Incremental neural networks for function approximation

A new strategy for incremental building of multilayer feedforward neural networks is proposed in the context of approximation of functions from R-p to R-q using noisy data. A stopping criterion based on the properties of the noise is also proposed. Experimental results for both artificial and real data are performed and two alternatives of the proposed construction strategy are compared.

A comparison of methods for differential expression analysis of RNA-seq data.

BACKGROUND: Finding genes that are differentially expressed between conditions is an integral part of understanding the molecular basis of phenotypic variation. In the past decades, DNA microarrays have been used extensively to quantify the abundance of mRNA corresponding to different genes, and more recently high-throughput sequencing of cDNA (RNA-seq) has emerged as a powerful competitor. As the cost of sequencing decreases, it is conceivable that the use of RNA-seq for differential expression analysis will increase rapidly. To exploit the possibilities and address the challenges posed by this relatively new type of data, a number of software packages have been developed especially for differential expression analysis of RNA-seq data. RESULTS: We conducted an extensive comparison of eleven methods for differential expression analysis of RNA-seq data. All methods are freely available within the R framework and take as input a matrix of counts, i.e. the number of reads mapping to each genomic feature of interest in each of a number of samples. We evaluate the methods based on both simulated data and real RNA-seq data. CONCLUSIONS: Very small sample sizes, which are still common in RNA-seq experiments, impose problems for all evaluated methods and any results obtained under such conditions should be interpreted with caution. For larger sample sizes, the methods combining a variance-stabilizing transformation with the 'limma' method for differential expression analysis perform well under many different conditions, as does the nonparametric SAMseq method.

Super-resolution of remotely sensed images with variable-pixel linear reconstruction

This paper describes the development and applications of a super-resolution method, known as Super-Resolution Variable-Pixel Linear Reconstruction. The algorithm works combining different lower resolution images in order to obtain, as a result, a higher resolution image. We show that it can make significant spatial resolution improvements to satellite images of the Earth¿s surface allowing recognition of objects with size approaching the limiting spatial resolution of the lower resolution images. The algorithm is based on the Variable-Pixel Linear Reconstruction algorithm developed by Fruchter and Hook, a well-known method in astronomy but never used for Earth remote sensing purposes. The algorithm preserves photometry, can weight input images according to the statistical significance of each pixel, and removes the effect of geometric distortion on both image shape and photometry. In this paper, we describe its development for remote sensing purposes, show the usefulness of the algorithm working with images as different to the astronomical images as the remote sensing ones, and show applications to: 1) a set of simulated multispectral images obtained from a real Quickbird image; and 2) a set of multispectral real Landsat Enhanced Thematic Mapper Plus (ETM+) images. These examples show that the algorithm provides a substantial improvement in limiting spatial resolution for both simulated and real data sets without significantly altering the multispectral content of the input low-resolution images, without amplifying the noise, and with very few artifacts.

Prediction of difficult intubation with advanced face recognition techniques: a preliminary study

Introduction: Difficult tracheal intubation remains a constant and significant source of morbidity and mortality in anaesthetic practice. Insufficient airway assessment in the preoperative period continues to be a major cause of unanticipated difficult intubation. Although many risk factors have already been identified, preoperative airway evaluation is not always regarded as a standard procedure and the respective weight of each risk factor remains unclear. Moreover the predictive scores available are not sensitive, moderately specific and often operator-dependant. In order to improve the preoperative detection of patients at risk for difficult intubation, we developed a system for automated and objective evaluation of morphologic criteria of the face and neck using video recordings and advanced techniques borrowed from face recognition. Method and results: Frontal video sequences were recorded in 5 healthy volunteers. During the video recording, subjects were requested to perform maximal flexion-extension of the neck and to open wide the mouth with tongue pulled out. A robust and real-time face tracking system was then applied, allowing to automatically identify and map a grid of 55 control points on the face, which were tracked during head motion. These points located important features of the face, such as the eyebrows, the nose, the contours of the eyes and mouth, and the external contours, including the chin. Moreover, based on this face tracking, the orientation of the head could also be estimated at each frame of the video sequence. Thus, we could infer for each frame the pitch angle of the head pose (related to the vertical rotation of the head) and obtain the degree of head extension. Morphological criteria used in the most frequent cited predictive scores were also extracted, such as mouth opening, degree of visibility of the uvula or thyreo-mental distance. Discussion and conclusion: Preliminary results suggest the high feasibility of the technique. The next step will be the application of the same automated and objective evaluation to patients who will undergo tracheal intubation. The difficulties related to intubation will be then correlated to the biometric characteristics of the patients. The objective in mind is to analyze the biometrics data with artificial intelligence algorithms to build a highly sensitive and specific predictive test.

Identification of the first nonsense CDSN mutation with expression of a truncated protein causing peeling skin syndrome type B.

BACKGROUND: Peeling skin disease (PSD), a generalized inflammatory form of peeling skin syndrome, is caused by autosomal recessive nonsense mutations in the corneodesmosin gene (CDSN). OBJECTIVES: To investigate a novel mutation in CDSN. METHODS: A 50-year-old white woman showed widespread peeling with erythema and elevated serum IgE. DNA sequencing, immunohistochemistry, Western blot and real-time polymerase chain reaction analyses of skin biopsies were performed in order to study the genetics and to characterize the molecular profile of the disease. RESULTS: Histology showed hyperkeratosis and acanthosis of the epidermis, and inflammatory infiltrates in the dermis. DNA sequencing revealed a homozygous mutation leading to a premature termination codon in CDSN: p.Gly142*. Protein analyses showed reduced expression of a 16-kDa corneodesmosin mutant in the upper epidermal layers, whereas the full-length protein was absent. CONCLUSIONS: These results are interesting regarding the genotype-phenotype correlations in diseases caused by CDSN mutations. The PSD-causing CDSN mutations identified heretofore result in total corneodesmosin loss, suggesting that PSD is due to full corneodesmosin deficiency. Here, we show for the first time that a mutant corneodesmosin can be stably expressed in some patients with PSD, and that this truncated protein is very probably nonfunctional.

New CP violation in neutrino oscillations

Measurements of CP-violating observables in neutrino oscillation experiments have been studied in the literature as a way to determine the CP-violating phase in the mixing matrix for leptons. Here we show that such observables also probe new neutrino interactions in the production or detection processes. Genuine CP violation and fake CP violation due to matter effects are sensitive to the imaginary and real parts of new couplings. The dependence of the CP asymmetry on the source-detector distance is different from the standard one and, in particular, enhanced at short distances. We estimate that future neutrino factories will be able to probe in this way new interactions that are up to four orders of magnitude weaker than the weak interactions. We discuss the possible implications for models of new physics.

Ellipsometric study of a-Si:H thin films deposited by square wave modulated rf glow discharge

Thin films of hydrogenated amorphous silicon (a‐Si:H), deposited by square wave modulated (SQWM) rf silane discharges, have been studied through spectroscopic and real time phase modulated ellipsometry. The SQMW films obtained at low mean rf power density (19 mW/cm2) have shown smaller surface roughness than those obtained in standard continuous wave (cw) rf discharges. At higher rf powers (≥56 mW/cm2), different behaviors depending on the modulating frequency have been observed. On the one hand, at low modulating frequencies (<40 Hz), the SQWM films have shown a significant increase of porosity and surface roughness as compared to cw samples. On the other, at higher modulating frequencies, the material density and roughness have been found to be similar in SQWM and cw films. Furthermore, the deposition rate of the films show more pronounced increases with the modulating frequency as the rf power is increased. Experimental results are discussed in terms of plasma negative charged species which can be relatively abundant in high rf power discharges and cause significant effects on the deposited layers through polymers, clusters, and powder formation.

Spatial planning to the test of property rights in Switzerland : an innovative land management approach to coordinate spatial planning goals with property rights interests

In Switzerland, the land management regime is characterized by a liberal attitude towards the institution of property rights, which is guaranteed by the Constitution. Under the present Swiss constitutional arrangement, authorities (municipalities) are required to take into account landowners' interests when implementing their spatial planning policy. In other words, the institution of property rights cannot be restricted easily in order to implement zoning plans and planning projects. This situation causes many problems. One of them is the gap between the way land is really used by the landowners and the way land should be used based on zoning plans. In fact, zoning plans only describe how landowners should use their property. There is no sufficient provision for handling cases where the use is not in accordance with zoning plans. In particular, landowners may not be expropriated for a non-conforming use of the land. This situation often leads to the opening of new building areas in greenfields and urban sprawl, which is in contradiction with the goals set into the Federal Law on Spatial Planning. In order to identify legal strategies of intervention to solve the problem, our paper is structured into three main parts. Firstly, we make a short description of the Swiss land management regime. Then, we focus on an innovative land management approach designed to implement zoning plans in accordance with property rights. Finally, we present a case study that shows the usefulness of the presented land management approach in practice. We develop three main results. Firstly, the land management approach brings a mechanism to involve landowners in planning projects. Coordination principle between spatial planning goals and landowners' interests is the cornerstone of all the process. Secondly, the land use is improved both in terms of space and time. Finally, the institution of property rights is not challenged, since there is no expropriation and the market stays free.

Bi-planar 2D-to-3D registration in Fourier domain for stereoscopic X-ray motion tracking

In this paper we present a new method to track bonemovements in stereoscopic X-ray image series of the kneejoint. The method is based on two different X-ray imagesets: a rotational series of acquisitions of the stillsubject knee that will allow the tomographicreconstruction of the three-dimensional volume (model),and a stereoscopic image series of orthogonal projectionsas the subject performs movements. Tracking the movementsof bones throughout the stereoscopic image series meansto determine, for each frame, the best pose of everymoving element (bone) previously identified in the 3Dreconstructed model. The quality of a pose is reflectedin the similarity between its simulated projections andthe actual radiographs. We use direct Fourierreconstruction to approximate the three-dimensionalvolume of the knee joint. Then, to avoid the expensivecomputation of digitally rendered radiographs (DRR) forpose recovery, we reformulate the tracking problem in theFourier domain. Under the hypothesis of parallel X-raybeams, we use the central-slice-projection theorem toreplace the heavy 2D-to-3D registration of projections inthe signal domain by efficient slice-to-volumeregistration in the Fourier domain. Focusing onrotational movements, the translation-relevant phaseinformation can be discarded and we only consider scalarFourier amplitudes. The core of our motion trackingalgorithm can be implemented as a classical frame-wiseslice-to-volume registration task. Preliminary results onboth synthetic and real images confirm the validity ofour approach.

eVOC: a controlled vocabulary for unifying gene expression data.

Expression data contribute significantly to the biological value of the sequenced human genome, providing extensive information about gene structure and the pattern of gene expression. ESTs, together with SAGE libraries and microarray experiment information, provide a broad and rich view of the transcriptome. However, it is difficult to perform large-scale expression mining of the data generated by these diverse experimental approaches. Not only is the data stored in disparate locations, but there is frequent ambiguity in the meaning of terms used to describe the source of the material used in the experiment. Untangling semantic differences between the data provided by different resources is therefore largely reliant on the domain knowledge of a human expert. We present here eVOC, a system which associates labelled target cDNAs for microarray experiments, or cDNA libraries and their associated transcripts with controlled terms in a set of hierarchical vocabularies. eVOC consists of four orthogonal controlled vocabularies suitable for describing the domains of human gene expression data including Anatomical System, Cell Type, Pathology and Developmental Stage. We have curated and annotated 7016 cDNA libraries represented in dbEST, as well as 104 SAGE libraries,with expression information,and provide this as an integrated, public resource that allows the linking of transcripts and libraries with expression terms. Both the vocabularies and the vocabulary-annotated libraries can be retrieved from http://www.sanbi.ac.za/evoc/. Several groups are involved in developing this resource with the aim of unifying transcript expression information.

L'exploitation des traces dans les accidents de la circulation

Les investigations dans le milieu des accidents de la circulation sont très complexes. Elles nécessitent la mise en oeuvre d'un grand nombre de spécialités venant de domaines très différents. Si certains de ces domaines sont déjà bien exploités, d'autres demeurent encore incomplets et il arrive de nos jours d'observer des lacunes dans la pratique, auxquelles il est primordial de remédier.Ce travail de thèse, intitulé « l'exploitation des traces dans les accidents de la circulation », est issu d'une réflexion interdisciplinaire entre de multiples aspects des sciences forensiques. Il s'agit principalement d'une recherche ayant pour objectif de démontrer les avantages découlant d'une synergie entre les microtraces et l'étude de la dynamique d'un accident. Afin de donner une dimension très opérationnelle à ce travail, l'ensemble des démarches entreprises a été axé de manière à optimiser l'activité des premiers intervenants sur les lieux.Après une partie introductive et ayant trait au projet de recherche, traitant des aspects théoriques de la reconstruction d'une scène d'accident, le lecteur est invité à prendre connaissance de cinq chapitres pratiques, abordés selon la doctrine « du général au particulier ». La première étape de cette partie pratique concerne l'étude de la morphologie des traces. Des séquences d'examens sont proposées pour améliorer l'interprétation des contacts entre véhicules et obstacles impliqués dans un accident. Les mécanismes de transfert des traces de peinture sont ensuite étudiés et une série de tests en laboratoire est pratiquée sur des pièces de carrosseries automobiles. Différents paramètres sont ainsi testés afin de comprendre leur impact sur la fragilité d'un système de peinture. Par la suite, une liste de cas traités (crash-tests et cas réels), apportant des informations intéressantes sur le traitement d'une affaire et permettant de confirmer les résultats obtenus est effectuée. Il s'ensuit un recueil de traces, issu de l'expérience pratique acquise et ayant pour but d'aiguiller la recherche et le prélèvement sur les lieux. Finalement, la problématique d'une banque de données « accident », permettant une gestion optimale des traces récoltées est abordée.---The investigations of traffic accidents are very complex. They require the implementation of a large number of specialties coming from very different domains. If some of these domains are already well exploited, others remain still incomplete and it happens nowadays to observe gaps in the practice, which it is essential to remedy. This thesis, entitled "the exploitation of traces in traffic accidents", arises from a multidisciplinary reflection between the different aspects of forensic science. It is primarily a research aimed to demonstrate the benefits of synergy between microtrace evidence and accidents dynamics. To give a very operational dimension to this work, all the undertaken initiatives were centred so as to optimise the activity of the first participants on the crime scene.After an introductory part treating theoretical aspects of the reconstruction of an accident scene the reader is invited to get acquainted with five practical chapters, according to the doctrine "from general to particular". For the first stage of this practical part, the problem of the morphology of traces is approached and sequences of examinations are proposed to improve the interpretation of the contacts between vehicles and obstacles involved in an accident. Afterwards, the mechanisms of transfer of traces of paint are studied and a series of tests in laboratory is practised on pieces of automobile bodies. Various parameters are thus tested to understand their impact on the fragility of a system of paint. It follows that a list of treated cases (crash-tests and real cases) is created, allowing to bring interesting information on the treatment of a case and confirm the obtained results. Then, this work goes on with a collection of traces, stemming from the acquired experience that aims to steer the research and the taking of evidence on scenes. Finally, the practical part of this thesis ends with the problem of a database « accident », allowing an optimal management of the collected traces.

Faire la ville juste : une analyse "in itinere" de la maîtrise publique d'ouvrage du projet urbain Carré de Soie (métropole lyonnaise)

Dans certaines portions des agglomérations (poches de pauvreté de centre-ville, couronnes suburbaines dégradées, espaces périurbains sans aménité), un cumul entre des inégalités sociales (pauvreté, chômage, etc.) et environnementales (exposition au bruit, aux risques industriels, etc.) peut être observé. La persistance de ces inégalités croisées dans le temps indique une tendance de fond : la capacité d'accéder à un cadre de vie de qualité n'est pas équitablement partagée parmi les individus. Ce constat interroge : comment se créent ces inégalités ? Comment infléchir cette tendance et faire la ville plus juste ?¦Apporter des réponses à cette problématique nécessite d'identifier les facteurs de causalités qui entrent en jeu dans le système de (re)production des inégalités urbaines. Le fonctionnement des marchés foncier et immobilier, la « tyrannie des petites décisions » et les politiques publiques à incidence spatiale sont principalement impliqués. Ces dernières, agissant sur tous les éléments du système, sont placées au coeur de ce travail. On va ainsi s'intéresser précisément à la manière dont les collectivités publiques pilotent la production de la ville contemporaine, en portant l'attention sur la maîtrise publique d'ouvrage (MPO) des grands projets urbains.¦Poser la question de la justice dans la fabrique de la ville implique également de questionner les référentiels normatifs de l'action publique : à quelle conception de la justice celle-ci doit- elle obéir? Quatre perspectives (radicale, substantialiste, procédurale et intégrative) sont caractérisées, chacune se traduisant par des principes d'action différenciés. Une méthodologie hybride - empruntant à la sociologie des organisations et à l'analyse des politiques publiques - vient clore le volet théorique, proposant par un détour métaphorique d'appréhender le projet urbain comme une pièce de théâtre dont le déroulement dépend du jeu d'acteurs.¦Cette méthodologie est utilisée dans le volet empirique de la recherche, qui consiste en une analyse de la MPO d'un projet urbain en cours dans la première couronne de l'agglomération lyonnaise : le Carré de Soie. Trois grands objectifs sont poursuivis : descriptif (reconstruire le scénario), analytique (évaluer la nature de la pièce : conte de fée, tragédie ou match d'improvisation ?) et prescriptif (tirer la morale de l'histoire). La description de la MPO montre le déploiement successif de quatre stratégies de pilotage, dont les implications sur les temporalités, le contenu du projet (programmes, morphologies) et les financements publics vont être déterminantes. Sur la base de l'analyse, plusieurs recommandations peuvent être formulées - importance de l'anticipation et de l'articulation entre planification et stratégie foncière notamment - pour permettre à la sphère publique de dominer le jeu et d'assurer la production de justice par le projet urbain (réalisation puis entretien des équipements et espaces publics, financement de logements de qualité à destination d'un large éventail de populations, etc.). Plus généralement, un décalage problématique peut être souligné entre les territoires stratégiques pour le développement de l'agglomération et les capacités de portage limitées des communes concernées. Ce déficit plaide pour le renforcement des capacités d'investissement de la structure intercommunale.¦La seule logique du marché (foncier, immobilier) mène à la polarisation sociale et à la production d'inégalités urbaines. Faire la ville juste nécessite une forte volonté des collectivités publiques, laquelle doit se traduire aussi bien dans l'ambition affichée - une juste hiérarchisation des priorités dans le développement urbain - que dans son opérationnalisation - une juste maîtrise publique d'ouvrage des projets urbains.¦Inner-city neighborhoods, poor outskirts, and peri-urban spaces with no amenities usually suffer from social and environmental inequalities, such as poverty, unemployment, and exposure to noise and industrial hazards. The observed persistence of these inequalities over time points to an underlying trend - namely, that access to proper living conditions is fundamentally unequal, thus eliciting the question of how such inequalities are effected and how this trend can be reversed so as to build a more equitable city.¦Providing answers to such questions requires that the causal factors at play within the system of (re)production of urban inequalities be identified. Real estate markets, "micromotives and macrobehavior", and public policies that bear on space are mostly involved. The latter are central in that they act on all the elements of the system. This thesis therefore focuses on the way public authorities shape the production of contemporary cities, by studying the public project ownership of major urban projects.¦The study of justice within the urban fabric also implies that the normative frames of reference of public action be questioned: what conception of justice should public action refer to? This thesis examines four perspectives (radical, substantialist, procedural, and integrative) each of which results in different principles of action. This theoretical part is concluded by a hybrid methodology that draws from sociology of organizations and public policy analysis and that suggests that the urban project may be understood as a play, whose outcome hinges on the actors' acting.¦This methodology is applied to the empirical analysis of the public project ownership of an ongoing urban project in the Lyon first-ring suburbs: the Carré de Soie. Three main objectives are pursued: descriptive (reconstructing the scenario), analytical (assessing the nature of the play - fairy tale, tragedy or improvisation match), and prescriptive (drawing the moral of the story). The description of the public project ownership shows the successive deployment of four control strategies, whose implications on deadlines, project content (programs, morphologies), and public funding are significant. Building on the analysis, several recommendations can be made to allow the public sphere to control the process and ensure the urban project produces equity (most notably, anticipation and articulation of planning and real- estate strategy, as well as provision and maintenance of equipment and public spaces, funding of quality housing for a wide range of populations, etc.). More generally, a gap can be highlighted between those territories that are strategic to the development of the agglomeration and the limited resources of the municipalities involved. This deficit calls for strengthening the investment abilities of the intermunicipal structure.¦By itself, the real-estate market logic brings about social polarization and urban inequalities. Building an equitable city requires a strong will on the part of public authorities, a will that must be reflected both in the stated ambition - setting priorities of urban development equitably - and in its implementation managing urban public projects fairly.

Involvement of microRNAs in the cytotoxic effects exerted by proinflammatory cytokines on pancreatic beta-cells.

OBJECTIVE: Pancreatic beta-cells exposed to proinflammatory cytokines display alterations in gene expression resulting in defective insulin secretion and apoptosis. MicroRNAs are small noncoding RNAs emerging as key regulators of gene expression. Here, we evaluated the contribution of microRNAs to cytokine-mediated beta-cell cytotoxicity. RESEARCH DESIGN AND METHODS: We used global microarray profiling and real-time PCR analysis to detect changes in microRNA expression in beta-cells exposed to cytokines and in islets of pre-diabetic NOD mice. We assessed the involvement of the microRNAs affected in cytokine-mediated beta-cell failure by modifying their expression in insulin-secreting MIN6 cells. RESULTS: We found that IL-1beta and TNF-alpha induce the expression of miR-21, miR-34a, and miR-146a both in MIN6 cells and human pancreatic islets. We further show an increase of these microRNAs in islets of NOD mice during development of pre-diabetic insulitis. Blocking miR-21, miR-34a, or miR-146a function using antisense molecules did not restore insulin-promoter activity but prevented the reduction in glucose-induced insulin secretion observed upon IL-1beta exposure. Moreover, anti-miR-34a and anti-miR-146a treatment protected MIN6 cells from cytokine-triggered cell death. CONCLUSIONS: Our data identify miR-21, miR-34a, and miR-146a as novel players in beta-cell failure elicited in vitro and in vivo by proinflammatory cytokines, notably during the development of peri-insulitis that precedes overt diabetes in NOD mice.

Methylation of the hTERT Promoter: A Novel Cancer Biomarker for Leptomeningeal Metastasis Detection in Cerebrospinal Fluids.

PURPOSE: The diagnosis of leptomeningeal metastases is usually confirmed by the finding of malignant cells by cytologic examination in the cerebrospinal fluid (CSF). More sensitive and specific cancer biomarkers may improve the detection of tumor cells in the CSF. Promoter methylation of the human telomerase reverse transcriptase (hTERT) gene characterizes most cancer cells. The aim of this study was to develop a sensitive method to detect hTERT methylation and to explore its use as a cancer biomarker in CSF. EXPERIMENTAL DESIGN: In 77 CSF specimens from 67 patients, hTERT promoter methylation was evaluated using real-time methylation-sensitive high-resolution melting (MS-HRM) and real-time TaqMan PCR and MS-HRM in a single-tube assay. RESULTS: Real-time MS-HRM assay was able to detect down to 1% hTERT-methylated DNA in a background of unmethylated DNA. PCR products were obtained from 90% (69/77) of CSF samples. No false positive hTERT was detected in the 21 non-neoplastic control cases, given to the method a specificity of 100%. The sensitivity of the real-time MS-HRM compared with the cytologic gold standard analysis was of 92% (11/12). Twenty-six CSFs from 22 patients with an hTERT-methylated primary tumor showed cytologic results suspicious for malignancy; in 17 (65%) of them, a diagnosis of leptomeningeal metastases could be confirmed by the hTERT methylation test. CONCLUSION: The hTERT real-time MS-HRM approach is fast, specific, sensitive, and could therefore become a valuable tool for diagnosis of leptomeningeal metastases as an adjunct to the traditional examination of CSF. Clin Cancer Res; 19(8); 2216-23. ©2013 AACR.

Large-scale clustering through functional embedding

We present a new framework for large-scale data clustering. The main idea is to modify functional dimensionality reduction techniques to directly optimize over discrete labels using stochastic gradient descent. Compared to methods like spectral clustering our approach solves a single optimization problem, rather than an ad-hoc two-stage optimization approach, does not require a matrix inversion, can easily encode prior knowledge in the set of implementable functions, and does not have an ?out-of-sample? problem. Experimental results on both artificial and real-world datasets show the usefulness of our approach.