Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Jorge Amaya has a total of 137 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Prediction of Adverse effects of Geomagnetic storms and Energetic Radiation (PAGER) project

    by Yuri Y. Shprits, Tony Arber, Michael Liemohn, OndřEj Santolík, Keith Bennett, Stefano Bianco, Melanie Burns, Angelica Castillo, Marina García Peñaranda, Bernhard Haas, et al.

  • Open Access

    Bypassing the static input size of neural networks in flare forecasting by using spatial pyramid pooling

    by P. Vong, L. R. Dolla, A. Koukras, J. Gustin, J. Amaya, E. Dineva, G. Lapenta

    Context. The spatial extension of active regions of the Sun (hence their associated images) can strongly vary from one case to the next. This inhomogeneity is a problem when using convolutional neural networks (CNNs) to study solar flares, as they generally use input images of a fixed size. Different processes can be performed to retrieve a database with homogeneous-sized images, such as coarse resizing, cropping, or padding of raw images. Unfortunately, key features can be lost or distorted beyond recognition during these processes. This can lead to a deterioration of the ability of CNNs to classify flares of different soft X-ray classes, especially those from active regions with structures of great complexity. Aims. This study aims to implement and test a CNN architecture that retains features of characteristic scales as fine as the original resolution of the input images. Methods. We compared the performance of two CNN architectures for solar flare prediction. The first one is a traditional CNN with convolution layers, batch normalization layers, max pooling layers, and resized input, whereas the other implements a spatial pyramid pooling (SPP) layer instead of a max pooling layer before the flattening layer and without any input resizing. Both were trained on the Spaceweather HMI Active Region Patch (SHARP) line of sight magnetogram database, which was generated from data collected by the Helioseismic and Magnetic Imager on board the Solar Dynamics Observatory from May 2010 to August 2021 only using images within 45°of the central meridian of the Sun. We also studied two cases of binary classification. In the first case, our model had to distinguish active regions producing flares in less than 24 h of class ≥C1.0 from active regions producing flares in more than 24 h or never. In the second case, it had to distinguish active regions producing flares in less than 24 h of class ≥M1.0 from active regions producing flares in more than 24 h or never, or flares in less than 24 h but of class < M1.0. The impact of the use of a score-oriented loss (SOL) function optimizing the true skill statistics (TSS) metric instead of a binary cross-entropy (BCE) loss function is also studied and discussed in this work. Results. Our models implementing an SPP layer and trained using a BCE loss function outperform the traditional CNN models, with an average increase of 0.1 in TSS and 0.17 in precision metrics when predicting flares ≥C1.0 within 24 h. However, their performances degrade sharply along the other models studied in this paper when trained to classify images of ≥M1.0 flares. Conclusions. We prove the higher efficiency of a CNN model that includes an SPP layer in predicting solar flares. The degradation of prediction performance of this model when the images of active regions producing a C class flare are classified as negative may be attributed to its success in identifying features that appear in active regions only a few hours before the flare, independent of their soft X-ray class. The development of explainable artificial intelligence tools adapted to this architecture in future projects will be interesting for the study of solar flare-triggering mechanisms.


  • Open Access

    Addressing the closure problem using supervised Machine Learning

    by Sophia Köhne, Brecht Laperre, Jorge Amaya, Sara Jamal, Simon Lautenbach, Rainer Grauer, Giovanni Lapenta, Maria Elena Innocenti

    When deriving fluid equations from the Vlasov equation for collisionless plasmas, one runs into the so-called closure problem: each equation for the temporal evolution of one particle moment (density, current, pressure, heat flux, …) includes terms depending on the next order moment. Therefore, when choosing to truncate the description at the nth order, one must approximate the terms related to the (n+1)th order moment included in the evolution equation for the nth order moment. The order at which the hierarchy is closed and the assumption behind the approximations used determine how accurately a fluid description can reproduce kinetic processes.In this work, we aim at reconstructing specific particle moments from kinetic simulations, using as input the electric and magnetic field and the lower moments. We use fully kinetic Particle In Cell simulations, where all physical information is available, as the ground truth. The approach we present here uses supervised machine learning to enable a neural network to learn how to reconstruct higher moments from lower moments and fields.Starting from the work of Laperre et al., 2022 we built a framework which makes it possible to train feedforward multilayer perceptrons on kinetic simulations to learn to predict the higher moments of the Vlasov equation from the lower moments, which would also be available in fluid simulations. We train on simulations of magnetic reconnection in a double Harris current sheet with varying background guide field obtained with the semi-implicit Particle-in-Cell code iPiC3D (Markidis et al, 2010). We test the influence of data preprocessing techniques, of (hyper-)parameter variations and of different architectures of the neural networks on the quality of the predictions that are produced. Furthermore, we investigate which metrics are most useful to evaluate the quality of the outcome.

  • Open Access

    An analysis of the European space weather modelling capabilities

    by Jorge Amaya

    For many years the Space Weather Office of the European Space Agency has promoted the development of resources for the monitoring of solar activity and its effects on our planet and technology. Current and past activities include the gathering and distribution of data from ground and space observatories, the deployment of specialized instruments onboard hosting satellites, the proposal of new nanosat missions, the promotion of improvements in computer models, and the design of Vigil, the first operational space weather mission in Europe. Multiple international institutions have also developed their own space weather assets. However, in contrast to meteorological forecasts, no prediction centre is still capable to produce accurate forecasts of the space weather conditions in the next days. Private and public institutions around the world are working on the improvement of forecasting tools. For ESA this effort requires additional support on the development of critical modelling techniques, including: data assimilation, uncertainty quantification, ensemble forecasting, machine learning, and model coupling, among others.   To have a clear mapping between the existing modelling capabilities in Europe and the user requirements, we invited European experts to participate in the "Space Weather Modelling Workshop" held in March 2023 at ESOC. In this presentation we will show the results of this workshop, including a detailed summary of the European modelling capabilities, a description of the existing gaps and overlaps, an analysis of missing assets, a description of existing bottlenecks, and a plan to improve the accuracy of space weather forecasting. The ESA Space Weather Office has set goals for the next years in the area of modelling, based on the data collected during the workshop, on discussions with the community, and on data from the literature. We will describe how the community can contribute to reach these goals and to develop the new generation of European Space Weather forecasting services.

  • Open Access

    Investigation of Machine Learning Techniques for Disruption Prediction Using JET Data

    by Joost Croonen, Jorge Amaya, Giovanni Lapenta

    Disruption prediction and mitigation is of key importance in the development of sustainable tokamak reactors. Machine learning has become a key tool in this endeavour. In this paper, multiple machine learning models are tested and compared. A focus has been placed on the analysis of a transition to dimensionless input quantities. The methods used in this paper are the support vector machine, two-tiered support vector machine, random forest, gradient-boosted trees and long-short term memory. The performance between different models is remarkably similar, with the support vector machine attaining a slightly better accuracy score. The similarity could indicate issues with the dataset, but further study is required to confirm this. Both the two-tiered model and long-short term memory performed below expectations. The former could be attributed to an implementation which did not allow error propagation between tiers. The latter could be attributed to high noise and low frequency of the input signals. Dimensionless models experienced an expected decrease in performance, caused by a loss of information in the conversion. However, random forest and gradient boosted trees experienced a significantly lower decrease, making them more suitable for dimensionless predictors. From the disruption detection times, it was concluded that several disruptions could be predicted at more than 600 ms in advance. A feature importance study using the random forest indicated the negative impact of high noise and missing data in the database, suggesting improvements in data preparation for future work and the potential reevaluation of some of the selected portable features due to poor performance.


  • Open Access

    A versatile exploration method for simulated data based on Self Organizing Maps

    by Maria Elena Innocenti, Sophia Köhne, Simon Hornisch, Rainer Grauer, Jorge Amaya, Jimmy Raeder, Banafsheh Ferdousi, James Andy Edmond, Giovanni Lapenta

    The large amount of data produced by measurements and simulations of space plasmas has made it fertile ground for the application of classification methods, that can support the scientist in preliminary data analysis. Among the different classification methods available, Self Organizing Maps, SOMs [Kohonen, 1982] offer the distinct advantage of producing an ordered, lower-dimensional representation of the input data that preserves their topographical relations. The 2D map obtained after training can then be explored to gather knowledge on the data it represents. The distance between nodes reflects the distance between the input data: one can then further cluster the map nodes to identify large scale regions in the data where plasma properties are expected to be similar.In this work, we train SOMs using data from different simulations of different aspects of the heliospheric environment: a global magnetospheric simulation done with the OpenGGCM-CTIM-RCM code, a Particle In Cell simulation of plasmoid instability done with the semi-implicit code ECSIM, a fully kinetic simulation of single X point reconnection done with the Vlasov code implemented in MuPhy2.We examine the SOM feature maps, unified distance matrix and SOM node weights to unlock information on the input data. We then classify the nodes of the different SOMs into a lower and automatically selected number of clusters, and we obtain, in all three cases, clusters that map well to our a priori knowledge on the three systems. Results for the magnetospheric simulations are described in Innocenti et al, 2021. This classification strategy then emerges as a useful, relatively cheap and versatile technique for the analysis of simulation, and possibly observational, plasma physics data.Innocenti, M. E., Amaya, J., Raeder, J., Dupuis, R., Ferdousi, B., & Lapenta, G. (2021). Unsupervised classification of simulated magnetospheric regions. Annales Geophysicae Discussions, 1-28. https://angeo.copernicus.org/articles/39/861/2021/angeo-39-861-2021.pdf

  • Identification of high order closure terms from fully kinetic simulations using machine learning

    by B. Laperre, J. Amaya, S. Jamal, G. Lapenta

    Simulations of large-scale plasma systems are typically based on a fluid approximation approach. These models construct a moment-based system of equations that approximate the particle-based physics as a fluid, but as a result, they lack the small-scale physical processes available to fully kinetic models. Traditionally, empirical closure relations are used to close the moment-based system of equations, which typically approximate the pressure tensor or heat flux. The more accurate the closure relation, the stronger the simulation approaches kinetic-based results. In this paper, new closure terms are constructed using machine learning techniques. Two different machine learning models, a multi-layer perceptron and a gradient boosting regressor, synthesize a local closure relation for the pressure tensor and heat flux vector from fully kinetic simulations of a 2D magnetic reconnection problem. The models are compared to an existing closure relation for the pressure tensor, and the applicability of the models is discussed. The initial results show that the models can capture the diagonal components of the pressure tensor accurately and show promising results for the heat flux, opening the way for new experiments in multi-scale modeling. We find that the sampling of the points used to train both models plays a capital role in their accuracy.


  • Open Access

    Parametrization of Heliophysical Data Using Autoencoders

    by Jorge Amaya, Sara Jamal, Brecht Laperre, Giovanni Lapenta

    One of the most important steps in any AI/ML application is the pre-processing of the data. The objective of this step is to project the original data in a new basis, or in a new latent space, where the different features of the problem are comparable and where their distribution covers a large range of values. Using the data in its natural basis can lead to under-performing AI/ML models. While almost all papers in our domain are careful to normalize or standardize the data, it is less frequent to see the use of simple linear PCA transformations, and even less frequent the use of more complex non-linear projections in latent spaces. Here we show how our research team is using autoencoder neural networks to perform non-linear transformations of images, simulations and time-series used in heliophyisical applications. Autoencoder transformations allow to parametrize any type of data by projecting it onto a latent space of higher or lower dimension. In these latent spaces the transformed data commonly presents better statistical properties allowing improvements in the AI/ML modeling. In addition, autoencoders are also known as generative techniques, i.e. they can be used to produce “artificial” or “synthetic” data. We will present three particular examples of the use of autoencoders: 1) parametrization of solar wind observations using standard feed forward autoencoders, 2) parametrization of magnetosphere simulations using convolutional autoencoders, and 3) parametrization and generation of solar active regions using variational convolutional autoencoders. We will show how these parametrizations can then be used for AI/ML classification and forecasting. This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 776262 (AIDA).

  • Open Access

    Unsupervised classification of simulated magnetospheric regions

    by Maria Elena Innocenti, Jorge Amaya, Joachim Raeder, Romain Dupuis, Banafsheh Ferdousi, Giovanni Lapenta

    In magnetospheric missions, burst-mode data sampling should be triggered in the presence of processes of scientific or operational interest. We present an unsupervised classification method for magnetospheric regions that could constitute the first step of a multistep method for the automatic identification of magnetospheric processes of interest. Our method is based on self-organizing maps (SOMs), and we test it preliminarily on data points from global magnetospheric simulations obtained with the OpenGGCM-CTIM-RCM code. The dimensionality of the data is reduced with principal component analysis before classification. The classification relies exclusively on local plasma properties at the selected data points, without information on their neighborhood or on their temporal evolution. We classify the SOM nodes into an automatically selected number of classes, and we obtain clusters that map to well-defined magnetospheric regions. We validate our classification results by plotting the classified data in the simulated space and by comparing with k-means classification. For the sake of result interpretability, we examine the SOM feature maps (magnetospheric variables are called features in the context of classification), and we use them to unlock information on the clusters. We repeat the classification experiments using different sets of features, we quantitatively compare different classification results, and we obtain insights on which magnetospheric variables make more effective features for unsupervised classification.


  • Open Access

    Reply on RC2

    by Maria Elena Innocenti, Jorge Amaya, Joachim Raeder, Romain Dupuis, Banafsheh Ferdousi, Giovanni Lapenta

    In magnetospheric missions, burst mode data sampling should be triggered in the presence of processes of scientific or opera- tional interest. We present an unsupervised classification method for magnetospheric regions, that could constitute the first-step of a multi-step method for the automatic identification of magnetospheric processes of interest. Our method is based on Self Organizing Maps (SOMs), and we test it preliminarily on data points from global magnetospheric simulations obtained with the OpenGGCM-CTIM-RCM code. The classification relies exclusively on local plasma properties at the selected data points, without information on their neighborhood or on their temporal evolution. We classify the SOM nodes into an automatically selected number of classes, and we obtain clusters that map to well defined magnetospheric regions. For the sake of result interpretability, we examine the SOM feature maps (magnetospheric variables are called features in the context of classification), and we use them to unlock information on the clusters. We repeat the classification experiments using different sets of features, and we obtain insights on which magnetospheric variables make more effective features for unsupervised classification.