Searcharxiv⌕ Search

arXiv subjects

Martín de los Rios

Publications and source records attributed to Martín de los Rios.

17 recordsLinked to original sources
arXiv2026

Reconstructing orbits of galaxies in extreme regions (roger v2.0): an extension to intermediate mass systems

In this paper, we present an updated version of the roger code, called roger v2.0, developed to perform the orbital classification of galaxies residing in and around galaxy groups and clusters. In addition to the projected phase-space coordinates, the new version incorporates the host halo mass as an additional input parameter. Although the inclusion of the halo mass leads to only modest changes in the classification, it contributes to improving the overall robustness of the method. We also extend the range of host halo masses over which roger can be applied, enabling the analysis of systems with masses down to $10^{13.5} h^{-1} M_{\odot}$. We further provide a Python implementation of the new code, which will be made publicly available. This implementation enables users to efficiently and robustly classify arbitrary galaxy samples using either version of the method. Moreover, it allows users to train a customized classifier on an alternative training set, providing the flexibility to adapt the method to different datasets and scientific applications.

astro-ph.GA↗
arXiv2026

Linking X-ray emission to galaxy populations in GAMA groups

We study how the presence of a detectable hot intragroup medium affects galaxy colour and star-formation properties, and how these trends depend on dynamical stage within the group. We analysed galaxies in GAMA groups with and without X-ray detection (XG and NXG) at 0.1 <= z <= 0.2, using samples with similar halo-mass distributions, compared with a field control sample. Galaxies were classified as red, green, or blue by UV-optical colour, and group galaxies further classified into dynamical classes based on projected phase-space position and host halo mass. We examined galaxy populations as a function of orbital class, star-formation activity, nuclear activity, and cluster-centric distance. The galaxy population is systematically more evolved in XG than NXG, with lower blue and star-forming fractions and higher red and passive fractions. The green fraction is stable across environments and orbital classes, though high-mass recent arrivals show a mild excess of green systems while retaining relatively high star-forming and low passive fractions, suggesting an intermediate quenching stage. The active galactic nucleus fraction shows no environmental dependence, while the post-starburst fraction increases strongly in XG, reaching nearly 20 percent, consistent with enhanced rapid quenching. Galaxy properties show stronger radial segregation in XG than NXG, with colour fractions differing from field values even at R/R200 ~ 3, consistent with preprocessing. Despite similar halo-mass distributions, the systematic differences between XG and NXG show that halo mass alone does not determine galaxy properties. A detectable hot intragroup medium is associated with enhanced environmental processing, stronger quenching, and more frequent rapid transitions, highlighting the role of the intragroup medium's thermodynamical state, beyond halo mass, in shaping galaxy evolution at group scales.

astro-ph.GA↗
arXiv2026

Two-Stage Machine Learning Strategy for Scalar and Vector Leptoquark Discrimination at the LHC

We present a machine learning framework for the characterization of leptoquark (LQ) signals at the Large Hadron Collider, focusing on the discrimination between scalar (SLQ) and vector (VLQ) hypotheses. The method is based on a two-stage inference pipeline that combines a classifier trained to separate Standard Model backgrounds from a mixed LQ signal with a second classifier designed to distinguish between SLQ and VLQ scenarios, using the signal yield inferred from the first-stage classifier to guide the corresponding scalar and vector mass hypotheses. A test statistic is constructed from the classifier outputs and interpreted using reference probability density functions. The approach is applied to realistic LHC final states with hadronically decaying tau leptons, multiple jets, and missing transverse momentum, and its performance is assessed using simulated pseudo-experiments. We show that the proposed strategy provides a robust and statistically consistent procedure to discriminate between SLQ and VLQ signals across a wide range of masses and signal stregths. We find that the spin identification power closely follows the discovery potential, demonstrating that determining the spin nature of a newly discovered LQ does not require substantially larger datasets than those needed for discovery itself.

hep-ph↗
arXiv2026

Dark Matter profiles of "in silico" galaxies: deep learning inference

Machine learning has the potential to improve the reconstruction of the dark matter profile of galaxies with respect to traditional methods, like rotation curves. We demonstrate on the simulation suite Illustris-TNG that a steerable equivariant convolutional neural network (CNN) is able to infer the dark matter profiles within and around individual galaxies from photometric and interferometric data, improving on a standard CNN. Within the in silico environment of the simulations, our architecture is able to capture the dark matter distribution within galaxies without a parametrization of the profile. We perform an interpretability analysis to understand the internal mechanisms of the trained model and the most important data features used to estimate the dark matter profiles. The equivariant CNN recovers the dark matter profile of galaxies within the stellar mass range $[10^{10} - 10^{12} ]$ $M_{\odot}$ with excellent precision and accuracy: the mean squared error is reduced by a factor of ~ 3 from its value under the training distribution, demonstrating that the network has learnt from the data features. While this holds within the controlled 'in silico' environment of the simulation, we argue that few additional steps are needed before this method can be reliably applied to galaxies in the real field observations.

astro-ph.GA↗
arXiv2026

Shedding Light on Dark Matter at the LHC with Machine Learning

We investigate a WIMP dark matter (DM) candidate in the form of a singlino-dominated lightest supersymmetric particle (LSP) within the $Z_3$-symmetric Next-to-Minimal Supersymmetric Standard Model (NMSSM). This framework gives rise to regions of parameter space where DM is obtained via co-annihilation with nearby higgsino-like electroweakinos and DM direct detection~signals are suppressed, the so-called ``blind spots''. On the other hand, collider signatures remain promising due to enhanced radiative decay modes of higgsinos into the singlino-dominated LSP and photons, rather than into leptons or hadrons. Compared to MSSM scenarios with light bino- and wino-like electroweakinos, the NMSSM allows for final states with multiple photons arising from cascade radiative decays, providing a distinctive collider signature. This motivates searches for radiatively decaying neutralinos, however, these signals face substantial background challenges, as the decay products are typically soft due to the small mass-splits ($Δm$) between the LSP and the higgsino-like coannihilation partners. We apply a data-driven Machine Learning (ML) analysis that improves sensitivity to these subtle signals, offering a powerful complement to traditional search strategies to discover a new physics scenario. Using an LHC integrated luminosity of $100~\mathrm{fb}^{-1}$ at $14~\mathrm{TeV}$, the method achieves a $5σ$ discovery reach for higgsino masses up to $225~\mathrm{GeV}$ with $Δm\!\lesssim\!12~\mathrm{GeV}$, and a $2σ$ exclusion up to $285~\mathrm{GeV}$ with $Δm\!\lesssim\!20~\mathrm{GeV}$. These results highlight~the power of collider searches to probe DM candidates that remain hidden from current~direct detection experiments, and provide a motivation for a search by the LHC collaborations using ML methods.

hep-ph↗
arXiv2026

Rotary Masked Autoencoders are Versatile Learners

Applying Transformers to irregular time-series typically requires specializations to their baseline architecture, which can result in additional computational overhead and increased method complexity. We present the Rotary Masked Autoencoder (RoMAE), which utilizes the popular Rotary Positional Embedding (RoPE) method for continuous positions. RoMAE is an extension to the Masked Autoencoder (MAE) that enables interpolation and representation learning with multidimensional continuous positional information while avoiding any time-series-specific architectural specializations. We showcase RoMAE's performance on a variety of modalities including irregular and multivariate time-series, images, and audio, demonstrating that RoMAE surpasses specialized time-series architectures on difficult datasets such as the DESC ELAsTiCC Challenge while maintaining MAE's usual performance across other modalities. In addition, we investigate RoMAE's ability to reconstruct the embedded continuous positions, demonstrating that including learned embeddings in the input sequence breaks RoPE's relative position property.

cs.LG↗
arXiv2025

Machine-Learning Analysis of Radiative Decays to Dark Matter at the LHC

The search for weakly interacting matter particles (WIMPs) is one of the main objectives of the High Luminosity Large Hadron Collider (HL-LHC). In this work we use Machine-Learning (ML) techniques to explore WIMP radiative decays into a Dark Matter (DM) candidate in a supersymmetric framework. The minimal supersymmetric WIMP sector includes the lightest neutralino that can provide the observed DM relic density through its co-annihilation with the second lightest neutralino and lightest chargino. Moreover, the direct DM detection cross section rates fulfill current experimental bounds and provide discovery targets for the same region of model parameters in which the radiative decay of the second lightest neutralino into a photon and the lightest neutralino is enhanced. This strongly motivates the search for radiatively decaying neutralinos which, however, suffers from strong backgrounds. We investigate the LHC reach in the search for these radiatively decaying particles by means of cut-based and ML methods and estimate its discovery potential in this well-motivated, new physics scenario. We demonstrate that using ML techniques would enable access to most of the parameter space unexplored by other searches.

hep-ph↗
arXiv2025

Galaxy evolution in groups: Transition galaxies in the IllustrisTNG simulations

The evolution of galaxies is significantly influenced by the environments they inhabit. While high-density regions, such as clusters of galaxies have been widely studied, the dynamics and quenching processes in intermediate environments remain less explored. These systems provide a valuable context for understanding the transition of galaxies from active star formation to quiescence. This study aims to characterise the astrophysical properties of galaxies in intermediate-mass galaxy groups $(13.5 \leq \log(M_{200}/M_{\odot}) \leq 13.7)$, with a focus on their evolutionary pathways and the key processes driving their transition through the green valley (GV) and green zone (GZ). Using the Illustris TNG300-1 hydrodynamical cosmological simulations, we classified galaxies based on their trajectories and environment into five categories: group galaxies (GRs), backsplash galaxies (BSs), recent infallers (RINs), infall galaxies (INs), and field galaxies (FGs). We examined their optical colours in the $(u-r)$-stellar mass diagram, specific star formation rates (sSFRs), gas fractions, and stellar mass evolution from $z=0.5$ to $z=0$. At $z=0$, FGs dominate the blue cloud, while GRs show progressive reddening. BSs exhibit the highest fraction of green galaxies, highlighting their transitional nature, whereas RINs show a rapid quenching upon entering $R_{200}$. Our results reveal that the timing of group entry and environmental effects, such as gas depletion, play a critical role in galaxy quenching. Green BS and RINs follow distinct evolutionary tracks, with the latter undergoing more rapid changes due to later infall into more massive systems.

astro-ph.GA↗
arXiv2025

Reconstructing orbits of galaxies in extreme regions (ROGER). IV. Unveiling galaxy evolution patterns in OmegaWINGS clusters

Clusters of galaxies have proven to be efficient systems in modifying various properties of galaxies, such as star formation or morphology. However, projection effects impose serious challenges in determining how, when, and to what extent galaxies are affected by the cluster environment. Using innovative techniques to classify galaxies based on their history within the cluster, we aim to determine how galaxies of different classes are affected by the cluster environment. We applied the ROGER code to select trajectories of galaxies in the phase space for 35 galaxy clusters from the OmegaWINGS survey. A new algorithm was applied to minimize contamination effects. We found that both morphological transformation and the quenching of star formation begin shortly after galaxies enter the cluster. Even though over the last $2-3$ Gyr, galaxies entering clusters have undergone significant transformations in both their star formation and morphology these transformation processes are not complete, that is, they are not completely quenched and are not early type yet. Backsplash galaxies and recent infallers show a higher fraction of jellyfish galaxies compared to older cluster members, suggesting that the timescale of this phenomenon is typically less than 3 Gyr.

astro-ph.GA↗
arXiv2025

Improving the accuracy of observable distributions for galaxies classified in the Projected Phase Space Diagram

Studies of galaxy populations classified according to their kinematic behaviours and dynamical state using the Projected Phase Space Diagram (PPSD) are affected by misclassification and contamination, leading to systematic errors in determining the characteristics of the different galaxy classes. We propose a method to statistically correct the determination of galaxy properties' distributions accounting for the contamination caused by misclassified galaxies from other classes. Using a sample of massive clusters and galaxies in their surroundings taken from the MultiDark Planck 2 simulation combined with the semi-analytic model of galaxy formation SAG, we compute the confusion matrix associated to a classification scheme in the PPSD. Based on positions in the PPSD, galaxies are classified as cluster members, backsplash galaxies, recent infallers, infalling galaxies, and interlopers. This classification is determined using probabilities calculated by the code ROGER, along with a threshold criterion. By inverting the confusion matrix, we are able to get better determinations of distributions of galaxy properties such as colour. Compared to a direct estimation based solely on the predicted galaxy classes, our method provides better estimates of the mass-dependent colour distribution for the galaxy classes most affected by misclassification: cluster members, backsplash galaxies, and recent infallers. We apply the method to a sample of observed X-ray clusters and galaxies. Our method can be applied to any classification of galaxies in the PPSD, and to any other galaxy property besides colour, provided an estimation of the confusion matrix. Blue, low-mass galaxies in clusters are almost exclusively recent infaller galaxies that have not yet been quenched by the environmental action of the cluster. Backsplash galaxies are on average redder than expected.

astro-ph.GA↗
arXiv2023

Measuring the Sterile Neutrino Mass in Spallation Source and Direct Detection Experiments

We explore the complementarity of direct detection (DD) and spallation source (SS) experiments for the study of sterile neutrino physics. We focus on the sterile baryonic neutrino model: an extension of the Standard Model that introduces a massive sterile neutrino with couplings to the quark sector via a new gauge boson. In this scenario, the inelastic scattering of an active neutrino with the target material in both DD and SS experiments gives rise to a characteristic nuclear recoil energy spectrum that can allow for the reconstruction of the neutrino mass in the event of a positive detection. We first derive new bounds on this model based on the data from the COHERENT collaboration on CsI and LAr targets, which we find do not yet probe new areas of the parameter space. We then assess how well future SS experiments will be able to measure the sterile neutrino mass and mixings, showing that masses in the range 15-50 MeV can be reconstructed. We show that there is a degeneracy in the measurement of the sterile neutrino mixing that substantially affects the reconstruction of parameters for masses of the order of 40 MeV. Thanks to their lower energy threshold and sensitivity to the solar tau neutrino flux, DD experiments allow us to partially lift the degeneracy in the sterile neutrino mixings and considerably improve its mass reconstruction down to 9 MeV. Our results demonstrate the excellent complementarity between DD and SS experiments in measuring the sterile neutrino mass and highlight the power of DD experiments in searching for new physics in the neutrino sector.

hep-ph↗
arXiv2023

Backsplash galaxies and their impact on galaxy evolution: a three-stage, four-type perspective

We study the population of backsplash galaxies at $z=0$ in the outskirts of massive, isolated clusters of galaxies taken from the MDPL2-SAG semi-analytic catalogue. We consider four types of backsplash galaxies according to whether they are forming stars or passive at three stagesin their lifetimes: before entering the cluster, during their first incursion through the cluster, and after they exit the cluster. We analyse several geometric, dynamic, and astrophysical aspects of the four types at the three stages. Galaxies that form stars at all stages account for the majority of the backsplash population ($58\%$) and have stellar masses typically below $M_\star\sim 3\times 10^{10} h^{-1}{\rm M}_\odot$ that avoid the innermost cluster's regions and are only mildly affected by it. In a similar mass range, galaxies that become passive after exiting the cluster ($26\%$) follow orbits characterised by small pericentric distance and a strong deflection by the cluster potential well while suffering a strong loss of both dark matter and gas content. Only a small fraction of our sample ($4\%$) become passive while orbiting inside the cluster. These galaxies have experienced heavy pre-processing and the cluster's tidal stripping and ram pressure provide the final blow to their star formation. Finally, galaxies that are passive before entering the cluster for the first time ($12\%$) are typically massive and are not affected significantly by the cluster. Using the bulge/total mass ratio as a proxy for morphology, we find that a single incursion through a cluster do not result in significant morphological changes in all four types.

astro-ph.GA↗
arXiv2021

Reconstructing Orbits of Galaxies in Extreme Regions (ROGER) II: reliability of projected phase-space in our understanding of galaxy populations

We connect galaxy properties with their orbital classification by analysing a sample of galaxies with stellar mass $M_{\star} \geq 10^{8.5}h^{-1}M_\odot$ residing in and around massive and isolated galaxy clusters with mass $M_{200} > 10^{15}h^{-1}M_\odot$ at redshift $z=0$. The galaxy population is generated by applying the semi-analytic model of galaxy formation SAG on the cosmological simulation MultiDark Planck 2. We classify galaxies considering their real orbits (3D) and their projected phase-space position using the ROGER code (2D). We define five categories: cluster galaxies, galaxies that have recently fallen into a cluster, backsplash galaxies, infalling galaxies, and interloper galaxies. For each class, we analyse the $g-r$ colour, the specific star formation rate (sSFR), and the stellar age, as a function of the stellar mass. For the 3D classes, we find that cluster galaxies have the lowest sSFR, and are the reddest and the oldest, as expected from environmental effects. Backsplash galaxies have properties intermediate between the cluster and recent infaller galaxies. For each 2D class, we find an important contamination by other classes. We find it necessary to separate the galaxy populations in red and blue to perform a more realistic analysis of the 2D data. For the red population, the 2D results are in good agreement with the 3D predictions. Nevertheless, when the blue population is considered, the 2D analysis only provides reliable results for recent infallers, infalling galaxies and interloper galaxies.

astro-ph.GA↗
arXiv2020

ROGER: Reconstructing Orbits of Galaxies in Extreme Regions using machine learning techniques

We present the ROGER (Reconstructing Orbits of Galaxies in Extreme Regions) code, which uses three different machine learning techniques to classify galaxies in, and around, clusters, according to their projected phase-space position. We use a sample of 34 massive, $M_{200}>10^{15} h^{-1} M_{\odot}$, galaxy clusters in the MultiDark Planck 2 (MDLP2) simulation at redshift zero. We select all galaxies with stellar mass $M_{\star} \ge 10^{8.5} h^{-1}M_{\odot}$, as computed by the semi-analytic model of galaxy formation SAG, that are located in, and in the vicinity of, the clusters and classify them according to their orbits. We train ROGER to retrieve the original classification of the galaxies out of their projected phase-space positions. For each galaxy, ROGER gives as output the probability of being a cluster galaxy, a galaxy that has recently fallen into a cluster, a backsplash galaxy, an infalling galaxy, or an interloper. We discuss the performance of the machine learning methods and potential uses of our code. Among the different methods explored, we find the K-Nearest Neighbours algorithm achieves the best performance.

astro-ph.GA↗
arXiv2019

Faraday Rotation Measure Dependence on Galaxy Clusters Dynamics

We study the magnetic fields in galaxy clusters through Faraday rotation measurements crossing systems in different dynamical states. We confirm that magnetic fields are present in those systems and analyze the difference between relaxed and unrelaxed samples with respect to the dispersion between their inherent Faraday Rotation measurements. We found an increase of this RM dispersion and a higher RM overlapping frequency for unrelaxed clusters. This fact suggests that a large scale physical process is involved in the nature of unrelaxed systems and possible depolarization effects are present in the relaxed ones. We show that dynamically unrelaxed systems can enhance magnetic fields to large coherence lengths. In contrast, the results for relaxed systems suggests that small-scale dynamo can be a dominant mechanism for sustaining magnetic fields, leading to intrinsic depolarization.

astro-ph.CO↗
arXiv2018

I. Analysis of candidates for interacting galaxy clusters

Merging galaxy clusters allows to study the different mass components, dark and baryonic, separately. Also their occurrence enables to test the $Λ$CDM scenario and they could put constrains in the self interacting cross section of the dark matter particle. It is necessary to perform an homogeneous analysis of these systems. Hence, based in a recently presented sample of candidates for interacting galaxy clusters, we present the analysis of two of these cataloged systems. In this work, the first of a serie devoted to characterize galaxy clusters in merger process, we perform a weak lensing analysis of A1204 and A2029/2033 clusters to derive the total masses of each identified interacting structures together with a dynamical study based on a two-body model. We also describe the gas and the mass distributions in the field through a lensing and an X-ray analysis. This is the first of a series of works which will analyze these type of systems to characterize them. Both merging clusters candidates do not show evidence of having had a recent merger event. Nevertheless, there is dynamical evidence that these systems could be interacting or could interact in the future. It is necessary to include more constrains in order to improve the methodology to classify merging galaxy clusters. Characterization of these clusters is important in order to understand in deep the nature of these systems and their connection with dynamical studies.

astro-ph.CO↗
arXiv2015

The MeSsI (Merging Systems Identification) Algorithm & Catalogue

Merging galaxy systems provides observational evidence of the existence of dark matter and constraints on its properties. Therefore, statistical uniform samples of merging systems would be a powerful tool for several studies. In this work we presents a new methodology for merging systems identification and the results of its application to galaxy redshift surveys. We use as starting point a mock catalogue of galaxy systems, identified using traditional FoF algorithms, which experienced a major merger as indicated by its merger tree. Applying machine learning techniques in this training sample, and using several features computed from the observable properties of galaxy members, it is possible to select galaxy groups with a high probability of have been experienced a major merger. Next we apply clustering techniques on galaxy members in order to reconstruct the properties of the haloes involved in such merger. This methodology provides a highly reliable sample of merging systems with low contamination and precise recovered properties. We apply our techniques in samples of galaxy systems obtained from SDSS-DR7, WINGS and HeCS. Our results recover previously known merging systems and provide several new candidates. We present its measured properties and discuss future analysis on current and forthcoming samples.

astro-ph.CO↗