arXiv subjects

Michał Bejger

Publications and source records attributed to Michał Bejger.

At least 19 recordsLinked to original sources

arXiv2026

Deterministic denoising of long-duration gravitational-wave signal candidates with $α$-(de)blending

We present a deterministic machine-learning method for denoising candidate signals from all-sky searches for continuous gravitational waves from rotating neutron stars. Using the time-domain $F$-statistic search pipeline, we post-process the resulting $F(f,\dot{f})$ candidate patterns with iterative $α$-(de)blending, a deterministic diffusion-type generative model implemented as a U-Net, well suited to the non-Gaussian, correlated noise of the $F$-statistic output. Two models, trained on data based on simulated 6- and 12-day time-domain segments with software-injected signals similar to the LVK hardware injections, are tested by comparing deblended images to the library of expected signal patterns via the Structural Similarity Index Measure. In general the method recovers the correct sky-position-dependent pattern for signal-to-noise ratios $ρ\gtrsim 4$, moderately below typical all-sky detection thresholds, with recovery depending strongly on pattern morphology, demonstrating that deterministic diffusion-based denoising may serve as a consistency/veto tool ahead of the semi-coherent coincidence stage. We also discuss limitations and possible improvements of the method.

gr-qc↗

arXiv2026

NS-UNO: Neutron Star EoS Inference from an Unconstrained Number of Observations

Future multimessenger observations of neutron stars (NS) are expected to substantially increase both the number and precision of astrophysical constraints on the equation of state (EoS) of dense matter. This motivates inference frameworks capable of accommodating a variable, non fixed number of observations while preserving the posterior information associated with each measurement. In this work, we introduce NS-UNO, a Neural Posterior Estimation framework for NS EoS inference designed to accommodate an Unconstrained Number of Observations (UNO). NS-UNO combines a hierarchical DeepSets model with a conditional normalising flow, enabling a single trained model to perform inference from mass-radius observation sets of varying size, with each observation represented by a set of posterior samples. We demonstrate accurate and well calibrated posterior reconstructions using a model trained jointly on piecewise polytropic and non-parametric Gaussian process EoS ensembles. The reconstruction improves as observations probe a broader range of NS masses, while remaining robust to variations in the number and precision of the observations. The model also generalises to EoSs outside the families used during training. Finally, we qualitatively demonstrate the framework on current multimessenger constraints from NICER and GW170817. NS-UNO provides a flexible and scalable approach to NS EoS inference, naturally suited to the increasingly diverse observational datasets expected from next generation multimessenger astronomy.

nucl-th↗

arXiv2026

Probing globular clusters parameters through gravitational wave lensing with stellar-mass black hole binaries

Globular clusters (GCs) can act as gravitational lenses for gravitational waves(GWs) in the wave-optics regime, imprinting frequency-dependent signatures on the observed signal. We investigate whether such lensing effects can be used to probe intrinsic properties of GCs, in particular their central velocity dispersion. Modeling GCs as singular isothermal spheres, we simulate lensed GW150914-like signals and perform Bayesian parameter estimation using waveform templates that include both source and lens parameters. We show that the effective lensing mass can be recovered and, when combined with GW sky localization information and GC catalogs, allows for an estimate of the cluster velocity dispersion. For favorable source-lens alignments, the injected values are well recovered within credible intervals. Our results demonstrate that lensed GWs can provide a complementary probe of GC dynamics and motivate searches for such signatures in current and future observations.

gr-qc↗

arXiv2026

APRIL: Auxiliary Physically-Redundant Information in Loss -- A physics-informed framework for parameter estimation with a gravitational-wave case study

Physics-Informed Neural Networks (PINNs) embed the partial differential equations (PDEs) governing the system under study directly into the training of Neural Networks, ensuring solutions that respect physical laws. While effective for single-system problems, standard PINNs scale poorly to datasets containing many realizations of the same underlying physics with varying parameters. To address this limitation, we present a complementary approach by including auxiliary physically-redundant information in loss (APRIL), i.e. augment the standard supervised output-target loss with auxiliary terms which exploit exact physical redundancy relations among outputs. We mathematically demonstrate that these terms preserve the true physical minimum while reshaping the loss landscape, improving convergence toward physically consistent solutions. As a proof-of-concept, we benchmark APRIL on a fully-connected neural network for gravitational wave (GW) parameter estimation (PE). We use simulated, noise-free compact binary coalescence (CBC) signals, focusing on inspiral-frequency waveforms to recover the chirp mass $\mathcal{M}$, the total mass $M_\mathrm{tot}$, and symmetric mass ratio $η$ of the binary. In this controlled setting, we show that APRIL achieves up to an order-of-magnitude improvement in test accuracy, especially for parameters that are otherwise difficult to learn. This method provides physically consistent learning for large multi-system datasets and is well suited for future GW analyses involving realistic noise and broader parameter ranges.

gr-qc↗

arXiv2025

Dynamical tides in neutron stars with first-order phase transitions: the role of the discontinuity mode

During the late stages of a binary neutron star inspiral, dynamical tides induced in each star by its companion become significant and should be included in complete gravitational-wave (GW) modeling. We investigate the coupling between the tidal field and quasi-normal modes in hybrid stars and show that the discontinuity mode ($g$-mode) - intrinsically associated with first-order phase transitions and buoyancy - contributes non-negligibly compared with the fundamental $f$-mode. We find that the $g$-mode overlap integral can reach up to $\sim 10\%$ of the $f$-mode value for hybrid star masses in the range 1.4-2.0$M_{\odot}$, with the largest values generally associated with larger density jumps. This leads to a GW phase shift due to the $g$-mode of $Δϕ_g \lesssim 0.1$-$1$ rad (i.e., up to $\sim5\%-10\%$ of $Δϕ_f$), with the largest shifts occurring for masses near the phase transition. At higher masses, the shifts remain smaller and nearly constant, with $Δϕ_g \lesssim 0.1$ rad (roughly $\sim 1\%$ of $Δϕ_f$). These GW shifts may be relevant even at the design sensitivity of current second-generation GW detectors in the most optimistic cases. Moreover, if a $g$-mode is present and lies near the $f$-mode frequency, neglecting it in the GW modeling can lead to systematic biases in neutron star parameter estimation, resulting in radius errors of up to $1\%-2\%$. These results show the importance of dynamical tides to probe neutron stars' equation of state, and to test the existence of dense-matter phase transitions.

astro-ph.HE↗

arXiv2025

Microlensing of long-duration gravitational wave signals originating from Galactic sources

Detection of quasi-monochromatic, long-duration (continuous) gravitational wave radiation emitted by, e.g., asymmetric rotating neutron stars in our Galaxy requires a long observation time to distinguish it from the detector's noise. If this signal is additionally microlensed by a lensing object located in the Galaxy, its magnitude would be temporarily magnified, which may lead to its discovery and allow probing of the physical nature of the lensing object and the source. We study the observational effect of microlensing of continuous gravitational wave signals for Galactic sources and lenses in the point mass lens approximation. In particular, we examine the regions of the parameter space that are promising for lensed CW searches, and perform example simulations to demonstrate how the lensing effect affects the continuous-wave signal. We show that an analytical lensing pattern can be identified from the lensed continuous wave signal using the Time-Domain F-statistic search, as the estimated signal-to-noise ratio in each time-domain segment scales directly with the amplification factor.

gr-qc↗

arXiv2025

Neural Posterior Estimation of Neutron Star Equations of State

We present a simulation-based inference (SBI) framework to constrain the neutron star (NS) equation of state (EoS) from astrophysical observations of masses, radii and tidal deformabilities, using Neural posterior estimation (NPE) with Conditional Normalising Flows (CNF). To ensure that the model conforms with reality, physics-informed constraints are embedded directly into the training loss. This enables efficient, likelihood-free inference of full posterior distributions for key thermodynamic quantities-including pressure, squared speed of sound, and the trace anomaly-conditioned on observational data. Our models are trained on synthetic datasets generated from two agnostic EoS priors: polytropic parametrizations (PT) and gaussian process (GP) reconstructions. These datasets span various scenarios, including the presence or absence of tidal deformability information and observational noise. Across all settings, the method produces accurate and well-calibrated posteriors, with uncertainties reduced when tidal deformability constraints are included. Furthermore, we find that the behavior of normalized predictive dispersions is strongly correlated with the maximum central density inside NSs, suggesting that the model can indirectly infer this physically meaningful quantity. The approach generalizes well across EoS families and accurately reconstructs derivative quantities such as the polytropic index, demonstrating its robustness and potential for probing dense matter in NS cores.

nucl-th↗

arXiv2025

PINNGraPE: Physics Informed Neural Network for Gravitational wave Parameter Estimation

Weakly-modelled searches for gravitational waves are essential for ensuring that all potential sources are accounted for in detection efforts, as they make minimal assumptions regarding source morphology. While these searches primarily target generic transient sources, they are also highly effective at identifying a broad range of compact binary coalescences, demonstrated by the weakly-modelled search algorithm coherent WaveBurst being the first to detect GW150914. Despite their ability to detect compact binaries with diverse properties, the accurate estimation of source parameters from their output remains to be a challenging task. To overcome this, we leverage physics-informed neural networks, which serve as a powerful tool for parameter estimation by applying physical constraints through the universal differential equation governing a compact binary system. With this approach, we rapidly infer the mass parameters of binary black hole merger systems to within 7% from only the time-frequency representation of the gravitational wave signal.

gr-qc↗

arXiv2025

Approximating neutron-star radii using gravitational-wave only measurements with symbolic regression

Gravitational waves emitted by binary neutron-star inspirals carry information on components' masses and tidal deformabilities, but not directly radii, which are measured by electromagnetic observations of neutron stars. To improve the multi-messenger astronomy studies of neutron stars, an expression for neutron-star radii as a function of gravitational-wave only data would be advantageous, as it would allow to compare information from two different channels. In order to do so, a symbolic regression method, pySR, is trained on TOV solutions to piecewise polytropic EOS input to discover an approximate symbolic expression for the neutron-star radius as a function of gravitational-wave measurements only. The approximation is tested on piecewise polytropic EOS NS data, as well as on NS sequences based on selected realistic (non-polytropic) dense-matter theory EOSs, achieving consistent agreement between the ground truth values and the symbolic approximation for a broad range of NS parameters covering current astrophysical observations, with average radii differences of few hundred meters. Additionally, the approximation is applied to the GW170817 gravitational-wave mass and tidal deformability posteriors, and compared to reported inferred radius distributions.

gr-qc↗

arXiv2025

Explainable autoencoder for neutron star dense matter parameter estimation

We present a physics-informed autoencoder designed to encode the equation of state of neutron stars into an interpretable latent space. In particular the input will be encoded in the mass, radius, and tidal deformability values of a neutron star. Unlike traditional black-box models, our approach incorporates additional loss functions to enforce explainability in the encoded representations. This method enhances the transparency of machine learning models in physics, providing a robust proof-of-concept tool to study compact stars data. Our results demonstrate that the proposed autoencoder not only accurately estimates the EoS parameters and central density/pressure but also offers insights into the physical connection between equation of state and observable physical quantities. This framework conceptualizes the physical differential equations themselves as the ``encoders", allowing interpretability of the latent space.

physics.comp-ph↗

arXiv2025

Conditional variational autoencoder inference of neutron star equation of state from astrophysical observations

We present a new inference framework for neutron star astrophysics based on conditional variational autoencoders. Once trained, the generator block of the model reconstructs the neutron star equation of state from a given set of mass-radius observations. While the pressure of dense matter is the focus of the present study, the proposed model is flexible enough to accommodate the reconstructing of any other quantity related to dense matter equation of state. Our results show robust reconstructing performance of the model, allowing to make instantaneous inference from any given observation set.

nucl-th↗

arXiv2024

Evidence for 3XMM J185246.6+003317 as a massive magnetar with a low magnetic field

3XMM J185246.6+003317 is a transient magnetar located in the vicinity of the supernova remnant Kes\,79. So far, observations have only set upper limits to its surface magnetic field and spindown, and there is no estimate for its mass and radius. Using ray-tracing modelling and Bayesian inference for the analysis of several light curves spanning a period of around three weeks, we have found that it may be one of the most massive neutron stars to date. In addition, our analysis suggests a multipolar magnetic field structure with a subcritical field strength and a carbon atmosphere composition. Due to the time-resolution limitation of the available light curves, we estimate the surface magnetic field and the mass to be $\log_{10} (B/{\rm G}) = 11.89^{+0.19}_{-0.93}$ and $M=2.09^{+0.16}_{-0.09}$~$M_{\odot}$ at $1σ$ confidence level, while the radius is estimated to be $R=12.02^{+1.44}_{-1.42}$ km at $2σ$ confidence level. They were verified by simulations, i.e., data injections with known model parameters, and their subsequent recovery. The best-fitting model has three small hot spots, two of them in the southern hemisphere. These are, however, just first estimates and conclusions, based on a simple ray-tracing model with anisotropic emission; we also estimate the impact of modelling on the parameter uncertainties and the relevant phenomena on which to focus in more precise analyses. We interpret the above best-fitting results as due to accretion of supernova layers/interstellar medium onto 3XMM J185246.6+003317 leading to burying and a subsequent re-emergence of the magnetic field, and a carbon atmosphere being formed possibly due to hydrogen/helium diffusive nuclear burning. Finally, we briefly discuss some consequences of our findings for superdense matter constraints.

astro-ph.HE↗

arXiv2024

Detecting the third family of compact stars with normalizing flows

We explore the anomaly detection framework based on Normalizing Flows (NF) models introduced in \cite{PhysRevC.106.065802} to detect the presence of a large (destabilising) dense matter phase transition in neutron star (NS) observations of masses and radii, and relate the feasibility of detection with parameters of the underlying mass-radius sequence, which is a functional of the dense matter equation of state. Once trained on simulated data featuring continuous $M(R)$ solutions (i.e., no phase transitions), NF is used to determine the likelihood of a first-order phase transition in a given set of $M(R)$ observations featuring a discontinuity, i.e., perform the anomaly detection. Different mock test sets, featuring two branch solutions in the $M(R)$ diagram, were parameterized by the NS mass at which the phase transition occurs, $M_c$, and the radius difference between the heaviest hadronic star and lightest hybrid star, $ΔR$. We analyze the impact of these parameters on the NF performance in detecting the presence of a first-order phase transition. Among the results, we report that given a set of 15 stars with radius uncertainty of $0.2$ km, a detection of a two-branch solution is possible with 95\% accuracy if $ΔR > 0.4$ km.

nucl-th↗

arXiv2023

Search for Postmerger Gravitational Waves from Binary Neutron Star Mergers Using a Matched-filtering Statistic

In this paper, we present a new method to search for a short, a few tens of milliseconds long, postmerger gravitational-wave signal following the merger of two neutron stars. Such a signal could follow the event GW170817 observed by LIGO and Virgo detectors. Our method is based on a matched filtering statistic and an approximate template of the postmerger signal in the form of a damped sinusoid. We test and validate our method using postmerger numerical simulations from the CoRe database. We find no evidence of the short postmerger signal in the LIGO data following the GW170817 event and we obtain upper limits. For short postmerger signals investigated, our best upper limit on the root sum square of the gravitational-wave strain emitted from 1.15 kHz to 4 kHz is $h_{\text{rss}}^{50\%}=1.8\times 10^{-22}/\sqrt{\text{Hz}}$ at 50% detection efficiency. The distance corresponding to this best upper limit is 4.64 Mpc.

gr-qc↗

arXiv2023

Crustal failure as a tool to probe hybrid stars

It is currently unknown if neutron stars (NSs) are composed of nucleons only or are hybrid stars, i.e., in addition to nucleonic crusts and outer cores, they also possess quark cores. Quantum chromodynamics allows for such a possibility, but accurate calculations relevant for compact stars are still elusive. Here we investigate some crust-breaking aspects of hybrid stars. We show that the crust-breaking frequency and maximum fiducial ellipticity are sensitive to the quark-hadron density jump and equation of state stiffness. Remarkably, the crust-breaking frequency related to static tides scales linearly with the mass of the star (for a given companion's mass), and its slope encompasses information about the microphysics of the star. However, for precise crust-breaking frequency predictions, relativistic corrections to Kepler's third law and the Newtonian tidal field should not be ignored. When a liquid quark core touches an elastic hadronic phase (the result of a significant energy-density jump), the maximum ellipticity can increase around an order of magnitude when compared to a liquid quark core touching a liquid hadronic phase. That is relevant because it would increase the odds of detecting continuous gravitational waves from NSs. Our order-of-magnitude analysis also suggests that a given upper limit to the ellipticity(crust-breaking frequency) could have representatives in stars with either small or intermediate(large) energy-density jumps. Therefore, when upper limits to the ellipticity for isolated stars are better constrained or electromagnetic radiation (e.g., gamma-ray precursors) is detected along with gravitational waves in inspiraling binary systems, they may help constrain some aspects of phase transitions in NSs.

astro-ph.HE↗

arXiv2022

Detecting dense-matter phase transition signatures in neutron star mass-radius measurements as data anomalies using normalising flows

Observations of neutron stars may be used to study aspects of extremely dense matter, specifically a possibility of phase transitions to exotic states, such as de-confined quarks. We present a novel data analysis method for detecting signatures of dense-matter phase transitions in sets of mass-radius measurements, and study its sensitivity with respect to the size of observational errors and the number of observations. The method is based on machine learning anomaly detection coupled with normalizing flows technique: the algorithm trained on samples of astrophysical observations featuring no phase transition signatures interprets a phase transition sample as an ''anomaly''. For the sake of this study, we focus on dense-matter equations of state leading to detached branches of mass-radius sequences (strong phase transitions), use an astrophysically-informed neutron-star mass function, and various magnitudes of observational errors and sample sizes. The method is shown to reliably detect cases of mass-radius relations with phase transition signatures, while increasing its sensitivity with decreasing measurement errors and increasing number of observations. We discuss marginal cases, when the phase transition mass is located near the edges of the mass function range. Evaluated on the current state-of-art selection of real measurements of electromagnetic and gravitational-wave observations, the method gives inconclusive results, which we interpret as due to small available sample size, large observational errors and complex systematics.

astro-ph.HE↗

arXiv2022

Differentiating between sharp and smoother phase transitions in neutron stars

The internal composition of neutron stars is still an open issue in astrophysics. Their innermost regions are impervious to light propagation and gravitational waves mostly carry global aspects of stars, meaning that only indirect inferences of their interiors could be obtained. Here we assume a hypothetical future scenario in which an equation of state softening due to a phase transition is identified and estimate the observational accuracy to differentiate a sharp phase transition from a smoother one (associated with a mixed phase/state due to the unknown value of the surface tension of dense matter) in a region of a hybrid star by means of some electromagnetic and gravitational wave observables. We show that different transition constructions lead to similar sequences of stellar configurations due to their shared thermodynamic properties. In the most optimistic case - a strong quark-hadron density jump phase transition - radius observations require fractional uncertainties smaller than $1\%-2\%$ to differentiate mixed states from sharp phase transitions. For tidal deformabilities, relative uncertainties should be smaller than $5\%-10\%$. However, for masses around the onset of stable quark cores, relative tidal deformability differences associated with strong sharp phase transitions and mixed states could be much larger (up to around $20\%-30\%$). All the above suggests that 2.5- and 3rd generation gravitational wave detectors and near-term electromagnetic missions may be able to start assessing some particular aspects of phase transitions in neutron stars. In addition, it points to some limitations on the equation of state recovery using typical neutron star observables and the impact of systematic uncertainties on modellings of the equation of state of hybrid stars. Finally, we briefly discuss other observables that may also be relevant for the probe of mixed states in stars.

astro-ph.HE↗

arXiv2021

Anomaly Detection in Gravitational Waves data using Convolutional AutoEncoders

As of this moment, fifty gravitational waves (GW) detections have been announced, thanks to the observational efforts of the LIGO-Virgo Collaboration, working with the Advanced LIGO and the Advanced Virgo interferometers. The detection of signals is complicated by the noise-dominated nature of the data. Conventional approaches in GW detection procedures require either precise knowledge of the GW waveform in the context of matched filtering searches or coincident analysis of data from multiple detectors. Furthermore, the analysis is prone to contamination by instrumental or environmental artifacts called glitches which either mimic astrophysical signals or reduce the overall quality of data. In this paper, we propose an alternative generic method of studying GW data based on detecting anomalies. The anomalies we study are transient signals, different from the slow non-stationary noise of the detector. Presented in the manuscript anomalies are mostly based on the GW emitted by the mergers of binary black hole systems. However, the presented study of anomalies is not limited only to GW alone, but also includes glitches occurring in the real LIGO/Virgo dataset available at the Gravitational Waves Open Science Center.

astro-ph.IM↗