Searcharxiv⌕ Search

arXiv subjects

Isidro Gómez-Vargas

Publications and source records attributed to Isidro Gómez-Vargas.

16 recordsLinked to original sources
arXiv2026

Variational autoencoder for generating realistic $N$-body simulations for dark matter halos

In this paper, we present a deep-learning approach to generate synthetic cosmological images by training a convolutional variational autoencoder on two-dimensional dark matter density slices projected from $Λ$CDM $N$-body simulations. The model learns a compact latent representation that enables accurate reconstructions and fast generation of new synthetic realizations through a single forward pass through the decoder. We validate the generated fields using cosmology-based summary statistics, focusing on the matter power spectrum and related Fourier space diagnostics, and found good agreement with the reference simulation across the range of scales where the maps exhibit good resolution. Thanks to its low inference cost and stable training target, this variational-autoencoder approach provides a lightweight and reproducible basis for generative modeling of large-scale projected structures and can support downstream tasks such as fast simulation generation and data augmentation.

astro-ph.CO↗
arXiv2026

Modeling Doppler Shifts in Radial-Velocity Data with Deep Learning toward Earth-mass Exoplanet Detection

Detecting the tiny Doppler shifts induced by Earth-mass planets in stellar radial-velocity measurements remains extremely challenging due to stellar activity. Many deep-learning methods performing well on simulated data remain difficult to apply reliably on real stellar spectra. The aim of this work is to develop a deep-learning framework that generalizes to real, unseen spectra and improves the detectability of Earth-mass planets in radial-velocity data. We train artificial neural networks on HARPS-N solar spectra with injected planetary signals, using physics-motivated spectral representations based on flux and line-formation temperature, together with their velocity gradients. Two training strategies are explored: hold-out testing and cross-validation. Model robustness is enhanced through genetic-algorithm-based hyperparameter optimization, and predictive uncertainty is quantified using Monte Carlo dropout. Our most precise neural network model reliably retrieves, under the cross-validation strategy, the amplitudes, phases, and orbital periods of planetary signals with amplitudes greater than or equal to 25 cm/s and periods between 10 and 550 days. In addition, in all cases tested here, the successfully recovered signals correspond to the most significant peaks in the periodograms of the Doppler-shift predictions. Temperature-based spectral-shell representations consistently outperform flux-based shells. We also release doppleriann, a Python package implementing the proposed framework. Our results demonstrate that combining physically motivated spectral representations with deep learning provides a promising pathway toward the detection of Earth-mass planets in radial-velocity data from real observations, supported by a modeling framework that is both physically grounded and statistically rigorous, incorporating uncertainty quantification and optimized training strategies.

astro-ph.IM↗
arXiv2026

Evidence for evolving dark energy from DESI DR2 BAO and Pantheon$^+$, DES-Dovekie, and Union3

Evidences for evolving dark energy are shown using baryon acoustic oscillation measurements from the recent Dark Energy Spectroscopic Instrument Data Release 2 , combined with different Type Ia supernova datasets (Pantheon$^+$, DES-Dovekie, and Union3) and the CMB compressed likelihood. We examine several dark energy parameterizations, including the Logarithmic, Exponential, CPL, BA, JBP, Thawing, Mirage, and GEDE models. Analyzing the DESI DR2 measurements alone, we find that evidence for evolving dark energy is primarily driven by the LRG1-2 tracers, as their inclusion yields a preferred value of $w_0 > -1$. However, as each tracer provides only limited observables, this preference can result in an underconstrained and potentially unstable inference. Further, we find that each dark energy model predicts values in the $w_0 > -1$, $w_a < 0$ quadrant, a region characterized by the Quintom-B type dark energy scenario. The logarithmic bayes factor shows that, among all models, the Mirage model shows the inconclusive-to-moderate evidence across all dataset combinations. Consistently, the statistical significance remains modest, with $Nσ\sim 1.1$-$2.3$, and no model showing a robust preference for dynamical dark energy using late-time datasets alone. The evolution of $w(z)$ shows a phantom crossing around $z \sim 0.5$ in most dynamical dark energy models, and the evolution of $f_{\mathrm{DE}}(z)$ converges to $f_{\mathrm{DE}}(0) = 1$ in all dark energy models.

astro-ph.CO↗
arXiv2026

Data-driven modeling of rotation curves with artificial neural networks

Galactic rotation curves are crucial for understanding the distribution of mass in galaxies. Despite advances in precision observations, there are discrepancies between the inferred mass from luminosity and the observed rotational velocities, often attributed to dark matter. While traditional parametric models provide valuable insights, they struggle with complex galactic features like prominent bulges and non-circular motions. In this study, we apply artificial neural networks to generate robust, data-driven models, tailored to each galaxy, for the rotation curves of spiral galaxies using high-quality observational data. Our approach demonstrates that neural networks can effectively capture the intricate structure of rotation curves without relying on predefined astrophysical assumptions. By comparing the data-based models with the Navarro-Frenk-White model under two different assumptions for the stellar component, we classify galaxies based on the model that best fits their rotation curves, offering insights into the limitations and strengths of both theoretical and data-based methods. This work highlights the potential of machine learning techniques in identifying galaxies whose dynamics are not well captured by standard theoretical models, pointing to the need for more refined physical descriptions.

astro-ph.GA↗
arXiv2025

Equivalence of Dark Energy Models: A Theoretical and Bayesian Perspective

We explore the background equivalence among three dark energy models by constructing explicit mappings between dynamical dark energy (DDE), interacting dark energy (IDE), and running vacuum (RV). In our approach, the dark sector functions that characterize each model-such as the equation of state parameter $\bar{w}(a)$ for DDE, the interaction term $Q$ for IDE, and the functional form $Λ(H)$ for RV-are transformed into one another under specific assumptions. Extending previous work by von Marttens et al. (2020), we demonstrate that running vacuum models, characterized by $Λ(H) = a_0 + a_1 \dot{H} + a_2 H^2$, can be reinterpreted as an interacting dark energy model with $Q = 3Hγ\hatρ_c$, which in turn is equivalent to a dynamic dark energy model with an appropriately defined $\bar{w}(a)$. Using Bayesian analysis with Type Ia supernovae, Baryon Acoustic Oscillations, and Cosmic Chronometers, our observational constraints confirm that these theoretical equivalences hold at the background level. This study underscores the importance of seeking convergence in dark energy models, facilitating a better understanding of the dark sector.

astro-ph.CO↗
arXiv2024

Dark energy reconstruction analysis with artificial neural networks: Application on simulated Supernova Ia data from Rubin Observatory

In this paper, we present an analysis of Supernova Ia (SNIa) distance moduli $μ(z)$ and dark energy using an Artificial Neural Network (ANN) reconstruction based on LSST simulated three-year SNIa data. The ANNs employed in this study utilize genetic algorithms for hyperparameter tuning and Monte Carlo Dropout for predictions. Our ANN reconstruction architecture is capable of modeling both the distance moduli and their associated statistical errors given redshift values. We compare the performance of the ANN-based reconstruction with two theoretical dark energy models: $Λ$CDM and Chevallier-Linder-Polarski (CPL). Bayesian analysis is conducted for these theoretical models using the LSST simulations and compared with observations from Pantheon and Pantheon+ SNIa real data. We demonstrate that our model-independent ANN reconstruction is consistent with both theoretical models. Performance metrics and statistical tests reveal that the ANN produces distance modulus estimates that align well with the LSST dataset and exhibit only minor discrepancies with $Λ$CDM and CPL.

astro-ph.CO↗
arXiv2024

Deep Learning and genetic algorithms for cosmological Bayesian inference speed-up

In this paper, we present a novel approach to accelerate the Bayesian inference process, focusing specifically on the nested sampling algorithms. Bayesian inference plays a crucial role in cosmological parameter estimation, providing a robust framework for extracting theoretical insights from observational data. However, its computational demands can be substantial, primarily due to the need for numerous likelihood function evaluations. Our proposed method utilizes the power of deep learning, employing feedforward neural networks to approximate the likelihood function dynamically during the Bayesian inference process. Unlike traditional approaches, our method trains neural networks on-the-fly using the current set of live points as training data, without the need for pre-training. This flexibility enables adaptation to various theoretical models and datasets. We perform simple hyperparameter optimization using genetic algorithms to suggest initial neural network architectures for learning each likelihood function. Once sufficient accuracy is achieved, the neural network replaces the original likelihood function. The implementation integrates with nested sampling algorithms and has been thoroughly evaluated using both simple cosmological dark energy models and diverse observational datasets. Additionally, we explore the potential of genetic algorithms for generating initial live points within nested sampling inference, opening up new avenues for enhancing the efficiency and effectiveness of Bayesian inference methods.

astro-ph.IM↗
arXiv2024

Exploring the Evolution of Nonlinear Electrodynamics in the Universe: A Dynamical Systems Approach

This paper investigates the dynamics of cosmological models incorporating nonlinear electrodynamics (NLED), focusing on their stability and causality. We explore two specific NLED models: the Power-Law and Rational Lagrangians. We assess these models' viability in describing the universe's evolution using dynamical systems theory and Bayesian inference. We present the theoretical framework of NLED coupled with general relativity, followed by an analysis of the stability and causality through the squared sound speed of the NLED Lagrangians. We then conduct a detailed dynamical analysis to identify the universe's evolution with this matter content. Our results show that the Power-Law Lagrangian model transitions through various cosmological phases from a Maxwell radiation-dominated state to a matter-dominated state. For the Rational Lagrangian model, including the Maxwell term, stable and causal behavior is observed within specific parameter ranges, with critical points indicating the evolutionary pathways of the universe. To validate our theoretical findings, we perform Bayesian parameter estimation using a comprehensive set of observational data, including cosmic chronometers, Baryon Acoustic Oscillation (BAO) measurements, and Type Ia Supernovae (SNeIa). The estimated parameters for both models align with expected values for the current universe, particularly the matter density $Ω_m$ and the Hubble parameter $h$. However, the parameters $α$ and $b$ are not tightly constrained within the prior ranges. Our model comparison strongly favors the $Λ$CDM model over the NLED models for late-universe observations, as the NLED model does not exhibit a cosmological constant behavior. Our results highlight the need for further refinement and exploration of NLED-based cosmological models to fully integrate them into the standard cosmological framework.

astro-ph.CO↗
arXiv2024

Improving Earth-like planet detection in radial velocity using deep learning

Many novel methods have been proposed to mitigate stellar activity for exoplanet detection as the presence of stellar activity in radial velocity (RV) measurements is the current major limitation. Unlike traditional methods that model stellar activity in the RV domain, more methods are moving in the direction of disentangling stellar activity at the spectral level. The goal of this paper is to present a novel convolutional neural network-based algorithm that efficiently models stellar activity signals at the spectral level, enhancing the detection of Earth-like planets. We trained a convolutional neural network to build the correlation between the change in the spectral line profile and the corresponding RV, full width at half maximum (FWHM) and bisector span (BIS) values derived from the classical cross-correlation function. This algorithm has been tested on three intensively observed stars: Alpha Centauri B (HD128621), Tau ceti (HD10700), and the Sun. By injecting simulated planetary signals at the spectral level, we demonstrate that our machine learning algorithm can achieve, for HD128621 and HD10700, a detection threshold of 0.5 m/s in semi-amplitude for planets with periods ranging from 10 to 300 days. This threshold would correspond to the detection of a $\sim$4$\mathrm{M}_{\oplus}$ in the habitable zone of those stars. On the HARPS-N solar dataset, our algorithm is even more efficient at mitigating stellar activity signals and can reach a threshold of 0.2 m/s, which would correspond to a 2.2$\mathrm{M}_{\oplus}$ planet on the orbit of the Earth. To the best of our knowledge, it is the first time that such low detection thresholds are reported for the Sun, but also for other stars, and therefore this highlights the efficiency of our convolutional neural network-based algorithm at mitigating stellar activity in RV measurements.

astro-ph.EP↗
arXiv2024

Coupled Multi Scalar Field Dark Energy

The main aim of this paper is to present the multi scalar field components as candidates to be the dark energy of the universe and their observational constraints. We start with the canonical Quintessence and Phantom fields with quadratic potentials and show that a more complex model should bear in mind to satisfy current cosmological observations. Then we present some implications for a combination of two fields, named as Quintom models. We consider two types of models, one as the sum of the quintessence and phantom potentials and other including an interacting term between fields. We find that adding one extra degree of freedom, by the interacting term, the dynamics enriches considerably and could lead to an improvement in the fit of $-2\lnΔ\Like_{\rm max}= 5.19$, compared to $Λ$CDM. The resultant effective equation of state is now able to cross the phantom divide line, and in several cases present an oscillatory or discontinuous behavior, depending on the interaction value. The parameter constraints of the scalar field models (quintessence, phantom, quintom and interacting quintom) were performed using Cosmic Chronometers, Supernovae Ia and Baryon Acoustic Oscillations data; and the Log-Bayes factors were computed to compare the performance of the models. We show that single scalar fields may face serious troubles and hence the necessity of a more complex models, i.e. multiple fields.

astro-ph.CO↗
arXiv2024

Cosmological parameter estimation with Genetic Algorithms

Genetic algorithms are a powerful tool in optimization for single and multi-modal functions. This paper provides an overview of their fundamentals with some analytical examples. In addition, we explore how they can be used as a parameter estimation tool in cosmological models to maximize the likelihood function, complementing the analysis with the traditional Markov Chain Monte Carlo methods. We analyze that genetic algorithms provide fast estimates by focusing on maximizing the likelihood function, although they cannot provide confidence regions with the same statistical meaning as Bayesian approaches. Moreover, we show that implementing sharing and niching techniques ensures an effective exploration of the parameter space, even in the presence of local optima, always helping to find the global optima. This approach is invaluable in the cosmological context, where exhaustive space exploration of parameters is essential. We use dark energy models to exemplify the use of genetic algorithms in cosmological parameter estimation, including a multimodal problem, and we also show how to use the output of a genetic algorithm to obtain derived cosmological functions. This paper concludes that genetic algorithms are a handy tool within cosmological data analysis, without replacing the traditional Bayesian methods but providing different advantages.

astro-ph.CO↗
arXiv2023

Analysis of Dark Matter Halo Structure Formation in $N$-body Simulations with Machine Learning

The properties of the matter density field in the initial conditions have a decisive impact on the features of the large-scale structure of the Universe as observed today. These need to be studied via $N$-body simulations, which are imperative to analyze high density collapsed regions into dark matter halos. In this paper, we train Machine Learning algorithms with information from N -body simulations to infer two properties: dark matter particle halo classification that leads to halo formation prediction with the characteristics of the matter density field traced back to the initial conditions, and dark matter halo formation by calculating the Halo Mass Function (HMF), which offers the number density of dark matter halos with a given threshold. We map the initial conditions of the matter density field into classification labels of dark matter halo structures. The Halo Mass Function of the simulations is calculated and reconstructed with theoretical methods as well as our trained algorithms. We test several Machine Learning techniques where we could find that the Random Forest and Neural Networks proved to be the better performing tools to classify dark matter particles in cosmological simulations. We also show that that it is not compulsory to use a high amount of data to train the algorithms in order to reconstruct the HMF, giving us a very good fitting function for both simulation and theoretical results.

astro-ph.CO↗
arXiv2023

Neural network reconstructions for the Hubble parameter, growth rate and distance modulus

This paper introduces a new approach to reconstruct cosmological functions using artificial neural networks based on observational measurements with minimal theoretical and statistical assumptions. By using neural networks, we can generate computational models of observational datasets, and then we compare them with the original ones to verify the consistency of our method. This methodology is applicable to even small-size datasets. In particular, we test the proposed method with data coming from cosmic chronometers, $fσ_8$ measurements, and the distance modulus of the Type Ia supernovae. Furthermore, we introduce a first approach to generate synthetic covariance matrices through a variational autoencoder, using the systematic covariance matrix of the Type Ia supernova compilation.

astro-ph.CO↗
arXiv2023

Neural Networks Optimized by Genetic Algorithms in Cosmology

The applications of artificial neural networks in the cosmological field have shone successfully during the past decade, this is due to their great ability of modeling large amounts of datasets and complex nonlinear functions. However, in some cases, their use still remains controversial because their ease of producing inaccurate results when the hyperparameters are not carefully selected. In this paper, to find the optimal combination of hyperparameters to artificial neural networks, we propose to take advantage of the genetic algorithms. As a proof of the concept, we analyze three different cosmological cases to test the performance of the architectures achieved with the genetic algorithms and compare them with the standard process, consisting of a grid with all possible configurations. First, we carry out a model-independent reconstruction of the distance modulus using a type Ia supernovae compilation. Second, the neural networks learn to infer the equation of state for the quintessence model, and finally with the data from a combined redshift catalog the neural networks predict the photometric redshift given six photometric bands (urgizy). We found that the genetic algorithms improve considerably the generation of the neural network architectures, which can ensure more confidence in their physical results because of the better performance in the metrics with respect to the grid method.

astro-ph.IM↗
arXiv2022

Observational cosmology with Artificial Neural Networks

In cosmology, the analysis of observational evidence is very important to test theoretical models of the Universe. Artificial neural networks are powerful and versatile computational tools for data modelling and are recently being considered in the analysis of cosmological data. The main goal of this paper is to provide an introduction to artificial neural networks and to describe some applications to cosmology. We present an overview on the fundamentals of neural networks and their technical details. Throughout three examples, we show their capabilities in modelling cosmological data, saving computational time in numerical tasks, and classifying stellar objects. Artificial neural networks offer interesting qualities that make them a viable alternative method for data analysis in cosmological research.

astro-ph.CO↗