• Netherlands
• Publicationsclose
• CA close
• English close
• Hyper Article en Ligne close
• HAL Paris Nanterre close

Several chronic lymphocytic leukaemia (CLL) susceptibility loci have been reported; however, much of the heritable risk remains unidentified. Here we perform a meta-analysis of six genome-wide association studies, imputed using a merged reference panel of 1,000 Genomes and UK10K data, totalling 6,200 cases and 17,598 controls after replication. We identify nine risk loci at 1p36.11 (rs34676223, P=5.04 × 10−13), 1q42.13 (rs41271473, P=1.06 × 10−10), 4q24 (rs71597109, P=1.37 × 10−10), 4q35.1 (rs57214277, P=3.69 × 10−8), 6p21.31 (rs3800461, P=1.97 × 10−8), 11q23.2 (rs61904987, P=2.64 × 10−11), 18q21.1 (rs1036935, P=3.27 × 10−8), 19p13.3 (rs7254272, P=4.67 × 10−8) and 22q13.33 (rs140522, P=2.70 × 10−9). These new and established risk loci map to areas of active chromatin and show an over-representation of transcription factor binding for the key determinants of B-cell development and immune response. Chronic lymphocytic leukaemia has a hereditary component, much of which remains to be identified. Here, the authors perform a genome-wide association study and find new risk loci for the disease, which are associated with genes involved in immune function.

International audience; This article outlines the methods of a teacher of mathematics who works with a class of students with language disorders related to autism. By adopting the theoretical framework of the Dual Approach (ergonomics and didactics) and Maturanian epistemology, we show how an analysis of language interactions allows for a better understanding of teaching practices and their influence on the mathematical activity of students. We explore the nature of the language interactions, both verbal and non-verbal, within a specific learning project, when it comes to reinvesting, in a non-explicit way, basic mathematical objects. Our analysis highlights the richness of a practice in which the teacher proposes different types of help to trigger the mathematical activity of her students.

International audience; This paper draws from year one of a 5-year research study that seeks to examine how post- secondary mathematics students learn to use programming as a computational thinking instrument for mathematics. It focuses on how post-secondary mathematics students’ identities as mathematics learners are enacted as they engage in a programming-based mathematical investigations and applications learning environment. Specifically, the paper offers a discussion of a case of one student’s enactment of his identity while simultaneously learning to program and to use it for this kind of mathematical work. This paper highlights the importance of identity in learning mathematics and its role in the development of productive dispositions in learning to program for mathematics investigation and modeling.

In freight distribution services a required quality level may have a relevant effect on transportation costs. For this reason an evaluation tool is useful to compare different service settings and support the decision, on the base of quantitative indicators. This paper proposes a method for cost evaluation in this context and presents an application to a case study concerning a freight distribution service, which operates on a wide road network having a city centre, a peripheral urban area and a peri-urban rural zone. A simulation method is proposed to obtain real-life scenarios in order to test the method and its indicators. The performance of each indicator has been evaluated in an experimental context to produce realistic test cases, using a trip planning tool and a demand generator. First, the behaviour of the indicators is analysed with regard to the time windows width planned for the service. Then, their ability in estimating the total transportation cost to satisfy all the requests, under different time period profiles, is shown. The results confirm the ability of the set of indicators to predict with a good approximation the transportation costs and therefore to be used in supporting the service quality planning decisions. European Journal of Transport and Infrastructure Research, Vol 12 No 3 (2012)

International audience; In this paper I discuss some concepts that are part of the epistemology dimension of the Theory of Objectification—a Vygotskian theory of teaching and learning whose philosophical background comes from dialectical materialism. The epistemological concepts are presented in the first part of the paper. In the second part, they are illustrated through a classroom example.

International audience; This paper focuses on the process of university mathematics students engaging in a sequence of programming-based mathematical project tasks as part of a course. Data of this naturalistic research was collected mainly through four student projects and semi-structured individual interviews. The analysis led to narratives of students’ development process (instrumental genesis) in which enacted strategies (instrumented actions) are highlighted. In this paper we discuss a participant’s development process. Results suggest that the student, after 1 course, has appropriated programming as instrument for creating a tool for pragmatic purposes; however, not yet as instrument for mathematics investigations and applications, i.e., as an object-to-think-with (Papert, 1980).

L'apprentissage automatique (ML) a connu d'énormes succès ces dernières années et repose sur un nombre toujours croissant d'applications réelles. Cependant, la conception d'algorithmes prometteurs pour un problème spécifique nécessite toujours un effort humain considérable. L'apprentissage automatique (AutoML) a pour objectif de sortir l'homme de la boucle. AutoML est généralement traité comme un problème de sélection d’algorithme / hyper-paramètre. Les approches existantes incluent l’optimisation Bayésienne, les algorithmes évolutionnistes et l’apprentissage par renforcement. Parmi eux, auto-sklearn, qui intègre des techniques de meta-learning à l'initialisation de la recherche, occupe toujours une place de choix dans les challenges AutoML. Cette observation a orienté mes recherches vers le domaine du meta-learning. Cette orientation m'a amené à développer un nouveau cadre basé sur les processus de décision Markovien (MDP) et l'apprentissage par renforcement (RL). Après une introduction générale (chapitre 1), mon travail de thèse commence par une analyse approfondie des résultats du Challenge AutoML (chapitre 2). Cette analyse a orienté mon travail vers le meta-learning, menant tout d’abord à proposer une formulation d’AutoML en tant que problème de recommandation, puis à formuler une nouvelle conceptualisation du problème en tant que MDP (chapitre 3). Dans le cadre du MDP, le problème consiste à remplir de manière aussi rapide et efficace que possible une matrice S de meta-learning, dans laquelle les lignes correspondent aux tâches et les colonnes aux algorithmes. Un élément de matrice S (i, j) est la performance de l'algorithme j appliqué à la tâche i. La recherche efficace des meilleures valeurs dans S nous permet d’identifier rapidement les algorithmes les mieux adaptés à des tâches données. Dans le chapitre 4, nous examinons d’abord le cadre classique d’optimisation des hyper-paramètres. Au chapitre 5, une première approche de meta-learning est introduite, qui combine des techniques d'apprentissage actif et de filtrage collaboratif pour prédire les valeurs manquantes dans S. Nos dernières recherches appliquent RL au problème du MDP défini pour apprendre une politique efficace d’exploration de S. Nous appelons cette approche REVEAL et proposons une analogie avec une série de jeux pour permettre de visualiser les stratégies des agents pour révéler progressivement les informations. Cette ligne de recherche est développée au chapitre 6. Les principaux résultats de mon projet de thèse sont : 1) Sélection HP / modèle : j'ai exploré la méthode Freeze-Thaw et optimisé l'algorithme pour entrer dans le premier challenge AutoML, obtenant la 3ème place du tour final (chapitre 3). 2) ActivMetaL : j'ai conçu un nouvel algorithme pour le meta-learning actif (ActivMetaL) et l'ai comparé à d'autres méthodes de base sur des données réelles et artificielles. Cette étude a démontré qu'ActiveMetaL est généralement capable de découvrir le meilleur algorithme plus rapidement que les méthodes de base. 3) REVEAL : j'ai développé une nouvelle conceptualisation du meta-learning en tant que processus de décision Markovien et je l'ai intégrée dans le cadre plus général des jeux REVEAL. Avec un stagiaire en master, j'ai développé des agents qui apprennent (avec l'apprentissage par renforcement) à prédire le meilleur algorithme à essayer. Le travail présenté dans ma thèse est de nature empirique. Plusieurs méta-données du monde réel ont été utilisées dans cette recherche. Des méta-données artificielles et semi-artificielles sont également utilisées dans mon travail. Les résultats indiquent que RL est une approche viable de ce problème, bien qu'il reste encore beaucoup à faire pour optimiser les algorithmes et les faire passer à l’échelle aux problèmes de méta-apprentissage plus vastes. Machine Learning (ML) has enjoyed huge successes in recent years and an ever- growing number of real-world applications rely on it. However, designing promising algorithms for a specific problem still requires huge human effort. Automated Machine Learning (AutoML) aims at taking the human out of the loop and develop machines that generate / recommend good algorithms for a given ML tasks. AutoML is usually treated as an algorithm / hyper-parameter selection problems, existing approaches include Bayesian optimization, evolutionary algorithms as well as reinforcement learning. Among them, auto-sklearn which incorporates meta-learning techniques in their search initialization, ranks consistently well in AutoML challenges. This observation oriented my research to the Meta-Learning domain. This direction led me to develop a novel framework based on Markov Decision Processes (MDP) and reinforcement learning (RL).After a general introduction (Chapter 1), my thesis work starts with an in-depth analysis of the results of the AutoML challenge (Chapter 2). This analysis oriented my work towards meta-learning, leading me first to propose a formulation of AutoML as a recommendation problem, and ultimately to formulate a novel conceptualisation of the problem as a MDP (Chapter 3). In the MDP setting, the problem is brought back to filling up, as quickly and efficiently as possible, a meta-learning matrix S, in which lines correspond to ML tasks and columns to ML algorithms. A matrix element S(i, j) is the performance of algorithm j applied to task i. Searching efficiently for the best values in S allows us to identify quickly algorithms best suited to given tasks. In Chapter 4 the classical hyper-parameter optimization framework (HyperOpt) is first reviewed. In Chapter 5 a first meta-learning approach is introduced along the lines of our paper ActivMetaL that combines active learning and collaborative filtering techniques to predict the missing values in S. Our latest research applies RL to the MDP problem we defined to learn an efficient policy to explore S. We call this approach REVEAL and propose an analogy with a series of toy games to help visualize agents’ strategies to reveal information progressively, e.g. masked areas of images to be classified, or ship positions in a battleship game. This line of research is developed in Chapter 6. The main results of my PhD project are: 1) HP / model selection: I have explored the Freeze-Thaw method and optimized the algorithm to enter the first AutoML challenge, achieving 3rd place in the final round (Chapter 3). 2) ActivMetaL: I have designed a new algorithm for active meta-learning (ActivMetaL) and compared it with other baseline methods on real-world and artificial data. This study demonstrated that ActiveMetaL is generally able to discover the best algorithm faster than baseline methods. 3) REVEAL: I developed a new conceptualization of meta-learning as a Markov Decision Process and put it into the more general framework of REVEAL games. With a master student intern, I developed agents that learns (with reinforcement learning) to predict the next best algorithm to be tried. To develop this agent, we used surrogate toy tasks of REVEAL games. We then applied our methods to AutoML problems. The work presented in my thesis is empirical in nature. Several real world meta-datasets were used in this research. Artificial and semi-artificial meta-datasets are also used in my work. The results indicate that RL is a viable approach to this problem, although much work remains to be done to optimize algorithms to make them scale to larger meta-learning problems.

Live (Rose Bengal stained) foraminifera were analysed at 8 stations along a latitudinal bathymetric transect (282-4987 m) across the Portuguese margin open slope. A succession of foraminiferal assemblaghes occurs with increasing depth, related to the decrease of trophic level. Standing stocks, species richness and diversity decrease more rapidly with depth for calcareous than for agglutinated taxa. Calcareous taxa are generally restricted to the uppermost cm in mesotrophic and oligotrophic sites. Agglutinated taxa have low diversity at the most shallow, eutrophic station due to competition with calcareous foraminifera and very shallow oxygen penetration. At greater depths, their diversity and ALD10 increase in response to deeper oxygen penetration and macrofaunal bioturbation. At the deepest, oligotrophic environments, where their diversity is low, agglutinated taxa are again restricted to shallow microhabitats. Another 7 stations from ~1000 m water depth were sampled along a longitudinal transect consisting of open slope and canyon environments. Total foraminiferal standing stocks, dominated by intermediate and deep infaunal species, are much higher in canyon stations than on the open slope, related to the elevated CPE concentrations within the canyons. Low oxygen conditions decrease foraminiferal diversity on the Nazaré canyon terrace by inhibiting many shallow infaunal taxa and increasing the dominance of intermediate and deep infaunal species. Temporal variability appears to be much greater in canyon environments, where both sedimentary disturbance and organic supplies are maximal, compared to the relatively stable open slope environment. Les faunes de foraminifères vivants (colorés au Rose Bengale) ont été analyses à 8 stations le long d'un transect bathymétrique (282-4987 m) sur la marge portugaise. Du plateau externe vers la plaine abyssale, une succession d'assemblages est trouvée en relation avec la diminuation du niveau trophique. La densité des faunes, la richesse spécifique et la diversité diminuent plus rapidement avec la profondeur pour les foraminifères carbonatés que pour les foraminifères agglutinés. Aux sites mésotrophes et oligotrophes les espèces carbonatées sont généralement limitées au centimètre supérieur. Les espèces agglutinées sont peu diversifies à la station la moins profonde, à cause d'une competition avec les espèces carbonatées et la pénétration d'oxygène très faible. Aux stations plus profondes, leurs diversité et ALD10 augmentent en réponse à la penetration d'oxygène et la bioturbation plus importantes. Aux stations les plus profondes, les foraminifères agglutinés montrent une faible diversité, et sont de nouveau limités à la surface du sediment. Un transect latitudinal regroupe 7 autres stations, de pente ouverte et de canyon ouvert, à une profondeur d'environ 1000 m. Dans les canyons sousmarins, les densités des faunes, dominées par des espèces endopéliques intermédiares et profondes, sont largement supérieures, en accord avec des teneurs en CPE plus élévées. Dans le canyon de Nazaré, des faibles teneurs en oxygène causent une diversité minimale. La variabilité temporelle apparait ètre beaucoup plus importante dans les canyons sous-marins, où des perturbations sédimentaires et les apports organiques sont maximaux.