Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
The number of satellites and sensors devoted to Earth observation has become increasingly elevated, delivering extensive data, especially images. At the same time, the access to such data and the tools needed to process them has considerably improved. In the presence of such data flow, we need automatic image interpretation methods, especially when it comes to the monitoring and prediction of environmental and societal changes in highly dynamic socio-environmental contexts. This could be accomplished via artificial intelligence. The concept described here relies on the induction of classification rules that explicitly take into account structural knowledge, using Aleph, an Inductive Logic Programming (ILP) system, combined with a multi-class classification procedure. This methodology was used to monitor changes in land cover/use of the French Guiana coastline. One hundred and fifty-eight classification rules were induced from 3 diachronic land cover/use maps including 38 classes. These rules were expressed in first order logic language, which makes them easily understandable by non-experts. A 10-fold cross-validation gave significant average values of 84.62%, 99.57% and 77.22% for classification accuracy, specificity and sensitivity, respectively. Our methodology could be beneficial to automatically classify new objects and to facilitate object-based classification procedures.
ISSN: 0098-3004

hal-01369822
https://hal.archives-ouvertes.fr/hal-01369822
https://hal.archives-ouvertes.fr/hal-01369822/document
https://hal.archives-ouvertes.fr/hal-01369822/file/CAGEO-D-13-00538R1_article.pdf
DOI : 10.1016/j.cageo.2014.08.013

Éditeur(s) : HAL CCSD
Résumé : International audience
This paper describes a blood alcohol content estimation method for car driver, based on a comportment analysis performed within a realistic simulation. An artificial neural network learns how to estimate subject's blood alcohol content. Low-level recording of user actions on the steering wheel and pedals are used to feed a multilayer perceptron, and a breathalyzer is used to build the learning examples set (desired output). Results are compared with a successful previous work based on a simple video game and demonstrate the ''complexity scalability'' of the approach.
The 20 th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning - Proceedings - Bruges, Belgium from 25 to 27 April 2012 .
Bruges, Belgium

hal-00850136
https://hal.archives-ouvertes.fr/hal-00850136

Éditeur(s) : HAL CCSD
Résumé : International audience
“Affective Computing is computing that relates to, arises from, or deliberately influences emotions” (R.W. Picard, MIT Press, 1997). Specifically, psychosocial equilibrium is now considered as a major issue for both individual and public health. Predicting human emotions via non-intrusive methods is a great challenge triggered by the rise of intelligent assisting systems. In order to determine the emotional state of a subject having a potential psychosocial disorder, it is required to process both neutral and expressive emotions. A computer-engineered method aiming to provide emotional information from videos is proposed. An original approach for affect classification and its interpretation is presented.The database is part of the HUMAINE database, and consists of a large volume of video clips, each displaying individuals, depressed or not, talking with a conversational agent. Both the high-dimensionality of data and its temporal aspect require a pre-training that consists of cleaning, sorting and feature selection, in order to make relevant information usable for further processing by MachineLearning methods. Next step consists in classifying the affective state of subjects according to their positivity or negativity (Russell's model, 1980). By means of six independent classifiers, each running a Support Vector Machine algorithm, separate performances are obtained for each class of emotions. The results reveal an interesting “neutral area phenomenon”, where emotions tend to be harder to be detected.Though more frequent, neutral emotions seem to carry less discriminant information than expressive ones. This issue will be a key for designing a tuning step in further research.
Caribbean Academy Of Science
Deshaies, France

hal-01408029
https://hal.univ-antilles.fr/hal-01408029

Éditeur(s) : HAL CCSD
Résumé : International audience
We instrumented a realistic car simulator to extract low level data related to the driver's use of the vehicle controls. After proceeding these data, we generated features that were fed to a Multi-Layer Perceptron (MLP) and Support Vector Machines (SVM). Our goal was determine if the driver's Blood Alcohol Content (BAC) was over $0.4g.l^{-1}$ or not, and even estimate the BAC value. Our device process the vehicle's controls data and then outputs the user BAC. We discuss the results of the prototype using the MLP and SVM algorithms in both single-user and multi-user context for detection of drunk drivers and estimation of the BAC value. The prototype performed better with single user base than with multi-user, and provided comparable results with MLP and SVM.
ESANN 2013 proccedings,European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning
Bruges, Belgium

hal-00850138
https://hal.archives-ouvertes.fr/hal-00850138

Éditeur(s) : HAL CCSD
Résumé : International audience
This paper focuses on the study of the influence of a newly implemented mechanism on a Pittsburgh-like classifier system. The Adapted Pittsburgh Classifier System is a learning classifier system that uses genetic algorithms to evolve its ruleset. The new mechanism discussed is inspired from Wilson work on the eXtended Classifier System (XCS): it allows the concerned LCS to adapt its rule set when facing a new signal by modifying an existing rule. This mechanism is called covering mechanism due to the fact that the rule is covered up by a new rule which is sensitive to this new signal. Effects of this covering mechanism are first measured on the performances of APCS on well known multi-step environments: maze-type environments (Woods 101 and E2). In addition, further measures presented in this paper indicate that we could possibly rely on the behavior of this covering mechanism to automate the research of a correct cognitive capacity needed by the APCS to solve a given multistep problem.
2008 Eighth International Conference on Hybrid Intelligent Systems
Barcelone, Spain

hal-00520608
https://hal.archives-ouvertes.fr/hal-00520608
DOI : 10.1109/HIS.2008.105

Éditeur(s) : HAL CCSD
Résumé : The concept of network provides a model for representing a wide variety of objects and systems, both natural and social, in which a set of homogeneous or heterogeneous entities interact. It is now widely used to describe various kinds of relational structures. However, if everyone has an idea of the concept of network, we often ignore the implications that these structures have in real world phenomena. This is for example the case of processes such as the spread of a rumor, the disease transmission, or even the emergence of subjects of common interest for a group of individuals, in which the relations maintained between individuals, and their nature, often prove to be the main factors determining the evolution of the phenomenon. This is the reason why the study of networks has become one of the emerging areas in the 21st century called the "Science of networks." In this thesis, we address three issues of the domain of the science of networks: the problem of diffusion in social networks, where we have addressed more particularly the impact of the network dynamics on the diffusion process, the problem of the analysis of social networks, in which we have proposed a solution to take full advantage of all information available on the network by combining information on both structure and node attributes and the central problem of the social data collection, for which we have focused on the particular case of the data collection in a wild environment.
Le concept de réseau offre un modèle de représentation pour une grande variété d'objets et de systèmes, aussi bien naturels que sociaux, dans lesquels un ensemble d'entités homogènes ou hétérogènes interagissent entre elles. Il est aujourd'hui employé couramment pour désigner divers types de structures relationnelles. Pourtant, si chacun a une idée plus ou moins précise de ce qu'est un réseau, nous ignorons encore souvent les implications qu'ont ces structures dans de nombreux phénomènes du monde qui nous entoure. C'est par exemple le cas de processus tels que la diffusion d'une rumeur, la transmission d'une maladie, ou même l'émergence de sujets d'intérêt commun à un groupe d'individus, dans lesquels les relations que maintiennent les individus entre eux et leur nature s'avèrent souvent être les principaux facteurs déterminants l'évolution du phénomène. C'est ainsi que l'étude des réseaux est devenue l'un des domaines émergents du 21e siècle appelé la "Science des réseaux". Dans ce mémoire, nous abordons trois problèmes de la science des réseaux: le problème de la diffusion dans les réseaux sociaux, où nous nous sommes intéressés plus particulièrement à l'impact de la dynamique du réseau sur le processus de diffusion, le problème de l'analyse des réseaux sociaux, dans lequel nous avons proposé une solution pour tirer parti de l'ensemble des informations disponibles en combinant les informations sur la structure du réseau et les attributs des noeuds et le problème central de la collecte de données sociales, où nous nous sommes intéressés au cas particulier de la collecte de données en milieux sauvages.
https://tel.archives-ouvertes.fr/tel-00830882

Droits : info:eu-repo/semantics/OpenAccess
tel-00830882
https://tel.archives-ouvertes.fr/tel-00830882
https://tel.archives-ouvertes.fr/tel-00830882/document
https://tel.archives-ouvertes.fr/tel-00830882/file/These_STATTNER_2012.pdf

Éditeur(s) : HAL CCSD
Résumé : International audience
La recherche de liens conceptuels fréquents (FCL) est une nouvelle approche de clustering de réseaux, qui exploite à la fois la structure et les attributs des noeuds. Bien que les travaux récents se soient déjà intéressés à l’optimisation des algorithmes de recherche des FCL, peu de travaux sont aujourd’hui menés sur la complémentarité qui existe entre les liens conceptuels et l’approche classique de clustering qui consiste en l’extraction de communautés. Ainsi dans ce papier, nous nous intéressons à ces deux approches. Notre objectif est d’évaluer les relations potentiellement existantes entre les communautés et les FCL pour comprendre la façon dont les motifs obtenus par chacune des méthodes peuvent correspondre ou s’intersecter ainsi que la connaissance utile résultant de la prise en compte de ces deux types de connaissance. Nous proposons pour cela un ensemble de mesures originales, basées sur la notion d’homogénéité, visant à évaluer le niveau d’intersection des FCL et des communautés lorsqu’ils sont extraits d’un même jeu de données. Notre approche est appliquée à deux réseaux et démontre l’importance de considérer simultanément plusieurs types de connaissance et leur intersection.
Extraction et gestion des connaissances
Rennes, France

hal-01312560
https://hal.archives-ouvertes.fr/hal-01312560

Éditeur(s) : HAL CCSD
Résumé : International audience
https://hal.archives-ouvertes.fr/hal-01436542

hal-01436542
https://hal.archives-ouvertes.fr/hal-01436542
DOI : 10.1109/IGARSS.2011.6049595

Résumé : Les méthodes classiques d'analyse d'images satellites sont inadaptées au volume actuel du flux de données. L'automatisation de l'interprétation de ces images devient donc cruciale pour l'analyse et la gestion des phénomènes observables par satellite et évoluant dans le temps et l'espace. Ce travail vise à automatiser la cartographie dynamique de l'occupation du sol à partir d'images satellites, par des mécanismes expressifs, facilement interprétables en prenant en compte les aspects structurels de l'information géographique. Il s'inscrit dans le cadre de l'analyse d'images basée objet. Ainsi, un paramétrage supervisé d'un algorithme de segmentation d'images est proposé. Dans un deuxième temps, une méthode de classification supervisée d'objets géographiques est présentée combinant apprentissage automatique par programmation logique inductive et classement par l'approche multi-class rule set intersection. Ces approches sont appliquées à la cartographie de la bande côtière Guyanaise. Les résultats démontrent la faisabilité du paramétrage de la segmentation, mais également sa variabilité en fonction des classes de la carte de référence et des données d'entrée. Les résultats de la classification supervisée montrent qu'il est possible d'induire des règles de classification expressives, véhiculant des informations cohérentes et structurelles dans un contexte applicatif donnée et conduisant à des valeurs satisfaisantes de précision et de KAPPA (respectivement 84,6% et 0,7). Ce travail de thèse contribue ainsi à l'automatisation de la cartographie dynamique à partir d'images de télédétection et propose des perspectives originales et prometteuses.
Classical methods for satellite image analysis are inadequate for the current bulky data flow. Thus, automate the interpretation of such images becomes crucial for the analysis and management of phenomena changing in time and space, observable by satellite. Thus, this work aims at automating land cover cartography from satellite images, by expressive and easily interpretable mechanism, and by explicitly taking into account structural aspects of geographic information. It is part of the object-based image analysis framework, and assumes that it is possible to extract useful contextual knowledge from maps. Thus, a supervised parameterization methods of a segmentation algorithm is proposed. Secondly, a supervised classification of geographical objects is presented. It combines machine learning by inductive logic programming and the multi-class rule set intersection approach. These approaches are applied to the French Guiana coastline cartography. The results demonstrate the feasibility of the segmentation parameterization, but also its variability as a function of the reference map classes and of the input data. Yet, methodological developments allow to consider an operational implementation of such an approach. The results of the object supervised classification show that it is possible to induce expressive classification rules that convey consistent and structural information in a given application context and lead to reliable predictions, with overall accuracy and Kappa values equal to, respectively, 84,6% and 0,7. In conclusion, this work contributes to the automation of the dynamic cartography from remotely sensed images and proposes original and promising perpectives

http://www.theses.fr/2013AGUY0671/document

Éditeur(s) : HAL CCSD
Résumé : International audience
This paper demonstrates that a human being using an interface can be efficiently evaluated - in real time - by embedding basic measurements in the interface and using a suitable trained artificial neural network. The approach is introduced through video games but is suitable for any machine capable of valuable measurements on user actions. Of course, the quality of the "diagnostic" depends of the learnability of the task and of the size and quality of the learning base. Typical applications include the detection of fatigue, stress, emotions, the influence of a drug or of medical treatments ; screening a deficit or adequateness to a task, etc. Two successful prototypes are presented, one to predict the mental age of children through a set of simple basic games, and the other to detect if a subject is right-handed of left-handed through a racing car simulation.
Proceeding of the Third International Conferences on Advances in Computer-Human Interactions
St. Maarten, Netherlands Antilles

hal-00601476
https://hal.archives-ouvertes.fr/hal-00601476

Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
These last years, many analysis methods have been proposed to extract knowledge from social networks. As for the traditional data mining domain, these network-based approaches can be classified according to two main families. The approaches based on predictive modelling, which encompass the techniques that analyse current and historical facts to make predictive assumptions about future or unknown events. The approaches based on descriptive modelling, which cover the set of techniques that aim to summarize the data by identifying some relevant features in order to describe how things organize and actually work. In this paper, we review the main descriptive modelling methods of social networks and show for each of them the resulting useful knowledge on a running example. We particularly emphasize on the most recent methods that combine information available on both the network structure and the node attributes in order to provide original description models taking into account the context.
EISSN: 1877-0509

hal-01312383
https://hal.univ-antilles.fr/hal-01312383
DOI : 10.1016/j.procs.2015.05.505

Éditeur(s) : HAL CCSD Springer International Publishing AG.
Résumé : International audience
Underwater Seascapes. From geographical to ecological perspectives

hal-01429034
https://hal.archives-ouvertes.fr/hal-01429034
DOI : 10.1007/978-3-319-03440-9_16

Éditeur(s) : HAL CCSD
Résumé : International audience
We instrumented a car simulator by gathering low level data and fed it to an artificial neural network in order to perform blood alcohol content (BAC) estimations. The results depend on the quality of the data extraction and processing, and also on the selected inputs. We explain our data extraction and processing methodology, and how we used it to generate reliable and comparable features. At last, we describe the performances of individual features and how they combine. In the end, the prototype was able to accurately estimate the BAC value of a subject after being trained with driving samples of this subject with various BAC values.
Proceedings of ACHI 2013, The Sixth International Conference on Advances in Computer-Human Interactions
Nice, France

hal-00850137
https://hal.archives-ouvertes.fr/hal-00850137

Éditeur(s) : HAL CCSD Ecological Society of America
Résumé : Reliable assessment of fish origin is of critical importance for exploited species, since nursery areas must be identified and protected to maintain recruitment to the adult stock. During the last two decades, otolith chemical signatures (or "fingerprints") have been increasingly used as tools to discriminate between coastal habitats. However, correct assessment of fish origin from otolith fingerprints depends on various environmental and methodological parameters, including the choice of the statistical method used to assign fish to unknown origin. Among the available methods of classification, Linear Discriminant Analysis (LDA) is the most frequently used, although it assumes data are multivariate normal with homogeneous within-group dispersions, conditions that are not always met by otolith chemical data, even after transformation. Other less constrained classification methods are available, but there is a current lack of comparative analysis in applications to otolith microchemistry. Here, we assessed stock identification accuracy for four classification methods (LDA, Quadratic Discriminant Analysis [QDA], Random Forests [RF], and Artificial Neural Networks [ANN]), through the use of three distinct data sets. In each case, all possible combinations of chemical elements were examined to identify the elements to be used for optimal accuracy in fish assignment to their actual origin. Our study shows that accuracy differs according to the model and the number of elements considered. Best combinations did not include all the elements measured, and it was not possible to define an ad hoc multielement combination for accurate site discrimination. Among all the models tested, RF and ANN performed best, especially for complex data sets (e. g., with numerous fish species and/or chemical elements involved). However, for these data, RF was less time-consuming and more interpretable than ANN, and far more efficient and less demanding in terms of assumptions than LDA or QDA. Therefore, when LDA and QDA assumptions cannot be reached, the use of machine learning methods, such as RF, should be preferred for stock assessment and nursery identification based on otolith microchemistry, especially when data set include multispecific otolith signatures and/or many chemical elements.
ISSN: 1051-0761

hal-00644549
https://hal.archives-ouvertes.fr/hal-00644549
DOI : 10.1890/09-1887.1

Éditeur(s) : HAL CCSD
Résumé : International audience
We present a novel approach towards signature recognition from spatio-temporal data. The data is obtained by recording gyroscope and accelerometer measurements from an embedded pen device. The idea of Universum learning was previously presented by Vapnik and recently popularized in machine learning community. It assumes that the decision boundary of a classifier lies close to data with high uncertainty. The quality of the final classifier strongly depends on a way how to choose the Universum data and also on the representation of original data. In our paper we use a novel approach of Universum learning to classify signature data, also we present our novel idea how to sample the Universum data. At last, we also find more effective representation of the signature data itself compared to the baseline method. These three novelties allow us to outperform previously published results by 4.89% / 5.58%.
17th Biennial Conference of the International Graphonomics Society
Pointe-à-Pitre, Guadeloupe

hal-01165925
https://hal.univ-antilles.fr/hal-01165925
https://hal.univ-antilles.fr/hal-01165925/document
https://hal.univ-antilles.fr/hal-01165925/file/IGS_2015_submission_38.pdf

Éditeur(s) : HAL CCSD Coastal Education and Research Foundation
Résumé : Special Issue 65 - International Coastal Symposium Volume 1
International audience
ISSN: 0749-0208

hal-01429044
https://hal.archives-ouvertes.fr/hal-01429044
DOI : 10.2112/SI65-117.1

Éditeur(s) : HAL CCSD
Résumé : The Voronoi diagram of a finite set of objects is a fundamental geometric structure that subdivides the embedding space into regions, each region consisting of the points that are closer to a given object than to the others. We may define many variants of Voronoi diagrams depending on the class of objects, the distance functions and the embedding space. In this paper, we investigate a framework for defining and building Voronoi diagrams for a broad class of distance functions called Bregman divergences. Bregman divergences include not only the traditional (squared) Euclidean distance but also various divergence measures based on entropic functions. Accordingly, Bregman Voronoi diagrams allow to define information-theoretic Voronoi diagrams in statistical parametric spaces based on the relative entropy of distributions. We define several types of Bregman diagrams, establish correspondences between those diagrams (using the Legendre transformation), and show how to compute them efficiently. We also introduce extensions of these diagrams, e.g. k-order and k-bag Bregman Voronoi diagrams, and introduce Bregman triangulations of a set of points and their connexion with Bregman Voronoi diagrams. We show that these triangulations capture many of the properties of the celebrated Delaunay triangulation. Finally, we give some applications of Bregman Voronoi diagrams which are of interest in the context of computational geometry and machine learning.
https://hal.inria.fr/inria-00137865

Report N° : RR-6154
inria-00137865
https://hal.inria.fr/inria-00137865
https://hal.inria.fr/inria-00137865v2/document
https://hal.inria.fr/inria-00137865/file/bregman.pdf

Éditeur(s) : HAL CCSD
Résumé : International audience
International Coral Reef Symposium 2012

hal-01429056
https://hal.archives-ouvertes.fr/hal-01429056

Éditeur(s) : HAL CCSD Public Library of Science
Résumé : International audience
ISSN: 1932-6203

hal-01520195
https://hal.archives-ouvertes.fr/hal-01520195
DOI : 10.1371/journal.pone.0176862

Éditeur(s) : HAL CCSD
Résumé : International audience
Dans ce travail, nous analysons les données concernant les articles publiés à la conférence EGC. Notre objectif est d’identifier et de comprendre les tendances en matière de collaborations. Pour ce faire, nous adoptons une modélisation descriptive, à travers une approche réseau qui consiste à générer tout d’abord le réseau de collaborations des auteurs à partir des données. Nous enrichissons ensuite les noeuds de ce réseau d’une dizaine d’attributs individuels extraits à partir des données. Enfin, nous recherchons des vues conceptuelles, une approche récente de clustering de liens, qui permet de synthétiser des réseaux en mettant en évidence les ensembles d’attributs retrouvés fréquemment liés dans le réseau. Les résultats obtenus montrent les tendances existantes dans les comportements de collaborations. Dans ce papier, nous présentons ces tendances et montrons comment elles évoluent selon différents seuils d’extraction.
Extraction et gestion des connaissances
Reims, France

hal-01312559
https://hal.archives-ouvertes.fr/hal-01312559