In this study, an experiment aimed to integrate Global Navigation Satellite System (GNSS) atmospheric data with meteorological data into a neural network system is performed. Precipitable Water Vapor (PWV) estimates derived from GNSS are combined with surface pressure, surface temperature and relative humidity obtained continuously from ground-based meteorological stations. The work aims to develop a methodology to forecast short-term intense rainfall. Hence, all the data is sampled at one hour interval. A continuous time series of 3 years of GNSS data from one station in Lisbon, Portugal, is processed. Meteorological data from a nearby meteorological station are collected. Remote sensing data of cloud top from SEVIRI is used, providing collocated data also on an hourly basis. A 3 year time series of hourly accumulated precipitation data are also available for evaluation of the neural network results. In previous studies, it was found that time varying PWV is correlated with rainfall, with a strong increase of PWV peaking just before intense rainfall, and with a strong decrease afterwards. However, a significant amount of false positives was found, meaning that the evolution of PWV does not contain enough information to infer future rain. In this work a multilayer fitting network is used to process the GNSS and meteorological data inputs in order to estimate the target outputs, given by the hourly precipitation. It is found that the combination of GNSS data and meteorological variables processed by neural network improves the detection of heavy rainfall events and reduces the number of false positives.

We propose the use of a mathematical model in order to denoise X-ray twodimensional patterns. The model, which makes use of a generalized diffusion equation whose diffusion constant depends on the image gradients, enables to obtain an efficient reduction of pattern noise as witnessed by the computed peak of signal to noise ratio. The corresponding MATLAB code is made available.

In this contribution we deal with the problem of learning an undi- rected graph which encodes the conditional dependence relationship be- tween variables of a complex system, given a set of observations of this system. This is a very central problem of modern data analysis and it comes out every time we want to investigate a deeper relationship be- tween random variables, which is different from the classical dependence usually measured by the covariance. In particular, in this contribution we deal with the case of Gaussian Graphical Models (GGMs) for which the system of variables has a mul- tivariate gaussian distribution. We revise some of the existing techniques for such a problem and propose a smart implementation of the symmetric parallel regression technique which turns out to be very competitive for learning sparse GGMs under high dimensional data regime.

The sparse multivariate method of simulated quantiles (S-MMSQ) is applied to solve a portfolio optimization problem under value-at-risk constraints where the joint returns follow a multivariate skew-elliptical stable distribution. The S-MMSQ is a simulation-based method that is particularly useful for making parametric inference in some pathological situations where the maximum likelihood estimator is difficult to compute. The method estimates parameters by minimizing the distance between quantile-based statistics evaluated on true and synthetic data, simulated from the postulated model, penalized by adding the smoothly clipped absolute deviation l-penalty in order to achieve sparsity. The S-MMSQ aims to efficiently handle the problem of estimating large-dimensional distributions with intractable likelihood, such as the stable distributions that have been widely applied in finance to model financial returns.

Undirected graphs are useful tools for the analysis of sparse and high-dimensional data sets. In this setting the sparsity helps in reducing the complexity of the model. However, sparse graphs are usually estimated under the Gaussian paradigm thereby leading to estimates that are very sensitive to the presence of outlying observations. In this paper we deal with sparse time-varying undirected graphs, namely sparse graphs whose structure evolves over time. Our contribution is to provide a robustification of these models, in particular we propose a robust estimator which minimises the ?-divergence. We provide an algorithm for the parameter estimation and we investigate the rate of convergence of the proposed estimator.

Parameter estimation of distributions with intractable density, such as the Elliptical Stable, often involves high-dimensional integrals requiring numerical integration or approximation. This paper introduces a novel Expectation-Maximisation algorithm for fitting such models that exploits the fast Fourier integration for computing the expectation step. As a further contribution we show that by slightly modifying the objective function, the proposed algorithm also handle sparse estimation of non-Gaussian models. The method is subsequently applied to the problem of selecting the asset within a sparse non-Gaussian portfolio optimisation framework.

MIPAS on ENVISAT performed almost continuous measurements of atmospheric composition for approximately 10 years, from June 2002 to April 2012. ESA processor, based on the algorithm ORM (Optimized Retrieval Model), originally designed for the Near Real Time analysis, is currently used for the reanalysis of the full MIPAS mission. Version 7 of the full mission data was released in 2016, but further improvements have been recently performed in ORM V8 to be used in next full mission reanalysis. For these latest releases (V7 and V8) L1 data corrected for reducing the instrumental drift are used.TheinstrumentaldriftisduetoMIPASphotometricdetectorsnonlinearitiesthatchangewithtimeduetothe ageing of the instrument. Numerous species are retrieved from MIPAS measurements. Among them, CCl4 has been recently studied. This species has received increasing interest due to the so called "mystery of CCl4", since it was found that its atmospheric concentration at the surface declines with a rate significantly smaller than its lifetime-limited rate. Indeed there is a discrepancy between the atmospheric observations and the estimated distribution based on the reported production and consumption. MIPAS products generated with Version 7 of the L2 ESA algorithm were used to estimate CCl4 distributions, its trend, and atmospheric lifetime in the upper troposphere / lower stratosphere (UTLS) region. The trends derived by these observations between 2002 and 2012 as a function of both latitude and altitude confirm the decline of atmospheric mixing ratios, in agreement with ground based observations. Stratospheric trend derived from the MIPAS data are non-uniform, with some positive trends even being found in the middle stratosphere, mainly at high altitudes in the Southern Hemisphere. The variability in stratospheric trends reflects the impact of variability in stratospheric transport on trace gases and their temporal evolution.In addition to CCl4, some preliminary results obtained with the latest version of the processor (V8), that performs the analysis of a larger number of species and takes into account horizontal inhomogeneities, will be shown.

We present a lattice Boltzmann model for charged leaky dielectric multiphase fluids in the context of electrified jet simulations, which are of interest for a number of production technologies including electrospinning. The role of nonlinear rheology on the dynamics of electrified jets is considered by exploiting the Carreau model for pseudoplastic fluids. We report exploratory simulations of charged droplets at rest and under a constant electric field, and we provide results for charged jet formation under electrospinning conditions.

In this paper we revisit the problem of performing a QR-step on an unreduced Hessenberg matrix H when we know an "exact" eigenvalue ?0 of H. Under exact arithmetic, this eigenvalue will appear on diagonal of the transformed Hessenberg matrix H~ and will be decoupled from the remaining part of the Hessenberg matrix, thus resulting in a deflation. But it is well known that in finite precision arithmetic the so-called perfect shift can get blurred and that the eigenvalue ?0 can then not be deflated and/or is perturbed significantly. In this paper, we develop a new strategy for computing such a QR step so that the deflation is almost always successful. We also show how to extend this technique to double QR-steps with complex conjugate shifts.

The beneficial effects of physical activity for the prevention and management of several chronic diseases are widely recognized. Mathematical modeling of the effects of physical exercise in body metabolism and in particular its influence on the control of glucose homeostasis is of primary importance in the development of eHealth monitoring devices for a personalized medicine. Nonetheless, to date only a few mathematical models have been aiming at this specific purpose. We have developed a whole-body computational model of the effects on metabolic homeostasis of a bout of physical exercise. Built upon an existing model, it allows to detail better both subjects' characteristics and physical exercise, thus determining to a greater extent the dynamics of the hormones and the metabolites considered.

Recent advances in omics profiling technologies yield ever larger amounts of molecular data. Yet, the elucidation of the molecular basis of human diseases remains an unsolved challenge. The analysis of multi-scale omics data requires integrative bioinformatic tools capable of multi-modal computing and multi-scale modeling. Unsupervised learning approaches are frequently employed to identify biomolecules and pathways involved in specific diseases. However, classical clustering is hardly suitable to analyse, e.g., gene expression data conjointly with experimental conditions and molecular pathway information. Since we are interested in gene sets displaying a consistent behaviour across different conditions, both genes and samples have to be clustered simultaneously employing models respecting the heterogeneity of such multi-scale data. To this end, we aim for extending bi-clustering approaches by including information encoded in biological networks. Methods BiCluE (Sun et al. 2013) has been the first software package tackling the weighted bi-cluster editing problem. It pro- vides an exact algorithm based on fixed-parameter tractability (FPT). The bi-cluster editing problem is formulated as a bi-partite graph connecting features and samples. We then transform this graph into a disjunct set of bi-cliques while minimizing the editing costs (e.g., number of edges to be added/removed). Even though BiCluE yields potent solutions in many scenarios such as novel genotype-phenotype associations in GWAS data, it does not consider intrinsic feature relationships, e.g., interactions between proteins or regulatory interactions between genes. Therefore, we propose an extension of the BiCluE algorithm by mapping molecular interaction networks onto the bi-partite graph such that we impose constraints that force bi-cliques to respect intrinsic feature relationships. This reduces the computational com- plexity from O(4k) to O(2k), with k being the cluster editing costs due to a drastic reduction of the search space. Ad- ditionally, this model straight-forwardly allows incorporation of multi-scale data depending on the integrated network. Results and conclusions We demonstrate the validity and efficiency of our extension to BiCluE on simulated data. In general, such network- constrained bi-clustering approaches do not only allow for more stable feature selection, they also lead to more coherent functional enrichment, improving interpretability with respect to systems biology and systems medicine while being straight-forwardly applicable to multi-scale omics data.

Biological interaction databases can be exploited by pathway-level enrichment methods for downstream analyses in biological and biomedical settings. Classical enrichment methods rely on predefined lists of pathways, biasing the search towards known pathways and risking to overlook unknown, yet important functional modules. To overcome this limitation, so-called de novo network enrichment approaches extract novel pathways from large, molecular interaction networks given molecular profiles of patients, e.g. gene expression, promoter methylation, etc. Network enrichment of molecular profiling data is challenging due to noise and incompleteness of both the data them- selves and the networks. KeyPathwayMiner (KPM) jointly considers multi-scale molecular profiles to extract subnet- works enriched for de-regulated genes, e.g. differentially expressed genes. KPM is available as a feature-rich, user-friend- ly Cytoscape app, standalone software, or web service for de novo network enrichment (http://www.keypathwayminer. compbio.sdu.dk/). Clinical cancer research often focuses on patient survival times. Thus, we developed a new strategy to identify sub- networks most significantly associated with differences in survival. Our approach is based on the Network of Muta- tions Associated with Survival (NoMAS) algorithm that extracts subnetworks enriched in mutations. NoMAS exploits colour-coding to identify candidate subnetworks that are then evaluated with a log-rank test. We adapted NoMAS for multi-scale omics data by introducing a k-means clustering step to split patients into two groups using the candidate subnetworks molecular profile. Next, we apply a log-rank test to assess the significance of the difference in survival times between the two groups. Our overall goal is to find subnetworks significantly associated with survival time, thus creating multi-scale models that connect molecular changes, e.g. on the level of gene expression, to changes in the time-scale of patient survival. The identified subnetworks can be expected to represent important disease mechanisms, making them interesting candidates for further investigation. We thus expect that extending KPM to survival data will make de novo network enrichment considerably more attractive as a systems medicine approach.

A general methodology, which consists in deriving two-dimensional finite-difference schemes which involve numerical fluxes based on Dirichlet-to-Neumann maps (or Steklov-Poincare operators), is first recalled. Then, it is applied to several types of diffusion equations, some being weakly anisotropic, endowed with an external source. Standard finite-difference discretizations are systematically recovered, showing that in absence of any other mechanism, like e.g. convection and/or damping (which bring Bessel and/or Mathieu functions inside that type of numerical fluxes), these well-known schemes achieve a satisfying multi-dimensional character. (C) 2018 Published by Elsevier Ltd.

A genuinely two-dimensional discretization of general drift-diffusion (including incompressible Navier--Stokes) equations is proposed. Its numerical fluxes are derived by computing the radial derivatives of "bubbles" which are deduced from available discrete data by exploiting the stationary Dirichlet--Green function of the convection-diffusion operator. These fluxes are reminiscent of Scharfetter and Gummel's in the sense that they contain modified Bessel functions which allow one to pass smoothly from diffusive to drift-dominating regimes. For certain flows, monotonicity properties are established in the vanishing viscosity limit (``asymptotic monotony'') along with second-order accuracy when the grid is refined. Practical benchmarks are displayed to assess the feasibility of the scheme, including the "western currents" with a Navier--Stokes--Coriolis model of ocean circulation. Read More: https://epubs.siam.org/doi/10.1137/17M1151353

Modal decomposition techniques are used to analyse the wake field past a marine propeller achieved by previous numerical simulations (Muscari et al. Comput. Fluids, vol. 73, 2013, pp. 65-79). In particular, proper orthogonal decomposition (POD) and dynamic mode decomposition (DMD) are used to identify the most energetic modes and those that play a dominant role in the inception of the destabilization mechanisms. Two different operating conditions, representative of light and high loading conditions, are considered. The analysis shows a strong dependence of temporal and spatial scales of the process on the propeller loading and correlates the spatial shape of the modes and the temporal scales with the evolution and destabilization mechanisms of the wake past the propeller. At light loading condition, due to the stable evolution of the wake, both POD and DMD describe the flow field by the non-interacting evolution of the tip and hub vortex. The flow is mainly associated with the ordered convection of the tip vortex and the corresponding dominant modes, identified by both decompositions, are characterized by spatial wavelengths and frequencies related to the blade passing frequency and its multiples, whereas the dynamic of the hub vortex has a negligible contribution. At high loading condition, POD and DMD identify a marked separation of the flow field close to the propeller and in the far field, as a consequence of wake breakdown. The tonal modes are prevalent only near to the propeller, where the flow is stable; on the contrary, in the transition region a number of spatial and temporal scales appear. In particular, the phenomenon of destabilization of the wake, originated by the coupling of consecutive tip vortices, and the mechanisms of hub-tip vortex interaction and wake meandering are identified by both POD and DMD.

This presentation describes activities developed by CNR within EoCoE

Computing eigenvectors of graph Laplacian is a main computational kernel in data clustering, i.e., in identifying different groups such that data in the same group are similar and points in different groups are dissimilar with respect to a given notion of similarity. Data clustering can be reformulated in terms of a graph partitioning problem when the given set of data is represented as a graph, also known as similarity graph. In this context, eigenvectors of the graph Laplacian are used to obtain a new geometric representation of the original data set which generally enhances cluster properties and improves cluster detection. In this work we apply a bootstrap Algebraic MultiGrid (AMG) method to compute an approximation of the eigenvectors corresponding to small eigenvalues of the graph Laplacian and analyse their ability to catch clusters both in synthetic and in realistic graphs.

The Smoothed Particle Hydrodynamics (SPH)method is revisited within a Large Eddy Simulation (LES)perspective following the recent work of [1]. To this aim, LESfiltering procedure is recast in a Lagrangian framework bydefining a filter centred at the particle position that moves withthe filtered fluid velocity. The Lagrangian formulation of LES isthen used to re-interpret the SPH approximation of differentialoperators as a specific model based on the decomposition of theLES filter into a spatial and time filter.The derived equations represent a general LES-SPH schemeand contain terms that in part come from LES filtering and inpart derive from SPH kernels. The last ones lead to additionalterms (with respect to LES filtering) that contain fluctuations inspace, requiring adequate modelling. Further, since the adoptedLES filter differs from the classical Favre averaging for thedensity field, fluctuation terms also appear in the continuityequation.In the paper, a closure model for all the terms is suggested andsome simplifications with respect to the full LES-SPH model areproposed. The simplified LES model is formulated in a fashionsimilar to the diffusive SPH scheme of Molteni & Colagrossi[2] and the diffusive parameter is reinterpreted as a turbulentdiffusive coefficient, namely ? ? . In analogy with the turbulentkinetic viscosity ? T , the diffusive coefficient is modelled througha Smagorinsky-like model and both ? T and ? ? are assumed todepend on the magnitude of the local strain rate tensor D.Some examples of the simplified model are reported forboth 2D and 3D free-decaying homogeneous turbulence andcomparisons with the full LES-SPH model are provided.

The rapid development of cyber insurance market brings forward the question about the effect of cyber insurance on cyber security. Some researchers believe that the effect should be positive as organisations will be forced to maintain a high level of security in order to pay lower premiums. On the other hand, other researchers conduct a theoretical analysis and demonstrate that availability of cyber insurance may result in lower investments in security. In this paper we propose a mathematical analysis of a cyber-insurance model in a non-competitive market. We prove that with a right pricing strategy it is always possible to ensure that security investments are at least as high as without insurance. Our general theoretical analysis is confirmed by specific cases using CARA and CRRA utility functions.

Aiming to study the bubble cavitation problem in quiescent and sheared liquids, a third-order isothermal lattice Boltzmann model that describes a two-dimensional (2D) fluid obeying the van der Waals equation of state, is introduced. The evolution equations for the distribution functions in this off-lattice model with 16 velocities are solved using the corner-transport-upwind (CTU) numerical scheme on large square lattices (up to 6144 x 6144 nodes). The numerical viscosity and the regularization of the model are discussed for first- and second-order CTU schemes finding that the latter choice allows to obtain a very accurate phase diagram of a nonideal fluid. In a quiescent liquid, the present model allows us to recover the solution of the 2D Rayleigh-Plesset equation for a growing vapor bubble. In a sheared liquid, we investigated the evolution of the total bubble area, the bubble deformation, and the bubble tilt angle, for various values of the shear rate. A linear relation between the dimensionless deformation coefficient D and the capillary number Ca is found at small Ca but with a different factor than in equilibrium liquids. A nonlinear regime is observed for Ca greater than or similar to 0.2.