Advanced Kernel Methods For Remote Sensing Image Classification

Résumé Suite aux recentes avancées technologiques, les archives d'images digitales ont connu une croissance qualitative et quantitative sans précédent. Malgré les énormes possibilités qu'elles offrent, ces avancées posent de nouvelles questions quant au traitement des masses de données saisies. Cette question est à la base de cette Thèse: les problèmes de traitement d'information digitale à très haute résolution spatiale et/ou spectrale y sont considérés en recourant à des approches d'apprentissage statistique, les méthodes à noyau. Cette Thèse étudie des problèmes de classification d'images, c'est à dire de catégorisation de pixels en un nombre réduit de classes refletant les propriétés spectrales et contextuelles des objets qu'elles représentent. L'accent est mis sur l'efficience des algorithmes, ainsi que sur leur simplicité, de manière à augmenter leur potentiel d'implementation pour les utilisateurs. De plus, le défi de cette Thèse est de rester proche des problèmes concrets des utilisateurs d'images satellite sans pour autant perdre de vue l'intéret des méthodes proposées pour le milieu du machine learning dont elles sont issues. En ce sens, ce travail joue la carte de la transdisciplinarité en maintenant un lien fort entre les deux sciences dans tous les développements proposés. Quatre modèles sont proposés: le premier répond au problème de la haute dimensionalité et de la redondance des données par un modèle optimisant les performances en classification en s'adaptant aux particularités de l'image. Ceci est rendu possible par un système de ranking des variables (les bandes) qui est optimisé en même temps que le modèle de base: ce faisant, seules les variables importantes pour résoudre le problème sont utilisées par le classifieur. Le manque d'information étiquétée et l'incertitude quant à sa pertinence pour le problème sont à la source des deux modèles suivants, basés respectivement sur l'apprentissage actif et les méthodes semi-supervisées: le premier permet d'améliorer la qualité d'un ensemble d'entraînement par interaction directe entre l'utilisateur et la machine, alors que le deuxième utilise les pixels non étiquetés pour améliorer la description des données disponibles et la robustesse du modèle. Enfin, le dernier modèle proposé considère la question plus théorique de la structure entre les outputs: l'intègration de cette source d'information, jusqu'à présent jamais considérée en télédétection, ouvre des nouveaux défis de recherche. Advanced kernel methods for remote sensing image classification Devis Tuia Institut de Géomatique et d'Analyse du Risque September 2009 Abstract The technical developments in recent years have brought the quantity and quality of digital information to an unprecedented level, as enormous archives of satellite images are available to the users. However, even if these advances open more and more possibilities in the use of digital imagery, they also rise several problems of storage and treatment. The latter is considered in this Thesis: the processing of very high spatial and spectral resolution images is treated with approaches based on data-driven algorithms relying on kernel methods. In particular, the problem of image classification, i.e. the categorization of the image's pixels into a reduced number of classes reflecting spectral and contextual properties, is studied through the different models presented. The accent is put on algorithmic efficiency and the simplicity of the approaches proposed, to avoid too complex models that would not be used by users. The major challenge of the Thesis is to remain close to concrete remote sensing problems, without losing the methodological interest from the machine learning viewpoint: in this sense, this work aims at building a bridge between the machine learning and remote sensing communities and all the models proposed have been developed keeping in mind the need for such a synergy. Four models are proposed: first, an adaptive model learning the relevant image features has been proposed to solve the problem of high dimensionality and collinearity of the image features. This model provides automatically an accurate classifier and a ranking of the relevance of the single features. The scarcity and unreliability of labeled. information were the common root of the second and third models proposed: when confronted to such problems, the user can either construct the labeled set iteratively by direct interaction with the machine or use the unlabeled data to increase robustness and quality of the description of data. Both solutions have been explored resulting into two methodological contributions, based respectively on active learning and semisupervised learning. Finally, the more theoretical issue of structured outputs has been considered in the last model, which, by integrating outputs similarity into a model, opens new challenges and opportunities for remote sensing image processing.

Analysis and spatial modeling of the indoor radon Swiss data

The present research deals with an important public health threat, which is the pollution created by radon gas accumulation inside dwellings. The spatial modeling of indoor radon in Switzerland is particularly complex and challenging because of many influencing factors that should be taken into account. Indoor radon data analysis must be addressed from both a statistical and a spatial point of view. As a multivariate process, it was important at first to define the influence of each factor. In particular, it was important to define the influence of geology as being closely associated to indoor radon. This association was indeed observed for the Swiss data but not probed to be the sole determinant for the spatial modeling. The statistical analysis of data, both at univariate and multivariate level, was followed by an exploratory spatial analysis. Many tools proposed in the literature were tested and adapted, including fractality, declustering and moving windows methods. The use of Quan-tité Morisita Index (QMI) as a procedure to evaluate data clustering in function of the radon level was proposed. The existing methods of declustering were revised and applied in an attempt to approach the global histogram parameters. The exploratory phase comes along with the definition of multiple scales of interest for indoor radon mapping in Switzerland. The analysis was done with a top-to-down resolution approach, from regional to local lev¬els in order to find the appropriate scales for modeling. In this sense, data partition was optimized in order to cope with stationary conditions of geostatistical models. Common methods of spatial modeling such as Κ Nearest Neighbors (KNN), variography and General Regression Neural Networks (GRNN) were proposed as exploratory tools. In the following section, different spatial interpolation methods were applied for a par-ticular dataset. A bottom to top method complexity approach was adopted and the results were analyzed together in order to find common definitions of continuity and neighborhood parameters. Additionally, a data filter based on cross-validation was tested with the purpose of reducing noise at local scale (the CVMF). At the end of the chapter, a series of test for data consistency and methods robustness were performed. This lead to conclude about the importance of data splitting and the limitation of generalization methods for reproducing statistical distributions. The last section was dedicated to modeling methods with probabilistic interpretations. Data transformation and simulations thus allowed the use of multigaussian models and helped take the indoor radon pollution data uncertainty into consideration. The catego-rization transform was presented as a solution for extreme values modeling through clas-sification. Simulation scenarios were proposed, including an alternative proposal for the reproduction of the global histogram based on the sampling domain. The sequential Gaussian simulation (SGS) was presented as the method giving the most complete information, while classification performed in a more robust way. An error measure was defined in relation to the decision function for data classification hardening. Within the classification methods, probabilistic neural networks (PNN) show to be better adapted for modeling of high threshold categorization and for automation. Support vector machines (SVM) on the contrary performed well under balanced category conditions. In general, it was concluded that a particular prediction or estimation method is not better under all conditions of scale and neighborhood definitions. Simulations should be the basis, while other methods can provide complementary information to accomplish an efficient indoor radon decision making.

International Society of Urological Pathology (ISUP) Consensus Conference on Handling and Staging of Radical Prostatectomy Specimens. Working group 4: seminal vesicles and lymph nodes.

The 2009 International Society of Urological Pathology Consensus Conference in Boston made recommendations regarding the standardization of pathology reporting of radical prostatectomy specimens. Issues relating to the infiltration of tumor into the seminal vesicles and regional lymph nodes were coordinated by working group 4. There was a consensus that complete blocking of the seminal vesicles was not necessary, although sampling of the junction of the seminal vesicles and prostate was mandatory. There was consensus that sampling of the vas deferens margins was not obligatory. There was also consensus that muscular wall invasion of the extraprostatic seminal vesicle only should be regarded as seminal vesicle invasion. Categorization into types of seminal vesicle spread was agreed by consensus to be not necessary. For examination of lymph nodes, there was consensus that special techniques such as frozen sectioning were of use only in high-risk cases. There was no consensus on the optimal sampling method for pelvic lymph node dissection specimens, although there was consensus that all lymph nodes should be completely blocked as a minimum. There was also a consensus that a count of the number of lymph nodes harvested should be attempted. In view of recent evidence, there was consensus that the diameter of the largest lymph node metastasis should be measured. These consensus decisions will hopefully clarify the difficult areas of pathological assessment in radical prostatectomy evaluation and improve the concordance of research series to allow more accurate assessment of patient prognosis.

Cytotoxic T-cell and NK-cell Lymphomas: Current Questions and Controversies.

The cytotoxic T-cell and natural killer (NK)-cell lymphomas and related disorders are important but relatively rare lymphoid neoplasms that frequently are a challenge for practicing pathologists. This selective review, based on a meeting of the International Lymphoma Study Group, briefly reviews T-cell and NK-cell development and addresses questions related to the importance of precise cell lineage (αβ-type T cell, γδ T cell, or NK cell), the implications of Epstein-Barr virus infection, the significance of anatomic location including nodal disease, and the question of further categorization of enteropathy-associated T-cell lymphomas. Finally, developments subsequent to the 2008 World Health Organization Classification, including the recognition of indolent NK-cell and T-cell disorders of the gastrointestinal tract are presented.

The brain tracks the energetic value in food images.

Do our brains implicitly track the energetic content of the foods we see? Using electrical neuroimaging of visual evoked potentials (VEPs) we show that the human brain can rapidly discern food's energetic value, vis à vis its fat content, solely from its visual presentation. Responses to images of high-energy and low-energy food differed over two distinct time periods. The first period, starting at approximately 165 ms post-stimulus onset, followed from modulations in VEP topography and by extension in the configuration of the underlying brain network. Statistical comparison of source estimations identified differences distributed across a wide network including both posterior occipital regions and temporo-parietal cortices typically associated with object processing, and also inferior frontal cortices typically associated with decision-making. During a successive processing stage (starting at approximately 300 ms), responses differed both topographically and in terms of strength, with source estimations differing predominantly within prefrontal cortical regions implicated in reward assessment and decision-making. These effects occur orthogonally to the task that is actually being performed and suggest that reward properties such as a food's energetic content are treated rapidly and in parallel by a distributed network of brain regions involved in object categorization, reward assessment, and decision-making.

An assurance-based model to holistically assess the information security posture

EXECUTIVE SUMMARY : Evaluating Information Security Posture within an organization is becoming a very complex task. Currently, the evaluation and assessment of Information Security are commonly performed using frameworks, methodologies and standards which often consider the various aspects of security independently. Unfortunately this is ineffective because it does not take into consideration the necessity of having a global and systemic multidimensional approach to Information Security evaluation. At the same time the overall security level is globally considered to be only as strong as its weakest link. This thesis proposes a model aiming to holistically assess all dimensions of security in order to minimize the likelihood that a given threat will exploit the weakest link. A formalized structure taking into account all security elements is presented; this is based on a methodological evaluation framework in which Information Security is evaluated from a global perspective. This dissertation is divided into three parts. Part One: Information Security Evaluation issues consists of four chapters. Chapter 1 is an introduction to the purpose of this research purpose and the Model that will be proposed. In this chapter we raise some questions with respect to "traditional evaluation methods" as well as identifying the principal elements to be addressed in this direction. Then we introduce the baseline attributes of our model and set out the expected result of evaluations according to our model. Chapter 2 is focused on the definition of Information Security to be used as a reference point for our evaluation model. The inherent concepts of the contents of a holistic and baseline Information Security Program are defined. Based on this, the most common roots-of-trust in Information Security are identified. Chapter 3 focuses on an analysis of the difference and the relationship between the concepts of Information Risk and Security Management. Comparing these two concepts allows us to identify the most relevant elements to be included within our evaluation model, while clearing situating these two notions within a defined framework is of the utmost importance for the results that will be obtained from the evaluation process. Chapter 4 sets out our evaluation model and the way it addresses issues relating to the evaluation of Information Security. Within this Chapter the underlying concepts of assurance and trust are discussed. Based on these two concepts, the structure of the model is developed in order to provide an assurance related platform as well as three evaluation attributes: "assurance structure", "quality issues", and "requirements achievement". Issues relating to each of these evaluation attributes are analysed with reference to sources such as methodologies, standards and published research papers. Then the operation of the model is discussed. Assurance levels, quality levels and maturity levels are defined in order to perform the evaluation according to the model. Part Two: Implementation of the Information Security Assurance Assessment Model (ISAAM) according to the Information Security Domains consists of four chapters. This is the section where our evaluation model is put into a welldefined context with respect to the four pre-defined Information Security dimensions: the Organizational dimension, Functional dimension, Human dimension, and Legal dimension. Each Information Security dimension is discussed in a separate chapter. For each dimension, the following two-phase evaluation path is followed. The first phase concerns the identification of the elements which will constitute the basis of the evaluation: ? Identification of the key elements within the dimension; ? Identification of the Focus Areas for each dimension, consisting of the security issues identified for each dimension; ? Identification of the Specific Factors for each dimension, consisting of the security measures or control addressing the security issues identified for each dimension. The second phase concerns the evaluation of each Information Security dimension by: ? The implementation of the evaluation model, based on the elements identified for each dimension within the first phase, by identifying the security tasks, processes, procedures, and actions that should have been performed by the organization to reach the desired level of protection; ? The maturity model for each dimension as a basis for reliance on security. For each dimension we propose a generic maturity model that could be used by every organization in order to define its own security requirements. Part three of this dissertation contains the Final Remarks, Supporting Resources and Annexes. With reference to the objectives of our thesis, the Final Remarks briefly analyse whether these objectives were achieved and suggest directions for future related research. Supporting resources comprise the bibliographic resources that were used to elaborate and justify our approach. Annexes include all the relevant topics identified within the literature to illustrate certain aspects of our approach. Our Information Security evaluation model is based on and integrates different Information Security best practices, standards, methodologies and research expertise which can be combined in order to define an reliable categorization of Information Security. After the definition of terms and requirements, an evaluation process should be performed in order to obtain evidence that the Information Security within the organization in question is adequately managed. We have specifically integrated into our model the most useful elements of these sources of information in order to provide a generic model able to be implemented in all kinds of organizations. The value added by our evaluation model is that it is easy to implement and operate and answers concrete needs in terms of reliance upon an efficient and dynamic evaluation tool through a coherent evaluation system. On that basis, our model could be implemented internally within organizations, allowing them to govern better their Information Security. RÉSUMÉ : Contexte général de la thèse L'évaluation de la sécurité en général, et plus particulièrement, celle de la sécurité de l'information, est devenue pour les organisations non seulement une mission cruciale à réaliser, mais aussi de plus en plus complexe. A l'heure actuelle, cette évaluation se base principalement sur des méthodologies, des bonnes pratiques, des normes ou des standards qui appréhendent séparément les différents aspects qui composent la sécurité de l'information. Nous pensons que cette manière d'évaluer la sécurité est inefficiente, car elle ne tient pas compte de l'interaction des différentes dimensions et composantes de la sécurité entre elles, bien qu'il soit admis depuis longtemps que le niveau de sécurité globale d'une organisation est toujours celui du maillon le plus faible de la chaîne sécuritaire. Nous avons identifié le besoin d'une approche globale, intégrée, systémique et multidimensionnelle de l'évaluation de la sécurité de l'information. En effet, et c'est le point de départ de notre thèse, nous démontrons que seule une prise en compte globale de la sécurité permettra de répondre aux exigences de sécurité optimale ainsi qu'aux besoins de protection spécifiques d'une organisation. Ainsi, notre thèse propose un nouveau paradigme d'évaluation de la sécurité afin de satisfaire aux besoins d'efficacité et d'efficience d'une organisation donnée. Nous proposons alors un modèle qui vise à évaluer d'une manière holistique toutes les dimensions de la sécurité, afin de minimiser la probabilité qu'une menace potentielle puisse exploiter des vulnérabilités et engendrer des dommages directs ou indirects. Ce modèle se base sur une structure formalisée qui prend en compte tous les éléments d'un système ou programme de sécurité. Ainsi, nous proposons un cadre méthodologique d'évaluation qui considère la sécurité de l'information à partir d'une perspective globale. Structure de la thèse et thèmes abordés Notre document est structuré en trois parties. La première intitulée : « La problématique de l'évaluation de la sécurité de l'information » est composée de quatre chapitres. Le chapitre 1 introduit l'objet de la recherche ainsi que les concepts de base du modèle d'évaluation proposé. La maniéré traditionnelle de l'évaluation de la sécurité fait l'objet d'une analyse critique pour identifier les éléments principaux et invariants à prendre en compte dans notre approche holistique. Les éléments de base de notre modèle d'évaluation ainsi que son fonctionnement attendu sont ensuite présentés pour pouvoir tracer les résultats attendus de ce modèle. Le chapitre 2 se focalise sur la définition de la notion de Sécurité de l'Information. Il ne s'agit pas d'une redéfinition de la notion de la sécurité, mais d'une mise en perspectives des dimensions, critères, indicateurs à utiliser comme base de référence, afin de déterminer l'objet de l'évaluation qui sera utilisé tout au long de notre travail. Les concepts inhérents de ce qui constitue le caractère holistique de la sécurité ainsi que les éléments constitutifs d'un niveau de référence de sécurité sont définis en conséquence. Ceci permet d'identifier ceux que nous avons dénommés « les racines de confiance ». Le chapitre 3 présente et analyse la différence et les relations qui existent entre les processus de la Gestion des Risques et de la Gestion de la Sécurité, afin d'identifier les éléments constitutifs du cadre de protection à inclure dans notre modèle d'évaluation. Le chapitre 4 est consacré à la présentation de notre modèle d'évaluation Information Security Assurance Assessment Model (ISAAM) et la manière dont il répond aux exigences de l'évaluation telle que nous les avons préalablement présentées. Dans ce chapitre les concepts sous-jacents relatifs aux notions d'assurance et de confiance sont analysés. En se basant sur ces deux concepts, la structure du modèle d'évaluation est développée pour obtenir une plateforme qui offre un certain niveau de garantie en s'appuyant sur trois attributs d'évaluation, à savoir : « la structure de confiance », « la qualité du processus », et « la réalisation des exigences et des objectifs ». Les problématiques liées à chacun de ces attributs d'évaluation sont analysées en se basant sur l'état de l'art de la recherche et de la littérature, sur les différentes méthodes existantes ainsi que sur les normes et les standards les plus courants dans le domaine de la sécurité. Sur cette base, trois différents niveaux d'évaluation sont construits, à savoir : le niveau d'assurance, le niveau de qualité et le niveau de maturité qui constituent la base de l'évaluation de l'état global de la sécurité d'une organisation. La deuxième partie: « L'application du Modèle d'évaluation de l'assurance de la sécurité de l'information par domaine de sécurité » est elle aussi composée de quatre chapitres. Le modèle d'évaluation déjà construit et analysé est, dans cette partie, mis dans un contexte spécifique selon les quatre dimensions prédéfinies de sécurité qui sont: la dimension Organisationnelle, la dimension Fonctionnelle, la dimension Humaine, et la dimension Légale. Chacune de ces dimensions et son évaluation spécifique fait l'objet d'un chapitre distinct. Pour chacune des dimensions, une évaluation en deux phases est construite comme suit. La première phase concerne l'identification des éléments qui constituent la base de l'évaluation: ? Identification des éléments clés de l'évaluation ; ? Identification des « Focus Area » pour chaque dimension qui représentent les problématiques se trouvant dans la dimension ; ? Identification des « Specific Factors » pour chaque Focus Area qui représentent les mesures de sécurité et de contrôle qui contribuent à résoudre ou à diminuer les impacts des risques. La deuxième phase concerne l'évaluation de chaque dimension précédemment présentées. Elle est constituée d'une part, de l'implémentation du modèle général d'évaluation à la dimension concernée en : ? Se basant sur les éléments spécifiés lors de la première phase ; ? Identifiant les taches sécuritaires spécifiques, les processus, les procédures qui auraient dû être effectués pour atteindre le niveau de protection souhaité. D'autre part, l'évaluation de chaque dimension est complétée par la proposition d'un modèle de maturité spécifique à chaque dimension, qui est à considérer comme une base de référence pour le niveau global de sécurité. Pour chaque dimension nous proposons un modèle de maturité générique qui peut être utilisé par chaque organisation, afin de spécifier ses propres exigences en matière de sécurité. Cela constitue une innovation dans le domaine de l'évaluation, que nous justifions pour chaque dimension et dont nous mettons systématiquement en avant la plus value apportée. La troisième partie de notre document est relative à la validation globale de notre proposition et contient en guise de conclusion, une mise en perspective critique de notre travail et des remarques finales. Cette dernière partie est complétée par une bibliographie et des annexes. Notre modèle d'évaluation de la sécurité intègre et se base sur de nombreuses sources d'expertise, telles que les bonnes pratiques, les normes, les standards, les méthodes et l'expertise de la recherche scientifique du domaine. Notre proposition constructive répond à un véritable problème non encore résolu, auquel doivent faire face toutes les organisations, indépendamment de la taille et du profil. Cela permettrait à ces dernières de spécifier leurs exigences particulières en matière du niveau de sécurité à satisfaire, d'instancier un processus d'évaluation spécifique à leurs besoins afin qu'elles puissent s'assurer que leur sécurité de l'information soit gérée d'une manière appropriée, offrant ainsi un certain niveau de confiance dans le degré de protection fourni. Nous avons intégré dans notre modèle le meilleur du savoir faire, de l'expérience et de l'expertise disponible actuellement au niveau international, dans le but de fournir un modèle d'évaluation simple, générique et applicable à un grand nombre d'organisations publiques ou privées. La valeur ajoutée de notre modèle d'évaluation réside précisément dans le fait qu'il est suffisamment générique et facile à implémenter tout en apportant des réponses sur les besoins concrets des organisations. Ainsi notre proposition constitue un outil d'évaluation fiable, efficient et dynamique découlant d'une approche d'évaluation cohérente. De ce fait, notre système d'évaluation peut être implémenté à l'interne par l'entreprise elle-même, sans recourir à des ressources supplémentaires et lui donne également ainsi la possibilité de mieux gouverner sa sécurité de l'information.

Defining risk thresholds for a 10-year probability of hip fracture model that combines clinical risk factors and quantitative ultrasound: results using the EPISEM cohort.

Using a large prospective cohort of over 12,000 women, we determined 2 thresholds (high risk and low risk of hip fracture) to use in a 10-yr hip fracture probability model that we had previously described, a model combining the heel stiffness index measured by quantitative ultrasound (QUS) and a set of easily determined clinical risk factors (CRFs). The model identified a higher percentage of women with fractures as high risk than a previously reported risk score that combined QUS and CRF. In addition, it categorized women in a way that was quite consistent with the categorization that occurred using dual X-ray absorptiometry (DXA) and the World Health Organization (WHO) classification system; the 2 methods identified similar percentages of women with and without fractures in each of their 3 categories, but the 2 identified only in part the same women. Nevertheless, combining our composite probability model with DXA in a case findings strategy will likely further improve the detection of women at high risk of fragility hip fracture. We conclude that the currently proposed model may be of some use as an alternative to the WHO classification criteria for osteoporosis, at least when access to DXA is limited.

Voir le monde à travers les lunettes du métacontraste: un modèle mathématique inspiré de la psychologie sociale

La théorie de l'autocatégorisation est une théorie de psychologie sociale qui porte sur la relation entre l'individu et le groupe. Elle explique le comportement de groupe par la conception de soi et des autres en tant que membres de catégories sociales, et par l'attribution aux individus des caractéristiques prototypiques de ces catégories. Il s'agit donc d'une théorie de l'individu qui est censée expliquer des phénomènes collectifs. Les situations dans lesquelles un grand nombre d'individus interagissent de manière non triviale génèrent typiquement des comportements collectifs complexes qui sont difficiles à prévoir sur la base des comportements individuels. La simulation informatique de tels systèmes est un moyen fiable d'explorer de manière systématique la dynamique du comportement collectif en fonction des spécifications individuelles. Dans cette thèse, nous présentons un modèle formel d'une partie de la théorie de l'autocatégorisation appelée principe du métacontraste. À partir de la distribution d'un ensemble d'individus sur une ou plusieurs dimensions comparatives, le modèle génère les catégories et les prototypes associés. Nous montrons que le modèle se comporte de manière cohérente par rapport à la théorie et est capable de répliquer des données expérimentales concernant divers phénomènes de groupe, dont par exemple la polarisation. De plus, il permet de décrire systématiquement les prédictions de la théorie dont il dérive, notamment dans des situations nouvelles. Au niveau collectif, plusieurs dynamiques peuvent être observées, dont la convergence vers le consensus, vers une fragmentation ou vers l'émergence d'attitudes extrêmes. Nous étudions également l'effet du réseau social sur la dynamique et montrons qu'à l'exception de la vitesse de convergence, qui augmente lorsque les distances moyennes du réseau diminuent, les types de convergences dépendent peu du réseau choisi. Nous constatons d'autre part que les individus qui se situent à la frontière des groupes (dans le réseau social ou spatialement) ont une influence déterminante sur l'issue de la dynamique. Le modèle peut par ailleurs être utilisé comme un algorithme de classification automatique. Il identifie des prototypes autour desquels sont construits des groupes. Les prototypes sont positionnés de sorte à accentuer les caractéristiques typiques des groupes, et ne sont pas forcément centraux. Enfin, si l'on considère l'ensemble des pixels d'une image comme des individus dans un espace de couleur tridimensionnel, le modèle fournit un filtre qui permet d'atténuer du bruit, d'aider à la détection d'objets et de simuler des biais de perception comme l'induction chromatique. Abstract Self-categorization theory is a social psychology theory dealing with the relation between the individual and the group. It explains group behaviour through self- and others' conception as members of social categories, and through the attribution of the proto-typical categories' characteristics to the individuals. Hence, it is a theory of the individual that intends to explain collective phenomena. Situations involving a large number of non-trivially interacting individuals typically generate complex collective behaviours, which are difficult to anticipate on the basis of individual behaviour. Computer simulation of such systems is a reliable way of systematically exploring the dynamics of the collective behaviour depending on individual specifications. In this thesis, we present a formal model of a part of self-categorization theory named metacontrast principle. Given the distribution of a set of individuals on one or several comparison dimensions, the model generates categories and their associated prototypes. We show that the model behaves coherently with respect to the theory and is able to replicate experimental data concerning various group phenomena, for example polarization. Moreover, it allows to systematically describe the predictions of the theory from which it is derived, specially in unencountered situations. At the collective level, several dynamics can be observed, among which convergence towards consensus, towards frag-mentation or towards the emergence of extreme attitudes. We also study the effect of the social network on the dynamics and show that, except for the convergence speed which raises as the mean distances on the network decrease, the observed convergence types do not depend much on the chosen network. We further note that individuals located at the border of the groups (whether in the social network or spatially) have a decisive influence on the dynamics' issue. In addition, the model can be used as an automatic classification algorithm. It identifies prototypes around which groups are built. Prototypes are positioned such as to accentuate groups' typical characteristics and are not necessarily central. Finally, if we consider the set of pixels of an image as individuals in a three-dimensional color space, the model provides a filter that allows to lessen noise, to help detecting objects and to simulate perception biases such as chromatic induction.

The REporting of Studies Conducted Using Observational Routinely-Collected Health Data (RECORD) Statement: Methods for Arriving at Consensus and Developing Reporting Guidelines.

OBJECTIVE: Routinely collected health data, collected for administrative and clinical purposes, without specific a priori research questions, are increasingly used for observational, comparative effectiveness, health services research, and clinical trials. The rapid evolution and availability of routinely collected data for research has brought to light specific issues not addressed by existing reporting guidelines. The aim of the present project was to determine the priorities of stakeholders in order to guide the development of the REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement. METHODS: Two modified electronic Delphi surveys were sent to stakeholders. The first determined themes deemed important to include in the RECORD statement, and was analyzed using qualitative methods. The second determined quantitative prioritization of the themes based on categorization of manuscript headings. The surveys were followed by a meeting of RECORD working committee, and re-engagement with stakeholders via an online commentary period. RESULTS: The qualitative survey (76 responses of 123 surveys sent) generated 10 overarching themes and 13 themes derived from existing STROBE categories. Highest-rated overall items for inclusion were: Disease/exposure identification algorithms; Characteristics of the population included in databases; and Characteristics of the data. In the quantitative survey (71 responses of 135 sent), the importance assigned to each of the compiled themes varied depending on the manuscript section to which they were assigned. Following the working committee meeting, online ranking by stakeholders provided feedback and resulted in revision of the final checklist. CONCLUSIONS: The RECORD statement incorporated the suggestions provided by a large, diverse group of stakeholders to create a reporting checklist specific to observational research using routinely collected health data. Our findings point to unique aspects of studies conducted with routinely collected health data and the perceived need for better reporting of methodological issues.

From bird to sparrow: Learning-induced modulations in fine-grained semantic discrimination.

Recognition of environmental sounds is believed to proceed through discrimination steps from broad to more narrow categories. Very little is known about the neural processes that underlie fine-grained discrimination within narrow categories or about their plasticity in relation to newly acquired expertise. We investigated how the cortical representation of birdsongs is modulated by brief training to recognize individual species. During a 60-minute session, participants learned to recognize a set of birdsongs; they improved significantly their performance for trained (T) but not control species (C), which were counterbalanced across participants. Auditory evoked potentials (AEPs) were recorded during pre- and post-training sessions. Pre vs. post changes in AEPs were significantly different between T and C i) at 206-232ms post stimulus onset within a cluster on the anterior part of the left superior temporal gyrus; ii) at 246-291ms in the left middle frontal gyrus; and iii) 512-545ms in the left middle temporal gyrus as well as bilaterally in the cingulate cortex. All effects were driven by weaker activity for T than C species. Thus, expertise in discriminating T species modulated early stages of semantic processing, during and immediately after the time window that sustains the discrimination between human vs. animal vocalizations. Moreover, the training-induced plasticity is reflected by the sharpening of a left lateralized semantic network, including the anterior part of the temporal convexity and the frontal cortex. Training to identify birdsongs influenced, however, also the processing of C species, but at a much later stage. Correct discrimination of untrained sounds seems to require an additional step which results from lower-level features analysis such as apperception. We therefore suggest that the access to objects within an auditory semantic category is different and depends on subject's level of expertise. More specifically, correct intra-categorical auditory discrimination for untrained items follows the temporal hierarchy and transpires in a late stage of semantic processing. On the other hand, correct categorization of individually trained stimuli occurs earlier, during a period contemporaneous with human vs. animal vocalization discrimination, and involves a parallel semantic pathway requiring expertise.

Facing Structural Disadvantage: The Role of Ingroup Connectedness

This thesis focuses on the social-psychological factors that help coping with structural disadvantage, and specifically on the role of cohesive ingroups and the sense of connectedness and efficacy they entail in this process. It aims to complement existing group-based models of coping that are grounded in a categorization perspective to groups and consequently focus exclusively on the large-scale categories made salient in intergroup contexts of comparisons. The dissertation accomplishes this aim through a reconsideration of between-persons relational interdependence as a sufficient and independent antecedent of a sense of groupness, and the benefits that a sense of group connectedness in one's direct environment, regardless of the categorical or relational basis of groupness, might have in the everyday struggles of disadvantaged group members. The three empirical papers aim to validate this approach, outlined in the first theoretical introduction, by testing derived hypotheses. They are based on data collected with youth populations (15-30) from three institutions in French-speaking Switzerland within the context of a larger project on youth transitions. Methods of data collection are paper-pencil questionnaires and in-depth interviews with a selected sub-sample of participants. The key argument of the first paper is that members of socially disadvantaged categories face higher barriers to their life project and that a general sense of connectedness, either based on categorical identities or other proximal groups and relations, mitigates the feeling of powerlessness associated with this experience. The second paper develops and tests a model that defines individual needs satisfaction as antecedent of self-group bonds and the efficacy beliefs derived from these intragroup bonds as the mechanism underlining the role of ingroups in coping. The third paper highlights the complexities that might be associated with the construction of a sense of groupness directly from intergroup comparisons and categorization-based disadvantage, and points out a more subtle understanding of the processes underling the emergence of groupness out of the situation of structural disadvantage. Overall, the findings confirm the central role of ingroups in coping with structural disadvantage and the importance of an understanding of groupness and its role that goes beyond the dominant focus on intergroup contexts and categorization processes.