fusion via graph Laplacian-guided coupled tensor decomposition / Yuanyang Bu in IEEE Transactions on geoscience and remote sensing, vol 59 n° 1 (January 2021)Permalink