Papers by Nhat-Quang Doan
International Journal of Engineering Research and, 2019
The architecture of service systems enables business functionalities to be invoked over a remote ... more The architecture of service systems enables business functionalities to be invoked over a remote network. This kind of systems is more and more developed nowadays. The modelling method for developing service-oriented solutions thus is required not only specifically but also precisely. Our work focuses on the method for modelling service systems having service-oriented architecture. Our approach is based on SoaML, a standard widespread notation introduced by OMG as a standard UML profile. In this paper, we propose a precise method for modelling the services and the service systems, named PreciseSoa, and illustrate the method using a case study.
Hierarchical Laplacian Score for unsupervised feature selection
In this paper, we address the problem of unsupervised feature selection. This is an important cha... more In this paper, we address the problem of unsupervised feature selection. This is an important challenge due to the absence of class labels that would guide the search for relevant information. Motivated by this challenge, we define the new method named Hierarchical Laplacian Score (HLS) that constrains the Laplacian Score using a tree topology structure. The purpose of using this structure is to automatically discover local data structure and local nearest neighbors for each data object. Experimental results on various datasets have demon- strated the effectiveness of the proposed algorithm in clustering and classification applications.
Hierarchical Laplacian Score for unsupervised feature selection
2018 International Joint Conference on Neural Networks (IJCNN), 2018
In this paper, we address the problem of unsupervised feature selection. This is an important cha... more In this paper, we address the problem of unsupervised feature selection. This is an important challenge due to the absence of class labels that would guide the search for relevant information. Motivated by this challenge, we define the new method named Hierarchical Laplacian Score (HLS) that constrains the Laplacian Score using a tree topology structure. The purpose of using this structure is to automatically discover local data structure and local nearest neighbors for each data object. Experimental results on various datasets have demon- strated the effectiveness of the proposed algorithm in clustering and classification applications.
Wireless Communications and Mobile Computing
Vehicular networks play a crucial role in Intelligent Transportation System (ITS), making transpo... more Vehicular networks play a crucial role in Intelligent Transportation System (ITS), making transportation safer and more convenient. Most applications in ITS require information carried by basic safety messages (BSMs) to be exchanged periodically among vehicles. However, BSMs are vulnerable to different attacks, especially jamming attacks, due to their limited short message length and life span. In this paper, we analyze the impact of a jamming attack on BSMs and initially propose a random channel surfing scheme to attempt to react to the attack. We investigate the scheme by a simple extendible probabilistic model and simulation in NS-3. Obtained results provide a reference to design an optimal channel surfing scheme that adapts to its supported applications.
Towards Real-Time Smile Detection Based on Faster Region Convolutional Neural Network
2018 1st International Conference on Multimedia Analysis and Pattern Recognition (MAPR), 2018
Real-time smile detection from digital images taken in unconstrained conditions of lighting and b... more Real-time smile detection from digital images taken in unconstrained conditions of lighting and background is still a challenge and can be applied in many real world applications, such as automatic capturing image on a mobile phone camera whenever smile is detected. Previous works usually consider this problem in two steps separately: face detection and smile detection. In this paper, we propose a new method to speed up computational performance of smile detection algorithm using a specialized architecture of Faster Region Convolutional Neural Network (Faster R-CNN). The evaluation from GENKI- 4K dataset shows that our network gains up to 50% faster inference performance and 2 times faster in training than the original Faster R-CNN with the accuracy of 84.5%, which is acceptable for predicting and classifying smile from given images.
Résumé. La sélection des variables a un rôle très important dans la fouille de données lorsqu’un ... more Résumé. La sélection des variables a un rôle très important dans la fouille de données lorsqu’un grand nombre de variables est disponible. Ainsi, certaines variables peuvent être peu significatives, corrélées ou non pertinentes. Une méthode de sélection a pour objectif de mesurer la pertinence d’un ensemble utilisant principalement un critère d’évaluation. Nous présentons dans cet article un critère non supervisé permettant de mesurer la pertinence d’un sous-ensemble de variables. Ce dernier repose sur l’utilisation du score Laplacien auquel nous avons ajouté des contraintes hiérarchiques. Travailler dans le cadre non supervisé est un vrai challenge dans ce domaine dû à l’absence des étiquettes de classes. Les résultats obtenus sur plusieurs bases de tests sont très encourageants et prometteurs.
Modèles hiérarchiques et topologiques pour le partitionnement et la visualisation des données
Cette these se concentre sur les approches hierarchiques et topologiques pour le clustering et la... more Cette these se concentre sur les approches hierarchiques et topologiques pour le clustering et la visualisation de donnees.Le probleme du clustering devient de plus en plus complique en raison de presence de donnees structureessous forme de graphes, arbres ou donnees sequentielles. Nous nous sommes particulierement interesses aux cartes auto-organisatrices et au modele hierarchique AntTree qui modelise la capacite des fourmis reelles. En combinant ces approches, l'objectif est de representer les donnees dans une structure hierarchique et topologique. Ce combine permet de visualiser les resultats de clustering sous forme de plusieurs arbres organises sur une carte topologique. Dans ce rapport, nous presentons trois modeles, dans le premier modele, nous montrons l'interet d’utiliser les structures hierarchiques et topologiques sur des ensembles de donnees structures sous formede graphes. Le second modele est une version incrementale qui n'impose pas de regles sur la preser...
2016 IEEE RIVF International Conference on Computing & Communication Technologies, Research, Innovation, and Vision for the Future (RIVF), 2016
Grounded in a self-reflexive, intersectional analysis of positionality, we examine emotions in fi... more Grounded in a self-reflexive, intersectional analysis of positionality, we examine emotions in fieldwork through the autobiographical accounts that we gathered during our postgraduate ethnographic research in the Global South. We show how we, two female early-career geographers, emotionally coped with instances that put us in a vulnerable position due to loneliness, commitment to the field, insistent questioning, violence, and violent threats. We argue that a culture of silence surrounding fieldwork difficulties and their emotional consequences tend to permeate our discipline. We contend that geography departments ought to provide mentorship that takes into account doctoral candidates' different positionalities, conflated vulnerability and privilege, and embodied intersectional axes. This renewed awareness will help not only to reveal possible risks and challenges connected with fieldwork but also ultimately to enrich the overall academic discussions within our discipline.
Dynamic indexing for Content-Based Image Retrieval systems using hierarchical and topological network
2016 Eighth International Conference on Knowledge and Systems Engineering (KSE), 2016
Over the past decade, Content-Based Image Retrieval (CBIR) has been an active research approach t... more Over the past decade, Content-Based Image Retrieval (CBIR) has been an active research approach to retrieve images based on their visual contents such as color, texture and shape. Image indexing is an important task to ensure the efficiency of CBIR systems. In this paper, we present a new stochastic clustering algorithm which allows automatic image indexing and quick neighbor search of similar images. Using a hierarchical and topological structure, this dynamic algorithm offers extensibility by handling new arrival images and training them without redoing the whole learning process again. The proposed structure provides advantages in improving the accuracy of retrieval results and reducing searching time with different kinds of search, i.e. searching in term of topological and hierarchical relationships. Dynamic indexing results are presented with the application of our proposed structure on a real-world image dataset.
2015 International Joint Conference on Neural Networks (IJCNN), 2015
Data stream clustering aims at studying large volumes of data that arrive continuously and the ob... more Data stream clustering aims at studying large volumes of data that arrive continuously and the objective is to build a good clustering of the stream, using a small amount of memory and time. Visualization is still a big challenge for large data streams. In this paper we present a new approach using a hierarchical and topological structure (or network) for both clustering and visualization. The topological network is represented by a graph in which each neuron represents a set of similar data points and neighbor neurons are connected by edges. The hierarchical component consists of multiple tree-like hierarchic of clusters which allow to describe the evolution of data stream, and then analyze explicitly their similarity. This adaptive structure can be exploited by descending top-down from the topological level to any hierarchical level. The performance of the proposed algorithm is evaluated on both synthetic and real-world datasets.
Self-Organizing Map and Tree Topology for Graph Summarization
Lecture Notes in Computer Science, 2012
In this paper, we present a novel approach called SOM-tree to summarize a given graph into a smal... more In this paper, we present a novel approach called SOM-tree to summarize a given graph into a smaller one by using a new decomposition of original graph. The proposed approach provides simultaneously a topological map and a tree topology of data using self-organizing maps. Unlike other clustering methods, the tree-structure aim to preserve the strengths of connections between graph vertices. The hierarchical nature of the summarization data structure is particularly attractive. Experiments evaluated by Accuracy and Normalized Mutual Information conducted on real data sets demonstrate the good performance of SOM-tree.
Self-organizing trees for visualizing protein dataset
The 2013 International Joint Conference on Neural Networks (IJCNN), 2013
ABSTRACT Clustering and visualizing multidimensional or structured data are important tasks for d... more ABSTRACT Clustering and visualizing multidimensional or structured data are important tasks for data analysis, especially in bioinformatics. Self-organizing models are often used to address both of these issues. In this paper we introduce a hierarchical and topological visualization technique called Self-organizing Trees (SoT) which is able to represent data in hierarchical and topological structure. The experiment is conducted on a real-world protein data set.
Graph Decomposition Using Self-organizing Trees
2012 16th International Conference on Information Visualisation, 2012
ABSTRACT In this paper, we present a new approach for graph decomposition using topological and h... more ABSTRACT In this paper, we present a new approach for graph decomposition using topological and hierarchical partitioning of data. Our method called GD-SOM-Tree (Graph Decomposition using Self-Organizing Trees) is based on self-organizing models. The benefit of this novel approach is to represent and visualize hierarchical relations which replace the original graph with a summary and gives a good understanding of the underlying problem.
Growing Self-organizing Trees for knowledge discovery from data
The 2012 International Joint Conference on Neural Networks (IJCNN), 2012
ABSTRACT In this paper, we propose a new unsupervised learning method based on growing neural gas... more ABSTRACT In this paper, we propose a new unsupervised learning method based on growing neural gas and using self-assembly rules to build hierarchical structures. Our method named GSoT (Growing Self-organizing Trees) depicts data in topological and hierarchical organization. This makes GSoT a good tool for data clustering and knowledge discovery. Experiments conducted on real data sets demonstrate the good performance of GSoT.
Neural Networks, 2013
This paper presents a new unsupervised learning method based on growing processes and autonomous ... more This paper presents a new unsupervised learning method based on growing processes and autonomous self-assembly rules. This method, called Growing Self-organizing Trees (GSoT), can grow both network size and tree topology to represent the topological and hierarchical dataset organization, allowing a rapid and interactive visualization. Tree construction rules draw inspiration from elusive properties of biological organization to build hierarchical structures. Experiments conducted on real datasets demonstrate good GSoT performance and provide visual results that are generated during the training process.
Uploads
Papers by Nhat-Quang Doan