1,721,094 research outputs found
Nowcasting of urban air pollutants by neural networks
Full text linkThe modelling of urban air quality prediction is a difficult task because: i) the processes are controlled by complex chemical and physical mechanisms; ii) its state is ascertained by measuring too few parameters for a sufficient chemical picture; iii) sampling measurements are generally collected at too few points without consideration of scaling (for example CO is a local phenomenon while O-3 is regional and both are often measured by the same monitoring network); iv) balances of chemical species are often forced to work far from "local equilibria". In order to overcome these problems, Artificial Neural Networks (ANNs) were used here because they are model free and require very little knowledge about the underlying system structure. CO, NO2, and O-3 concentrations at the time (t + Deltat) are variables that depend on their previous concentrations and of other external information, such as meteorological data, solar radiation, chemical precursors or vehicle traffic information. ANNs used in this work were able to explain over 90% of the variability of the pollutant concentrations considered at the next hour (CO, NO2, and O-3) and over 80% of that of the next three-hour O-3 concentration. The forecasting of CO peaks exceeding a given value has been successfully performed by transforming original concentration time series into a probability series and processing the transformed data by an ANN. Sensitivity analysis has provided useful insight into the most important forecasting variables and their relevant links
A mathematical model for image saturation with an application to the restoration of solar images via adaptive sparse deconvolution
No full textIn this paper we introduce a mathematical model of the image saturation phenomenon occurring in a charged coupled device (CCD), and we propose a novel computational method for restoring saturated images acquired by the atmospheric imaging assembly (AIA) telescope. The mathematical model takes into account both primary saturation, when the photon-induced charge reaches the CCD full well capacity, and the blooming effect, when the excess charge flows into adjacent pixels. The restoration of AIA saturated images is then formulated as an inverse problem with a forward operator encoding the standard diffraction of light rays by a convolution, the primary saturation by an upper limit to the number of photons and the blooming effect by the conservation of the photon-induced charge spilled over adjacent pixels. As a result of this theoretical formulation we propose an adaptive l 1 regularized inversion method improving the desaturation capabilities of the existing SE-DESAT method [Guastavino S et al 2019 Astrophys. J. 882 109]. We prove that this method has the consistency estimation property also in the case that a fixed unknown background is considered. We test the adaptive method both in the case of synthetic and real data, comparing the performance with the one of the SE-DESAT method, showing that the proposed method avoids edge effects and artifacts in reconstructions even when the background solar activity is particularly intense
Approximation of discontinuous inverse operators with neural networks
No full textIn this work we deal with parametric inverse problems, which consist in recovering a finite number of parameters describing the structure of an unknown object, from indirect measurements. State-of-the-art methods for approximating a regularizing inverse operator by using a dataset of input-output pairs of the forward model rely on deep learning techniques. In these approaches, a neural network (NN) is trained to predict the value of the sought parameters directly from the data. In this paper, we show that these methods provide suboptimal results when a regularizing inverse operator is discontinuous with respect to the Euclidean topology. Hence, we propose a two-step strategy for approximating it by means of a NN, which works under general topological conditions. First, we embed the parameters into a subspace of a low-dimensional Euclidean space; second, we use a NN to approximate a homeomorphism between the subspace and the image of the parameter space through the forward operator. The parameters are then retrieved by applying the inverse of the embedding to the network predictions. The results are shown for the problem of x-ray imaging of solar flares with data from the Spectrometer/Telescope for Imaging X-rays. In this case, the parameter space is homeomorphic to a Moebius strip. Our simulation studies show that the use of a NN for predicting the parameters directly from the data yields systematic errors due to the non-Euclidean topology of the parameter space. The proposed strategy overcomes the discontinuity issues and furnishes stable and accurate reconstructions
Joint myopic deconvolution
No full textAstronomical image reconstruction is an inverse problem based on the knowledge of the point-spread function (PSF). However, this knowledge is often only partial, and a myopic deconvolution process is required for reaching the estimation of the solution. In this paper we propose a new statistical model which incorporates the presence of noise both on the image of the object to retrieve and on the 'measured' PSF. This technique also takes into account the nonnegativity constraint on the solution and on the PSF. Deconvolution results are presented for simulated data. A comparison between the classical algorithms and that proposed in this paper is given. This method can also be extended when different measures of PSF with different sizes are available
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Operational solar flare forecasting via video-based deep learning
Full text linkOperational flare forecasting aims at providing predictions that can be used to make decisions, typically on a daily scale, about the space weather impacts of flare occurrence. This study shows that video-based deep learning can be used for operational purposes when the training and validation sets used for network optimization are generated while accounting for the periodicity of the solar cycle. Specifically, this article describes an algorithm that can be applied to build up sets of active regions that are balanced according to the flare class rates associated to a specific cycle phase. These sets are used to train and validate a long-term recurrent convolutional network made of a combination of a convolutional neural network and a long short-term memory network. The reliability of this approach is assessed in the case of two prediction windows containing the solar storms of March 2015, June 2015, and September 2017
- …
