arXiv subjects

Alejandro Torres-Forné

Publications and source records attributed to Alejandro Torres-Forné.

At least 19 recordsLinked to original sources

arXiv2026

DANSur_HM: Modularly incorporating higher modes in a deep learning based gravitational-wave surrogate

Numerical relativity (NR) simulations provide the most faithful representation of the gravitational waves (GWs) emitted by binary black hole (BBH) systems during merger. In the context of GW astronomy, tasks such as parameter estimation can require vast numbers of waveform evaluations per second across the entire parameter space. Since performing full NR simulations for each evaluation is not computationally feasible, interpolating methods for existing NR waveforms, known as surrogate models, have been developed with marked success. In this paper, we build on our previous work to introduce methods to train a fast surrogate model based on neural networks in order to generate BBH merger waveforms, including the fundamental (2,2) mode, as well as the (3,3), (2,1), (4,4), (3,2), (4,3) and (5,5) higher-order modes. Applying a pretraining step on approximant data before fine-tuning on NR data allows us to smooth out the parameter space, and making use of the parallelization ability of GPUs to project the NR waveforms in the inclination-phase $(ι, ϕ)$ sphere during training allows the fitting of all the explored modes simultaneously. The developed surrogate model, \texttt{DANSur\_HM}, achieves average mismatches of the order of $10^{-4}$, with the worst mismatch at $2.5\times10^{-3}$, and achieves throughput above $6\times10^5$ waveforms/second on an NVIDIA V100 GPU. Parameter estimation tests confirm the usefulness of the inclusion of higher modes.

gr-qc↗

arXiv2026

Automatic classification pipeline for glitches in the Virgo detector

Glitches frequently contaminate data in gravitational-wave detectors, complicating the observation and analysis of astrophysical signals. This work introduces VIGILant, an automatic pipeline for classification and visualization of glitches in the Virgo detector. Using a curated dataset of Virgo O3b glitches, two machine learning approaches are evaluated: tree-based models (Decision Tree, Random Forest and XGBoost) using structured Omicron parameters, and Convolutional Neural Networks (ResNet) trained on spectrogram images. While tree-based models offer higher interpretability and fast training, the ResNet34 model achieved superior performance, reaching a F1 score of 0.9772 and accuracy of 0.9833 in the testing set, with inference times of tens of milliseconds per glitch. The pipeline has been deployed for daily operation at the Virgo site since observing run O4c, providing the Virgo collaboration with an interactive dashboard to monitor glitch populations and detector behavior. This allows to identify low-confidence predictions, highlighting glitches requiring further attention.

gr-qc↗

arXiv2026

Toward More Realistic Machine-Learning Inference of the Dense-Matter Equation of State from Supernova Gravitational Waves

Gravitational waves from core-collapse supernovae offer a unique probe of the equation of state (EOS) of dense nuclear matter. For rapidly rotating stars, previous machine-learning studies demonstrated promising EOS classification accuracy. However, these analyses relied on several simplifying assumptions. In this work, we relax three key assumptions. First, we include real detector noise. Second, we expand the analysis from a single progenitor model to four models spanning 12 to 40 solar masses, and for each mass we consider multiple rotational configurations, from slow to rapid. Third, we introduce uncertainty in the core bounce time of up to 20 ms, rather than assuming it is known precisely. We find that none of these effects significantly degrades EOS classification performance. Instead, the larger dataset associated with multiple progenitor models and noise realizations improves training and classification accuracy. This study is a step in a broader effort to progressively incorporate more realistic conditions into gravitational-wave inference for core-collapse supernovae.

astro-ph.HE↗

arXiv2026

Strongly interacting matter with criticality induced by modified excluded volume in core-collapse supernova simulations

This article reviews critically the core-collapse supernova explosion mechanism associated with a sufficiently strong first-order phase transition from normal nuclear, in general hadronic matter to deconfined quark matter, which commonly assumes Gibbs conditions for the coexistence of phases and a phase transition construction accordingly. To this end, a novel class of multi-purpose equation of state (EOS) is developed, based on the modified excluded volume (MEV) approach employing a medium-dependent excluded-volume functional within the relativistic mean field framework with density-dependent meson-nucleon couplings. The chosen MEV parametrisation features the change in the number of degrees of freedom, mimicking the EOS softening in excess of nuclear saturation density, featuring a first-order phase transition with van der Waals like behaviour and the presence of a critical point at high temperatures. Simulations of core-collapse supernovae are performed, based on general relativistic neutrino radiation hydrodynamics in spherical symmetry, in order to explore the previously reported supernova explosion scenario within this class of phenomenological modified microscopic hadronic EOS. A burst-like neutrino signature is released, substantially longer than previously reported based on common hadron-quark hybrid model EOS with two-phase approach and Gibbs phase-transition construction, as observable signal, which is complemented by a gravitational wave mode analysis.

astro-ph.HE↗

arXiv2025

Classification of the equation of state of neutron stars via sparse dictionary learning

The post-merger phase of binary neutron star (BNS) mergers encodes valuable information about the equation of state (EOS) of supranuclear matter. Extracting this information from the analysis of the post-merger waveforms remains challenging due to the high-frequency limitations of current detectors. Future third-generation observatories, such as the Einstein Telescope (ET) and NEMO, will have the sensitivity required to resolve post-merger signals with high fidelity. In this work, we apply CLAWDIA, our recently developed sparse dictionary learning (SDL) framework, to classify different EOS models using only the post-merger gravitational-wave emission of simulated BNS mergers available in the CoRe database. Our dataset comprises five EOS models representative of a broad range of neutron star properties. The SDL framework is optimised under realistic detection conditions by injecting signals into simulated noise matching the sensitivity curves of ET and NEMO. Our results show that classification is primarily driven by the dominant post-merger frequency, $f_2$, which encodes EOS-dependent information. At a modest signal-to-noise ratio of 5, our method achieves $F_1$ scores of $0.76$ for ET and $0.70$ for NEMO, with performance improving for higher signal-to-noise ratios. The reliability and generalisation capabilities of the model are assessed with additional tests, including the classification of an EOS not included in the training dataset and the analysis of detector-specific biases.

astro-ph.HE↗

arXiv2025

Assessing the Distance for Probing the Nuclear Equation of State with Supernova Gravitational Waves

Gravitational waves from core-collapse supernovae provide a unique probe of the equation of state (EOS) of high density matter. In this work, we focus on the bounce signal from numerical simulations of rotating supernovae and explore its potential for EOS inference. We employ a support vector machine, previously shown to perform best among tested methods, to classify GW signals simulated for 18 EOS models. For optimally oriented sources, we estimate that the Advanced LIGO A+ detector can probe the EOS for Galactic events, while third-generation observatories such as the Einstein Telescope and Cosmic Explorer can reach substantially farther. For randomly oriented sources, only these next-generation detectors are expected to have sufficient sensitivity. These results represent the potential observational range for probing the nuclear EOS, although, due to the simplifying assumptions adopted, they should be regarded as approximate upper limits.

astro-ph.HE↗

arXiv2025

Proto-neutron star oscillations including accretion flows

The gravitational wave signature from core-collapse supernovae (CCSNe) is dominated by quadrupolar oscillation modes of the newly born proto-neutron star (PNS), and could be detectable at galactic distances. We have developed a framework for computing the normal oscillation modes of a PNS in general relativity, including, for the first time, the presence of an accretion flow and a surrounding stalled accretion shock. These new ingredients are key to understand PNS oscillation modes, in particular those related to the standing-accretion-shock instability (SASI). Their incorporation is an important step towards accurate PNS asteroseismology. For this purpose, we perform linear and adiabatic perturbations of a spherically symmetric background, in the relativistic Cowling approximation, and cast the resulting equations as an eigenvalue problem. We discretize the eigenvalue problem using collocation Chebyshev spectral methods, which is then solved by means of standard and efficient linear algebra methods. We impose boundary conditions at the accretion shock compatible with the Rankine-Hugoniot conditions. We present several numerical examples to assess the accuracy and convergence of the numerical code, as well as to understand the effect of an accretion flow on the oscillation modes, as a stepping stone towards a complete analysis of the CCSNe case.

astro-ph.HE↗

arXiv2025

CLAWDIA: A dictionary learning framework for gravitational-wave data analysis

Deep-learning methods are becoming increasingly important in gravitational-wave data analysis, yet their performance often relies on large training datasets and models whose internal representations are difficult to interpret. Sparse dictionary learning (SDL) offers a complementary approach: it performs well in scarce-data regimes and yields physically interpretable representations of gravitational-wave morphology. Here we present CLAWDIA (Comprehensive Library for the Analysis of Waves via Dictionary-based Algorithms), an open-source Python framework that integrates SDL-based denoising and classification under realistic detector noise. We systematise previously isolated SDL workflows into a unified, modular environment with a consistent, user-friendly interface. The current release provides several time-domain denoising strategies based on LASSO-regularised sparse coding and a classifier based on Low-Rank Shared Dictionary Learning. A companion toolbox, GWADAMA, supports dataset construction and realistic conditioning of real and simulated interferometer data. We demonstrate CLAWDIA's performance by denoising the signal from binary neutron star event GW170817 and by classifying families of instrumental glitches from LIGO's third observing run, highlighting robustness at low signal-to-noise ratios. CLAWDIA is intended as a community-driven, interoperable library extensible to additional tasks, including detection and parameter estimation.

astro-ph.IM↗

arXiv2025

A Deep Learning Powered Numerical Relativity Surrogate for Binary Black Hole Waveforms

Gravitational-wave approximants are essential for gravitational-wave astronomy, allowing the coverage binary black hole parameter space for inference or match filtering without costly numerical relativity (NR) simulations, but generally trading some accuracy for computational efficiency. To reduce this trade-off, NR surrogate models can be constructed using interpolation within NR waveform space. We present a 2-stage training approach for neural network-based NR surrogate models. Initially trained on approximant-generated waveforms and then fine-tuned with NR data, these dual-stage artificial neural surrogate (\texttt{DANSur}) models offer rapid and competitively accurate waveform generation, generating millions in under 20ms on a GPU while keeping mean mismatches with NR around $10^{-4}$. Implemented in the \textsc{bilby} framework, we show they can be used for parameter estimation tasks.

gr-qc↗

arXiv2025

Rapid detection of gravitational waves from binary black hole mergers using sparse dictionary learning

Current gravitational wave (GW) detection pipelines for compact binary coalescence based on matched-filtering have reported over 90 confident detections during the first three observing runs of the LIGO-Virgo-KAGRA (LVK) detector network. Decreasing the latency of detection, in particular for future detectors anticipated to have high detection rates, remains an ongoing effort. In this paper, we develop and test a sparse dictionary learning (SDL) algorithm for the rapid detection of GWs. We evaluate the algorithms biases and estimate its GW detection rate for an astrophysical population of binary black holes. The SDL algorithm is assessed using both, simulated data injected into the proposed A+ detector sensitivity and real data containing confident detections from the third LVK observing run. We find that our SDL algorithm can reconstruct a single binary black hole signal in less than 1 s. This suggests that SDL could be regarded as a promising approach for rapid, efficient GW detection in future observing runs of ground-based detectors.

gr-qc↗

arXiv2024

Deep-Learning Classification and Parameter Inference of Rotational Core-Collapse Supernovae

We test deep-learning (DL) techniques for the analysis of rotational core-collapse supernovae (CCSN) gravitational-wave (GW) signals by performing classification and parameter inference of the maximum (peak) frequency and the GW strain amplitude ($Δh$) multiplied by the luminosity distance ($D$) attained at core bounce, respectively, $(f_{peak})$ and $(D \cdot Δh)$. Our datasets are built from a catalog of numerically generated CCSN waveforms assembled by Richers et al. 2017. Those waveforms are injected into noise from the Advanced Laser Interferometer Gravitational Wave Observatory and Advanced Virgo detectors corresponding to the O2 and O3a observing runs. For a network signal-to-noise ratio (SNR) above 5, our classification network using time series detects Galactic CCSN GW signals buried in detector noise with a false positive rate of 0.10% and a 98% accuracy, being able to detect all signals with SNR>10. The inference of $f_{peak}$ is more accurate than for $D \cdot Δh $, particularly for our datasets with the shortest time window (0.25 s) and for a minimum SNR=15. From the calibration plots of predicted versus true values of the two parameters, the standard deviation ($σ$) and the slope deviation with respect to the ideal value are computed. We find $σ_{D \cdot Δh} = 52.6$ cm and $σ_{f_{peak}} = 18.3$ Hz, with respective slope deviations of 11.6% and 8.3%. Our best model is also tested on waveforms from a recent CCSN catalog built by Mitra et al. 2023, different from the one used for the training. For these new waveforms, the true values of the two parameters are mostly within the $1σ$ band around the network's predicted values. Our results show that DL techniques hold promise to infer physical parameters of Galactic rotational CCSN events.

astro-ph.HE↗

arXiv2024

High-speed reconstruction of long-duration gravitational waves from extreme mass ratio inspirals using sparse dictionary learning

Measuring accurate long-duration gravitational waves from extreme mass ratio inspirals (EMRIs) could provide scientifically fruitful knowledge of massive black hole populations and robust tests for general relatively during the LISA mission. However, the immense computational requirements surrounding EMRI data processing and analysis makes their detection and analysis challenging. We further develop and explore a sparse dictionary learning (SDL) algorithm to expeditiously reconstruct EMRI gravitational waveforms lasting as long as 1 year. A suite of year-long EMRI systems are studied to understand the detection and accurate waveform retrieval prospects of the method. We show that full-year EMRIs can be reconstructed within 2 minutes, some with a false alarm rate less than 0.001/yr and with 1.16 day time windows with mismatch as low as 0.06. This provides an encouraging prospect to use the SDL method for long-duration GW searches like that for EMRIs in this study.

gr-qc↗

arXiv2024

Numerical relativity surrogate models for exotic compact objects: the case of head-on mergers of equal-mass Proca stars

We present several high-accuracy surrogate models for gravitational-wave signals from equal-mass head-on mergers of Proca stars, computed through the Newman-Penrose scalar $ψ_4$. We also discuss the current state of the model extensions to mergers of Proca stars with different masses, and the particular challenges that these present. The models are divided in two main categories: two-stage and monolithic. In the two-stage models, a dimensional reduction algorithm is applied to embed the data in a reduced feature space, which is then interpolated in terms of the physical parameters. For the monolithic models, a single neural network is trained to predict the waveform from the input physical parameter. Our model displays mismatches below $10^{-3}$ with respect to the original numerical waveforms. Finally, we demonstrate the usage of our model in full Bayesian parameter inference through the accurate recovery of numerical relativity signals injected in zero-noise, together with the analysis of GW190521. For the latter, we observe excellent agreement with existing results that make use of full numerical relativity.

gr-qc↗

arXiv2024

Neutron stars in accreting systems -- signatures of the QCD phase transition

Neutron stars (NS) that are born in binary systems with a main-sequence star companion can experience mass transfer, resulting in the accumulation of material at the surface of the NS. This, in turn, leads to the continuous growth of the NS mass and the associated steepening of the gravitational potential. Supposing the central density surpasses the onset for the phase transition from nuclear, generally hadronic matter to deconfined quark-gluon plasma, which is a quantity currently constrained solely from an upper limit by asymptotic freedom in quantum chromodynamics (QCD), the system may experience a dynamic response due to the appearance of additional degrees of freedom in the equation of state (EOS). This dynamical response might give rise to a rapid softening of the EOS during the transition in the hadron-quark matter co-existence region. While this phenomenon has long been studied in the context of hydrostatic configurations, the dynamical implications of this problem are still incompletely understood. It is the purpose of the present paper to simulate the dynamics of NSs with previously accreted envelopes caused by the presence of a first-order QCD phase transition. Therefore, we employed the neutrino radiation hydrodynamics treatment based on the fully general relativistic approach in spherical symmetry, implementing a three-flavor Boltzmann neutrino transport and a microscopic model EOS that contains a first-order hadron-quark phase transition. The associated neutrino signal shows a sudden rise in the neutrino fluxes and average energies, becoming observable for the present generation of neutrino detectors for a galactic event, and a gravitational wave mode analysis revealed the behaviors of the dominant $f$ mode and the first and the second gravity $g$ modes that are excited during the NS evolution across the QCD phase transition.

astro-ph.HE↗

arXiv2024

Solving the Teukolsky equation with physics-informed neural networks

We use physics-informed neural networks (PINNs) to compute the first quasi-normal modes of the Kerr geometry via the Teukolsky equation. This technique allows us to extract the complex frequencies and separation constants of the equation without the need for sophisticated numerical techniques, and with an almost immediate implementation under the \texttt{PyTorch} framework. We are able to compute the oscillation frequencies and damping times for arbitrary black hole spins and masses, with accuracy typically below the percentual level as compared to the accepted values in the literature. We find that PINN-computed quasi-normal modes are indistinguishable from those obtained through existing methods at signal-to-noise ratios (SNRs) larger than 100, making the former reliable for gravitational-wave data analysis in the mid term, before the arrival of third-generation detectors like LISA or the Einstein Telescope, where SNRs of ${\cal O}(1000)$ might be achieved.

gr-qc↗

arXiv2024

Determining the core-collapse supernova explosion mechanism with current and future gravitational-wave observatories

Gravitational waves are emitted from deep within a core-collapse supernova, which may enable us to determine the mechanism of the explosion from a gravitational-wave detection. Previous studies suggested that it is possible to determine if the explosion mechanism is neutrino-driven or magneto-rotationally powered from the gravitational-wave signal. However, long duration magneto-rotational waveforms, that cover the full explosion phase, were not available during the time of previous studies, and explosions were just assumed to be magneto-rotationally driven if the model was rapidly rotating. Therefore, we perform an updated study using new 3D long-duration magneto-rotational core-collapse supernova waveforms that cover the full explosion phase, injected into noise for the Advanced LIGO, Einstein Telescope and NEMO gravitational-wave detectors. We also include a category for failed explosions in our signal classification results. We then determine the explosion mechanism of the signals using three different methods: Bayesian model selection, dictionary learning, and convolutional neural networks. The three different methods are able to distinguish between neutrino-driven explosions and magneto-rotational explosions, even if the neutrino-driven explosion model is rapidly rotating. However they can only distinguish between the non-exploding and neutrino-driven explosions for signals with a high signal to noise ratio.

astro-ph.HE↗

arXiv2024

Bayesian inference from gravitational waves in fast-rotating, core-collapse supernovae

Core-collapse supernovae (CCSNe) are prime candidates for gravitational-wave detectors. The analysis of their complex waveforms can potentially provide information on the physical processes operating during the collapse of the iron cores of massive stars. In this work we analyze the early-bounce rapidly rotating CCSN signals reported in the waveform catalog of Richers et al 2017, which comprises over 1800 axisymmetric simulations extending up to about 10~ms of post-bounce evolution. It was previously established that for a large range of progenitors, the amplitude of the bounce signal, $Δh$, is proportional to the ratio of rotational-kinetic energy to potential energy, T/|W|, and the peak frequency, $f_{\rm peak}$, is proportional to the square root of the central rest-mass density. In this work, we exploit these relations to suggest that it could be possible to use such waveforms to infer protoneutron star properties from a future gravitational wave observation, if the distance and inclination are well known. Our approach relies on the ability to describe a subset of the waveforms in the early post-bounce phase in a simple form depending only on two parameters, $Δh$ and $f_{\rm peak}$. We use this template to perform a Bayesian inference analysis of waveform injections in Gaussian colored noise for a network of three gravitational wave detectors formed by Advanced LIGO and Advanced Virgo. We show that, for a galactic event, it is possible to recover the peak frequency and amplitude with an accuracy better than 10% for about 80% and 60% of the signals, respectively, given known distance and inclination angle. However, inference on waveforms from outside the Richers catalog is not reliable, indicating a need for carefully verified waveforms of the first 10 ms after bounce of rapidly rotating supernovae of different progenitors with agreement between different codes.

astro-ph.HE↗

arXiv2023

Multi-messenger observations of core-collapse supernovae: Exploiting the standing accretion shock instability

The gravitational wave (GW) and neutrino signals from core-collapse supernovae (CCSNe) are expected to carry pronounced imprints of the standing accretion shock instability (SASI). We investigate whether the correlation between the SASI signatures in the GW and neutrino signals could be exploited to enhance the detection efficiency of GWs. We rely on a benchmark full-scale three-dimensional CCSN simulation with zero-age main sequence mass of $27\ M_\odot$. Two search strategies are explored: 1.~the inference of the SASI frequency range and/or time window from the neutrino event rate detectable at the IceCube Neutrino Observatory; 2.~the use of the neutrino event rate to build a matched filter template. We find that incorporating information from the SASI modulations of the IceCube neutrino event rate can increase the detection efficiency compared to standard GW excess energy searches up to $30\%$ for nearby CCSNe. However, we do not find significant improvements in the overall GW detection efficiency for CCSNe more distant than $1.5$~kpc. We demonstrate that the matched filter approach performs better than the unmodeled search method, which relies on a frequency bandpass inferred from the neutrino signal. The improved detection efficiency provided by our matched filter method calls for additional work to outline the best strategy for the first GW detection from CCSNe.

astro-ph.HE↗