HAL - Lille 3
Not a member yet
5967 research outputs found
Sort by
Indian Buffet process dictionary learning for image inpainting
No full textInternational audienceIll-posed inverse problems call for adapted models to define relevant solutions. Dictionary learning for sparse representation is often an efficient approach. In many methods, the size of the dictionary is fixed in advance and the noise level as well as regularization parameters need some tuning. Indian Buffet process dictionary learning (IBP-DL) is a Bayesian non para-metric approach which permits to learn a dictionary with an adapted number of atoms. The noise and sparsity levels are also inferred so that the proposed approach is really non para-metric: no parameters tuning is needed. This work adapts IBP-DL to the problem of image inpainting by proposing an accelerated collapsed Gibbs sampler. Experimental results illustrate the relevance of this approach
Analyse de la propagation basée sur les graphes logiciels et les données synthétiques
No full textPrograms are everywhere in our daily life: computers and phones but also fridges, planes and so on. The main actor in the process of creating these programs is human beings. As thorough as they can be, humans are known to make involuntary errors without their awareness. Thus, once finished an already hard phase of writing a program. they have to face the maintenance phase on which they have to deal with errors they had previously made. All long their development task, developers have to continuously face their (or their colleagues) errors. This key observation arises the need of aiding developers in their development/maintenance tasks.Les programmes sont partout dans notre vie quotidienne : les ordinateurs et les téléphones, mais aussi les frigo, les avions et ainsi de suite. L'acteur principal dans la création de ces programmes est humain les êtres. Aussi minutie qu'ils peuvent être, les humains sont connus pour faire des erreurs involontaires sans leur conscience. Ainsi, une fois une phase déjà difficile d'écriture d'un programme, ils doivent faire face à la phase de maintenance sur laquelle ils doivent faire face aux erreurs qu'ils ont eu précédemment réalisé. Toute la durée de leur tâche de développement, les développeurs doivent faire face continuellement leurs erreurs (ou leurs collègues). Cette observation clé soulève la nécessité d'aider les développeurs dans leurs tâches de développement / maintenance
On the Estimation of the Density of a Directional Data Stream
No full textInternational audienceMany directional data such as wind directions can be collected extremely easily so that experiments typically yield a huge number of data points that are sequentially collected. To deal with such big data, the traditional nonparametric techniques rapidly require a lot of time to be computed and therefore become useless in practice if real time or online forecasts are expected. In this paper, we propose a recursive kernel density estimator for directional data which (i) can be updated extremely easily when a new set of observations is available and (ii) keeps asymptotically the nice features of the traditional kernel density estimator. Our methodology is based on Robbins-Monro stochastic approximations ideas. We show that our estimator outperforms the traditional techniques in terms of computational time while being extremely competitive in terms of efficiency with respect to its competitors in the sequential context considered here. We obtain expressions for its asymptotic bias and variance together with an almost sure convergence rate and an asymptotic normality result. Our technique is illustrated on a wind dataset collected in Spain. A Monte-Carlo study confirms the nice properties of our recursive estimator with respect to its non-recursive counterpart. [ABSTRACT FROM AUTHOR
Usages de tablettes à l’école primaire : quelles contraintes sur l’activité pédagogique ?
No full textInternational audienceThe EXTATE project studied the introduction of tablets in eight primary schools. Observations and interviews were conducted with teachers. On the basis of these observations and interviews, a set of specific constraints is identified here. The first of these is ecosystemic constraints - those which stem from economic choices related to the purchase of applications. Then there are technical constraints, or ergonomic constraints, related to the design of applications. A diachronic analysis shows how these constraints alter the way teachers work and lead to changes in their instructional design. Dispersed data, the external storage of students’ work and the servicialisation of applications all impose constraints on the choice of educational activities.La recherche EXTATE s’est intéressée à l’implantation de tablettes dans huit écoles primaires françaises. À partir d’observations et d’entretiens, plusieurs contraintes spécifiques sont ici mises en évidence. Il s’agit de contraintes écosystémiques émanant de choix économiques relatifs à l’achat d’applications et de contraintes techniques, ou de contraintes ergonomiques, reliées à la conception des applications. Une analyse diachronique montre en quoi ces contraintes modifient l’activité des enseignants et les conduisent à des adaptations. La dispersion des données, la délocalisation du stockage des productions d’élèves, la servicialisation des applications contraignent les activités pédagogiques proposées
Transcriptional, epigenetic and retroviral signatures identify regulatory regions involved in hematopoietic lineage commitment
No full textInternational audienceGenome-wide approaches allow investigating the molecular circuitry wiring the genetic and epigenetic programs of human somatic stem cells. Hematopoietic stem/progenitor cells (HSPC) give rise to the different blood cell types; however, the molecular basis of human hematopoietic lineage commitment is poorly characterized. Here, we define the transcriptional and epigenetic profile of human HSPC and early myeloid and erythroid progenitors by a combination of Cap Analysis of Gene Expression (CAGE), ChIP-seq and Moloney leukemia virus (MLV) integration site mapping. Most promoters and transcripts were shared by HSPC and committed progenitors, while enhancers and super-enhancers consistently changed upon differentiation, indicating that lineage commitment is essentially regulated by enhancer elements. A significant fraction of CAGE promoters differentially expressed upon commitment were novel, harbored a chromatin enhancer signature, and may identify promoters and transcribed enhancers driving cell commitment. MLV-targeted genomic regions co-mapped with cell-specific active enhancers and super-enhancers. Expression analyses, together with an enhancer functional assay, indicate that MLV integration can be used to identify bona fide developmentally regulated enhancers. Overall, this study provides an overview of transcriptional and epigenetic changes associated to HSPC lineage commitment, and a novel signature for regulatory elements involved in cell identity
Spatially-resolved protein surface microsampling from tissue sections using liquid extraction surface analysis
No full textInternational audienceTissue microenvironment characterization presents a challenge for a better understanding of the full complexity of a pathology. Unfortunately, making a precise "picture" of the disease needs an efficient microsampling method coupled to an accurate localization for performing region-dependent proteomics. Here, we present a method that enables rapid and reproducible extraction of proteins from a tissue section to analyze a specific region at a millimeter scale. The method used a liquid-microjunction extraction with conventional detergent solution for proteomics analysis. We successfully performed immunoblotting experiments and showed the possibility to retrieve and identify more than 1400 proteins from a 1-mm diameter spot size on tissue sections with a high degree of reproducibility both qualitatively and quantitatively. Moreover, the small size of the extracted region achieved by this sampling method allows the possibility to perform multiple extractions on different tissue section points. Ten points on a sagittal rat brain tissue section were analyzed and the measured proteins clearly distinguished the different parts of the brain, thus permitting precise functional mapping. We thus demonstrate that with this technology, it is possible to map the tissue microenvironment and gain an understanding of the molecular mechanisms at millimeter resolution
