WorksPhysics and Astronomy
A gaussian-process reconstruction of gravitational-wave friction from dark standard sirens
Abstract
Standard sirens provide a direct probe of the cosmic distance-redshift relation and therefore a powerful test of whether gravitational waves (GWs) propagate as predicted by general relativity on cosmological scales. In many extensions of general relativity, GW propagation is modified by an anomalous friction term, which alters the amplitude evolution and leads to a GW luminosity distance that differs from its electromagnetic counterpart. In this work, we develop a non-parametric reconstruction of this anomalous friction from dark standard sirens. We reconstruct its redshift dependence, encoded in the friction function , with a Gaussian process (GP) and derive the corresponding GW-to-EM luminosity-distance ratio by integration, explicitly retaining the relation between the local friction modification and its accumulated effect on the observed distance, while enforcing a monotonic and invertible GW luminosity distance. We apply the method to the GWTC-5.0 catalog, jointly inferring the source population, modified propagation, and background cosmological parameters. We find no evidence for departures from GR, with and consistent with the posterior over the redshift range probed by the catalog. The result is stable under variations of the GP kernel, smoothness and amplitude priors, resolution prescription, and likelihood treatment. We further introduce function-space diagnostics for comparing flexible and parametric reconstructions, including prior calibration, prior-to-posterior contraction, common summaries of propagation-population degeneracies, and principal-component analysis. These tools make explicit which features are informed by the data and which are imposed by the function-space prior, providing a framework for interpretable non-parametric tests of gravitational-wave propagation.
Introduction
Gravitational-wave (GW) observations enable “standard siren” cosmology [1, 2]: direct measurements of luminosity distance from compact-binary signals can be combined with redshift information to infer the expansion history of the Universe. A powerful route is provided by spectral sirens, where features in the source-frame mass distribution of merging black holes break the intrinsic mass–redshift degeneracy of the GW waveform [3, 4, 5, 6]. State-of-the-art analyses jointly infer the BBH population and cosmological parameters, yielding independent measurements of the Hubble constant [7, 8, 9].
While much of the standard-siren program has focused on measuring , the same distance–redshift information also provides a direct test of GW propagation [10, 11, 12, 13, 14, 15, 16]. This is especially relevant for the physics of cosmic acceleration and dark energy, whose origin remains one of the central open questions in cosmology and is further sharpened by recent large-scale structure measurements, e.g. [17]. If gravity departs from GR on cosmological scales, or if dark energy is dynamical, the amplitude evolution of GWs can differ from the GR prediction through an additional friction term in the propagation equation [10, 11, 12, 13, 14, 15, 16]. The distance inferred from the GW amplitude is then not the standard electromagnetic (EM) luminosity distance, but a GW luminosity distance [14, 15, 18]. Equivalently, modified propagation can be described by the GW-to-EM distance ratio , or by the GW extra-friction function . This effect acts on top of possible modifications to the background expansion and accumulates along the propagation, making it directly constrainable with standard sirens [15].
Existing constraints on modified GW propagation [7, 19, 20, 21, 22, 23, 24] have mostly adopted theory-motivated parametric forms. One common approach parameterizes directly the GW-to-EM distance ratio, as in the – ansatz [15]. This form captures broad classes of luminal modified-gravity models, but does not span the full theory space, including DHOST, extra-dimensional, and bigravity scenarios [18, 25, 26, 27]. A second approach parameterizes the friction function itself, which is directly connected to the Planck-mass running in Horndeski gravity [28] and appears more generally as a free function in the effective field theory of dark energy [29]. In practice, one often assumes that tracks the onset of dark energy, for example by making it proportional to the background dark-energy density [7, 16, 17, 20, 23, 24, 30, 31, 32, 33]. This choice is well motivated but not unique, and the impact of such functional assumptions on the resulting constraints remains debated [34, 35, 36, 37].
This motivates a more general strategy: bypassing a fixed parametric form and allowing the data to inform the redshift dependence of modified GW propagation directly. Nonparametric, or weakly modelled, reconstructions provide a natural framework for this goal; in particular, see Ref. [38] for a reconstruction of the Hubble function from GW data. In the context of modified GW propagation and standard-siren techniques, related approaches have been first suggested in a forecasting context using Gaussian processes with bright sirens [39]. Only recently a Fourier-basis reconstruction of the GW-to-EM luminosity-distance ratio with spectral sirens was obtained from LVK catalogs [40]. In this work we suggest a complementary route and introduce a Gaussian-process reconstruction tailored directly to GW friction. Instead of modelling the GW-to-EM distance ratio as the primary object, we reconstruct the friction function and derive the distance ratio from it. This follows the physical structure of the propagation problem: describes the local modification to GW propagation, while is the cumulative effect measured by standard sirens. The approach therefore keeps the connection between the propagation equation and the observed luminosity-distance modification explicit, while still allowing a broad range of redshift dependences. After validating the method on simulated cases, we apply it to the GWTC–5.0 BBH catalog [41] and reconstruct both and .
A central focus of this work is the interpretation and robustness of the non-parametric reconstruction. We therefore vary several GP modelling choices, which act as effective changes of the function-space prior, and test whether the inferred propagation history is stable. In addition, we introduce a set of diagnostics to quantify how the data update that prior and to enable a controlled comparison with parametric models. These include a calibration of the GP amplitude in function space, prior-to-posterior contraction measures, common function-space summaries for studying degeneracies, and a principal-component analysis of the posterior covariance. We use these tools to compare with the standard – parameterization and to distinguish information supplied by the catalog from structure imposed by the assumed functional form.
Gaussian–Process Reconstruction of Modified Gravitational–Wave Propagation
In several extensions of GR, GWs propagate with an effective friction term that differs from the one implied by the standard background expansion. Here, we consider the propagation equation for the tensor modes
where primes denote derivatives with respect to conformal time , is the conformal Hubble parameter, labels the two tensor polarizations, and is the comoving wavenumber. The function modifies the friction experienced by GWs during propagation. In Horndeski gravity, with matter minimally coupled to the Jordan-frame metric, it is the logarithmic running of the effective Planck mass, , where is the coefficient of the tensor kinetic term [13, 28]. (1) corresponds to the luminal, massless and source-free limit of a more general tensor propagation equation; we assumed , , and , where is the tensor propagation speed, an effective tensor mass, and a possible source term, for instance from anisotropic stress [18, 42]. The assumption is motivated by the multimessenger observation of GW170817, which constrains the relative difference between the GW and electromagnetic propagation speeds to approximately one part in [43, 44].1 A more fundamental issue is relevant to the luminosity assumption. The currently employed spectral-siren likelihood uses the standard mass relation . This relation follows from the standard cosmological time dilation . For homogeneous propagation with a time-dependent speed , successive wavecrests instead satisfy (see e.g. Appendix A of Ref. [18]), which implies a modified mass scaling. Allowing an arbitrary distance-ratio function does not account for this modified conversion. Consequently, any constraint based on a likelihood making use of the standard source-to-detector frame-mass relation does not cover time-dependent nonluminal propagation. Such an extension requires a joint model for mass redshifting and amplitude propagation; we leave this for future work.
The friction term modifies the relation between the observed GW amplitude and the source redshift, and can be described by allowing the GW luminosity distance, , to differ from the EM luminosity distance, . Possible deviations between the GW and EM luminosity distances are modeled through the distance-ratio function
Under the propagation assumptions adopted above, the anomalous friction is the only modification of GW propagation, and the GW-to-EM luminosity-distance ratio is directly related to as
Our reconstruction should therefore be understood as a non-parametric reconstruction within the luminal, massless, friction-modified sector of GW propagation [18].
For the EM luminosity distance we assume a spatially flat CDM background,
where
Here, is the present matter density fraction, and is the dark energy equation-of-state parameter. We also define for later use the logarithmic EM-distance slope
A widely used phenomenological parametrization for modified GW propagation is the two-parameter – model [14, 15, 18],
This captures a transition from GR at low redshift to an asymptotic constant ratio at high redshift. Alternatively, one may parametrize directly, for instance by assuming that it scales with the dark-energy density, [16, 28]. The Gaussian-process reconstruction introduced below generalizes these parametric families by inferring the redshift dependence of , and therefore of , directly from the data.
Monotonic Gaussian-process construction
GW observations constrain luminosity distance rather than redshift. The likelihood therefore requires in general the inverse map , which is well defined only if is strictly increasing over the redshift range of interest. We enforce this condition by modeling
through a positive reparametrization.
We introduce a latent Gaussian process in the coordinate
with
where is the covariance kernel, and define
Here . This mapping guarantees and hence a monotonic .
The deterministic centering function is chosen so that the prior on and is centered on GR. Specifically,
so that implies
and therefore .
Combining Eqs. (2), (3), (8), and (10), we obtain
From Eqs. (10) and (16), the monotonicity condition implies
This bound is independent of , because the explicit dependence on cancels in the logarithmic derivative defining .
Given , or equivalently , the distance ratio is reconstructed by integrating Eq. (4). We impose the local boundary condition .
Resolution-adapted GP prior
Our baseline prior uses a non-stationary Gibbs kernel in . A stationary kernel assumes a single correlation length across the full redshift range, so that the prior allows the same characteristic variation scale at low and high redshift. A Gibbs kernel instead promotes the correlation length to a function of position, , and is therefore suited to a problem in which the resolving power of the data changes with redshift. This is the case for GW distances: at fixed intrinsic parameters, the fractional distance uncertainty scales approximately as the inverse signal-to-noise ratio, and nearby, louder events therefore provide better distance, and hence redshift, resolution than distant events. We use the Gibbs kernel to encode this information in the prior, allowing shorter correlation lengths where the catalog has better redshift resolution and enforcing smoother behavior where it does not.
Specifically, we write the local correlation length as
where is a dimensionless smoothness multiplier and is a data-driven resolution curve.
The resolution curve is constructed from the single-event distance measurement. For each event, we draw a fixed number of samples from its luminosity distance posterior, and convert these distances to redshifts using a fixed reference cosmology with GR propagation. This gives, for event , a set of reference samples and hence . We assign the event to the median redshift coordinate
and define its redshift-resolution width as half of the central 68% interval in ,
The pairs are then divided into a fixed number of equally spaced bins in . In each bin containing a minimum number of events, the bin center is taken to be the mean of the values in that bin, and the resolution value is taken to be a chosen quantile of the corresponding values. In the baseline analysis we use 1000 reference samples per event, 15 requested bins, a minimum of five events per retained bin, and the quantile of the event widths. Thus the baseline resolution curve tracks the better-measured events in each redshift bin. Variations using a smoothed resolution curve and the median width, , are discussed in § 1.4 D below.
The Gibbs covariance is2
As a robustness test, we also consider a stationary Matérn-5/2 kernel,
The Matérn length-scale prior is chosen to give a smoothness scale comparable to the baseline Gibbs model in the redshift range most constrained by the data. This kernel-family variation, together with other changes of the baseline configuration, is discussed in § 1.4 D below.
a. Amplitude prior. We adopt a penalized-complexity (PC) prior [45] for the GP amplitude ,
so that
This prior shrinks the reconstruction toward the GR limit, , while allowing deviations when supported by the data. The numerical values of and are specified together with the baseline configuration in the results section.
b. Smoothness prior. For the baseline Gibbs model, a smoothness prior is placed on the multiplier in Eq. (18). We use the same PC-prior motivation as for stationary Matérn fields [46]: short length scales are allowed, but penalized unless required by the data. In one dimension this gives the Fréchet form
We set by imposing
The cumulative distribution corresponding to Eq. (26) is
which gives
We additionally include a weak exponential penalty against very large values of ,
This avoids the nearly constant-kernel regime, which can lead to stiff geometries and inefficient sampling, without imposing a hard upper cutoff. We parametrize the strength of this penalty through an interpretable upper smoothness scale,
The adopted baseline values and the variations used to test robustness are given in § 1.4 D below.
Hierarchical Inference and Population Model
To describe the BBH population underlying the catalog, we adopt a probability density for source-frame component masses, spins, and redshift, , where denotes the physical source parameters and the corresponding population and cosmological hyperparameters. Our primary-mass model follows the default GWTC-4.0 and GWTC-5.0 BBH population model [47, 48], namely a broken power law plus two Gaussian components.3 The presence of mass scales in the population statistically breaks the mass–redshift degeneracy in GW data and enables cosmological inference. The secondary mass is modeled as a conditional power law in , normalized at fixed and restricted to . The mass model is assumed to be redshift-independent (see [49] for a demonstration of the limited impact of redshift evolution). The merger-rate density is described by a Madau–Dickinson functional form in redshift. In this analysis we do not introduce an explicit spin-population model.4
The full set of parameters and latent variables is denoted by and grouped as
For the background cosmology we sample
For modified GW propagation, the baseline GP model contains two hyperparameters , together with a finite set of latent GP coefficients. In the baseline configuration we represent the GP on nodes in , sampling the corresponding whitened latent variables and transforming them through the GP covariance.5 Therefore,
For the population, the Madau–Dickinson merger-rate density has three free parameters , while the mass model uses 15 free hyperparameters .
Thus, the model has
degrees of freedom.
The hyperparameters are inferred using hierarchical Bayesian inference [50, 51]. The analysis is based on a rate-marginalized likelihood that incorporates the impact of selection effects. Detection efficiency is estimated through reweighted Monte Carlo integration over simulated signals that are recovered by the search pipeline [52, 53] with default checks on the Monte Carlo variance of the estimator [54, 55].
Posterior sampling is performed with an extended implementation of the python package pymcpop-gw [56], introduced in Ref. [57]. The hierarchical posterior is explored using the No-U-Turn Sampler (NUTS) [58, 59] within numpyro [60, 61], with GPU acceleration. We monitor the convergence of the chains using the Gelman–Rubin statistic [62], requiring for all sampled parameters.
A complete description of the likelihood and its numerical implementation is provided in Appendix B. Priors are specified in Appendix § 1.4.3.
Validation on simulations
We first demonstrate that the GP construction is able to pinpoint deviations from GR that are described by more complex functional forms than those covered by standard parameterizations, while also correctly recovering GR in case that is the true underlying model.
We generate simulated catalogs of binary black hole (BBH) mergers assuming the same mass and redshift distributions used in the population model described in App. A, namely the default "Broken Power-Law plus Two Peaks" of the LVK GWTC-4.0 population paper. Fiducial values correspond to the median posterior of the corresponding data release. For simplicity we neglect spin degrees of freedom in the simulated sources.
We consider two cosmological scenarios governing gravitational-wave propagation. The first corresponds to standard CDM, in which the GW luminosity distance coincides with the EM distance. The second scenario is based on a modified-gravity model that cannot be described by the usual ansatz or by commonly adopted low-dimensional parameterizations of . Specifically, we adopt a model representative of DHOST theories [25], discussed in Sec. 3.1.2 of Ref. [18], using the polynomial–exponential (“polexp”) parameterization introduced there. In this framework modified GW propagation is described through
which leads to the friction term
as given in Eqs. (3.22)–(3.23) of Ref. [18]. For illustration we choose parameters corresponding to a large deviation from general relativity, and . While not completely ruled out by current observations, these are already in mild tension with current constraints [33]. This intentionally extreme choice provides a useful stress test of the reconstruction method, demonstrating that the Gaussian-process model can depart significantly from the GR-centered prior when supported by the data.
Mock observations and injection sets are generated assuming the LVK O5 design sensitivity. We use a modified version of the GWMocKCat package [63]. We analyze 3000 events, in order to make the effect of modified gravity distinguishable and to stress test the recovery of GR in a regime where the statistical uncertainty might not be large enough to mask possible residual biases induced by the GP construction.
Figure 1 shows the reconstruction of (left) and (right) for the fiducial CDM scenario (top) and for the polexp modified-propagation scenario (bottom). For these tests we adopt the same GP amplitude setting,
Figure 1. Reconstruction of GW friction (left) and the GW-to-EM distance ratio (right) from 3000 events at O5 design sensitivity. Shaded bands represent prior and posterior credible levels (colors in legend). Top panels: fiducial model is GR (dotted black line). Bottom panels: fiducial model is the DHOST-polexp theory described in the text (solid yellow line).
, used in the GWTC–5.0 analysis described in the next section. Dark shaded bands show the 50% and 90% posterior credible intervals, while the lighter region shows the 90% prior range. In the modified scenario, the injected polexp functions are shown in dark yellow.
For the GR injection, the reconstruction remains centered around and , with the posterior contracting relative to the prior over the redshift range where the simulated catalog is informative. For the polexp injection, the GP is able to depart substantially from its GR-centered prior and follows the main redshift dependence of both the local friction function and the accumulated distance ratio. As expected, is reconstructed less tightly and exhibits more local freedom, whereas the integrated quantity is smoother and more directly constrained.
These simulations are intended as validation tests of the reconstruction rather than forecasts for a particular observing run or estimates of the detectability of modified gravity. The catalog size, detector sensitivity, and injected departure from GR were chosen to provide informative tests of whether the method can recover both the GR limit and a non-standard propagation history outside the usual – functional form. A dedicated forecast of detection sensitivity is beyond the scope of this work.
Results with GWTC-5.0
We now turn to applying the GP reconstruction to data from the GWTC-5.0 catalog [41, 64].
Data
We analyze the GWTC-5.0 catalog following the event selection of the GWTC-5.0 population analysis [48]. Specifically, we use BBH events satisfying FAR in the LVK search pipelines. We use samples obtained with the IMRPhenomXPHM waveform model for events in O1 to O3 runs, and IMRPhenomXPHM_SpinTaylor for O4a-O4b events. The data products are available at [65, 66, 67, 68]. Selection effects are computed from the GWTC-5.0 injection campaigns, available at [69] (see [70] for a more complete description and [71, 72] for additional details).
Prior calibration
We calibrate the GP amplitude prior to the redshift range effectively probed by this data set. The GP prior uses the resolution curve inferred from the catalog, and we choose the amplitude hyperparameter by matching the integrated function-space prior width to that induced by the standard – parameterization. The priors adopted in this work have support over and , with sampled through a bounded transformation in logarithmic space. The specific prior distributions and their implementation are given in Appendix C.6 We calibrate the GP amplitude to obtain a comparable integrated prior width in event-supported log space.
This calibration matches the overall amplitude of the functional uncertainty, but does not impose identical prior distributions or correlations across redshift. We define equal-populated event-redshift bins from the event-redshift posteriors, obtained with a reference Planck cosmology. For each prior realization we compute
over the redshift interval , and summarize the prior width in each bin by
We then define the integrated prior-width metric
Scanning the GP amplitude hyperparameter over
we select the largest scanned value of for which does not exceed the corresponding value. This gives for the baseline analysis.
All other GP hyperparameters are kept fixed to the baseline Gibbs-resolution configuration: , , , , . Figure 2 shows the resulting catalog-calibrated GP prior together with the induced – prior.

Figure 2. Prior predictive distribution for the GWTC–5 analysis for the GW friction (left) and the GW-to-EM luminosity-distance ratio (right). Blue bands and thin blue curves show the GP prior predictive distribution and individual GP prior realizations for the selected amplitude hyperparameter . Purple bands and dashed purple curves show the induced prior of the standard – parameterization. The hatched region in the panel is excluded by the monotonicity condition.
The hatched region in the panel of Figure 2 shows the region excluded by monotonicity of the GW luminosity distance, as quantified by Eq. (17).
Results
Figure 3 shows the reconstruction of modified GW propagation from the GWTC–5 catalog. The left panel shows the friction function , while the right panel shows the accumulated distance ratio . The GP posterior is consistent with GR over the full event-supported redshift range: both and lie well within the posterior credible regions. The posterior is nevertheless visibly contracted relative to the prior, especially for , which is the quantity directly constrained by the luminosity-distance information.
Figure 3. Posterior reconstruction of modified GW propagation from the GWTC–5 catalog, compared with the standard – parameterization. The left panel shows the GW friction function , and the right panel shows the GW-to-EM luminosity-distance ratio . Blue bands show the calibrated GP posterior credible regions, while the pale blue band shows the GP prior predictive 90% region. Thin blue curves are GP posterior draws. Purple bands and the solid purple curve show the posterior induced by the – parameterization. The dashed dark-blue curve marks the GP posterior median, and the dotted black line denotes the GR prediction, and .
The two panels also illustrate the difference between the local and accumulated descriptions of modified propagation. The friction function is a local quantity and therefore allows larger pointwise fluctuations. By contrast, is obtained by integrating the propagation modification and is correspondingly smoother.
The same figure compares the GP reconstruction with the posterior obtained from the standard – parameterization. The two reconstructions are mutually consistent and both are compatible with GR. However, they distribute the allowed departures from GR differently. The – model induces a smooth, coherent trend in and correspondingly gives a positive already at low redshift. This behaviour is a consequence of the parameterization: is not independently reconstructed in the – model, but is derived from the assumed global functional form of . The GP reconstruction, instead, allows more local variation through the prior on the latent field governing , constrained through the integrated distance relation. This allows the friction function to remain closer to GR near , while still permitting large accumulated deviations in the distance ratio.
| Model | |||
| GP | 1.110 | 0.213 | 5.21 |
| – | 1.264 | 0.448 | 2.82 |
Table 1. Integrated prior and posterior widths in event-binned log space. The GP amplitude prior was calibrated so that the GP and – priors have comparable integrated widths in this metric. After conditioning on the GWTC–5 data, the GP posterior contracts more strongly and has an integrated posterior width about a factor of two smaller than the – posterior.
a. Integrated width and prior-to-posterior contraction. To quantify the information gained by the data, we use the same event-binned log metric introduced for the prior calibration. For each draw we use the quantities defined in Eqs. (38)–(39)–(40). In particular, measures the typical prior or posterior uncertainty in the event-supported part of .
The result is reported in Table I. The two priors are comparable in the calibration metric, with for the GP and for –. After conditioning on the data, however, the posterior widths differ more substantially: for the GP and for –. Thus, after prior calibration, the GP contracts by a factor , compared to for the – model. This shows that the smaller GP posterior width is not simply the result of assigning it a narrower prior amplitude. Instead, the two models organize the information in the catalog differently.
b. Propagation–rate degeneracy. A useful way to understand this difference is to inspect how modified propagation correlates with the population model. In the – parameterization, one can directly look at correlations involving the parameter . For the GP there is no analogous single propagation parameter. We therefore compress each reconstructed function into a common function-space summary,
where the two averages are taken over the highest and lowest event-supported redshift bins. This quantity measures the net redshift tilt of the accumulated propagation modification across the range where the catalog carries information. In the limit , , the summary reduces in the – parameterization to , independently of .
Figure 4 shows the joint posterior of this summary with the Madau–Dickinson rate parameter and with . Both reconstructions show a strong propagation–rate degeneracy: a change in the redshift evolution of the source population can be partially compensated by a change in the redshift dependence of the GW distance ratio. This degeneracy is therefore not removed by the GP. Rather, the GP expresses it in a function-space variable instead of a single parametric amplitude. Numerically, the correlations with the propagation tilt are comparable in the two models,
Figure 4. Function-space posterior corner for the GP and reconstructions. The propagation variable is , the difference between the high- and low-redshift averages of over the event-supported range. We compare this common propagation summary with the Madau–Dickinson rate parameter and with . Both models show the same qualitative propagation–rate degeneracy direction.
while the correlation with is weaker for the GP,
The important point is therefore not that the GP eliminates the degeneracy, but that the degeneracy does not force the GP posterior into the same global shape as the parameterization.
Function-space covariance structure
To understand more generally how posterior uncertainty is distributed across redshift, we perform a principal-component analysis (PCA) of the covariance matrix of the event-binned variables . This asks whether the posterior uncertainty is effectively controlled by a single coherent mode, or whether it contains additional independent directions in function space.
The PCA results are summarized in Table II. Let denote the eigenvalues of the posterior covariance matrix, ordered from largest to smallest. The entries labelled PC1 and PC1+PC2 give the fractions of the total posterior variance carried by the first principal component and by the first two components, respectively,
| Model | PC1 | PC1+PC2 | ||
| GP | 0.883 | 0.972 | 1.56 | 0.761 |
| 0.992 | 1.000 | 1.05 | 0.885 |
Table II. PCA summary of the posterior covariance of the event-binned variables . PC1 and PC1+PC2 give the fraction of total posterior variance explained by the first one and two principal components. The effective rank is computed from the entropy of the eigenvalue spectrum. The posterior is nearly rank-one in this function space, while the GP retains a higher-rank covariance structure.
Thus, values close to unity indicate that the posterior uncertainty is concentrated in only one or two coherent redshift-dependent modes.
To summarize the full eigenvalue spectrum with a single number, we also report the effective rank
This quantity can be interpreted as the effective number of principal components that contribute appreciably to the posterior covariance: corresponds to an exactly rank-one covariance, while would correspond, for example, to two equally important modes. It need not be an integer. Finally, is the posterior correlation between the amplitude of the leading principal component and the Madau–Dickinson rate parameter . It therefore measures how strongly the dominant propagation mode is coupled to the inferred redshift evolution of the merger rate.
For the model, the first principal component accounts for 99.2% of the posterior variance, and the effective rank is . This means that the posterior uncertainty is almost entirely a single coherent deformation of across all redshift bins. This is the expected behaviour for a fixed two-parameter functional form.
The GP posterior is also dominated by a leading mode, but less extremely: the first component accounts for 88.3% of the variance, the first two components for 97.2%, and the effective rank is . The GP therefore retains additional function-space directions beyond a single global tilt. This provides a quantitative explanation for the stronger contraction reported in Table I: after calibrating the prior amplitude, the GP can allocate posterior freedom across redshift more flexibly, whereas the model ties the accumulated distance modification to a nearly rank-one global shape.
Overall, the calibrated GP and the parameterization give consistent reconstructions and both are compatible with GR. The difference is in how the available information is organized. The model provides a compact global description of the accumulated distance ratio. The GP reconstruction keeps the physical relation between local friction and accumulated distance explicit and allows the posterior uncertainty to occupy a richer, though still data-constrained, function space.
Robustness to GP modelling choices
We test the stability of the GWTC–5 reconstruction by repeating the analysis under several controlled changes of the GP model, which probe different aspects of the effective function-space prior. Table III summarizes the robustness runs. Each variant changes one aspect of the GP reconstruction relative to the baseline, and is intended to test whether the inferred propagation history is stable under the corresponding modelling choice.
| Variant | Change | Test |
| Baseline | Calibrated Gibbs GP | Reference run |
| Wider | Broader smoothness prior | Smoothness sensitivity |
| Smoothed res. | Smoothed resolution curve | Resolution modelling |
| Median res. | Median resolution estimate | Resolution quantile |
| Matérn-5/2 | Stationary Matérn kernel | Kernel choice |
| Broader amplitude prior | Prior amplitude | |
| Lower | Shorter correlation scale | Local redshift structure |
| Marginal likelihood | Alternative likelihood | Likelihood implementation |
Table III. Robustness variations used for the GWTC–5 GP reconstruction. Each run changes one modelling choice relative to the calibrated baseline.
Figure 5 and Figure 6 show the result of this scan. In each case we compare the variant to the calibrated baseline reconstruction and to the – posterior. The distance-ratio reconstruction is the most stable quantity. Across all variants, the GP median for remains close to the calibrated baseline and close to GR, with differences that are small compared to the posterior uncertainty. The largest changes appear in variants that broaden or locally loosen the GP prior, such as the lower- and runs, but even in these cases the posterior remains consistent with the baseline and with .
Figure 5. Robustness of the reconstructed distance ratio under GP modelling variations. Each panel shows the calibrated baseline GP posterior in gray, one GP variant in color, and the posterior in red. Bands show the 68% credible regions and curves show posterior medians. The curve is repeated in each panel as a fixed parametric reference. The GP reconstruction of is stable across the variations and remains consistent with GR.
Figure 6. Robustness of the reconstructed GW friction function under GP modelling variations. The plotting convention is the same as in Figure 5: gray denotes the calibrated baseline GP reconstruction, colored bands and curves denote the GP variant, and red denotes the – posterior. The local quantity is more sensitive than to smoothness and likelihood choices, but the main conclusion remains unchanged: the reconstructions are compatible with GR and with one another over the event-supported redshift range.
The friction function is more sensitive to these choices, as expected for a local derivative-like quantity. Variants that allow shorter correlation scales or modify the likelihood treatment can produce more localized structure in , especially at low and intermediate redshift. However, all variations remain compatible with GR. Furthermore, these features do not translate into a qualitatively different accumulated distance ratio. This is important: the quantity most directly constrained by the luminosity-distance data is , and its posterior is stable under the GP modelling variations.
The – posterior is shown as reference in the same panels. It is consistent with the GP reconstructions in all cases. The robustness scan supports the interpretation of the main result: the GWTC–5 data are consistent with GR, and this conclusion is not driven by a specific GP kernel, resolution prescription, or amplitude choice.
Conclusions
In this work we have developed a non-parametric reconstruction of modified gravitational-wave propagation based on Gaussian processes. Rather than parametrizing
directly the accumulated GW-to-EM luminosity-distance ratio , we reconstruct the local friction function and obtain by integration. The construction is formulated so that the GW luminosity distance remains monotonic and therefore invertible, while the non-stationary GP kernel adapts its characteristic correlation scale to the redshift resolution of the catalog.
We validated the method on simulated catalogs containing both a GR fiducial and a non-trivial polexp propagation history that lies outside the standard – functional form. In both cases the reconstruction recovers the injected behavior within the posterior uncertainty, showing that the GR-centered GP prior can accommodate significant and non-standard departures when supported by sufficiently informative data.
We then applied the method to the GWTC–5 BBH catalog, jointly inferring the BBH population, the modified-propagation history, and the background cosmological parameters , , and . We find no evidence for departures from GR: both and remain consistent with the posterior over the redshift range probed by the catalog. This conclusion is stable under variations of the GP kernel, the correlation-scale prior, the resolution prescription, the GP amplitude prior, and the likelihood treatment. In particular, while the local reconstruction of is naturally more sensitive to these modelling choices, the accumulated distance ratio is remarkably stable.
Our results complement the recent weakly modelled spectral-siren analysis of Ref. [40], which reconstructs the distance ratio directly using a Fourier basis and likewise finds no evidence for modified propagation in GWTC–5. The two approaches differ in emphasis. Here the primary reconstructed quantity is the local friction , the background cosmology is varied jointly with the propagation and population parameters, including , and particular attention is devoted to the dependence of the inference on the choice of function-space prior and reconstruction settings.
More generally, an important aspect of this work is the development of diagnostics that allow a flexible reconstruction to be compared with a parametric model on a common footing. In particular, we have used:
a calibration of the GP amplitude prior through the integrated prior width of event-binned log , so that
the comparison with standard parametric models is not driven simply by different prior volumes;
the corresponding prior-to-posterior contraction in the same function-space metric, which shows that, after this calibration, the GP and reconstructions undergo different amounts of posterior contraction;
a common function-space summary, , which makes it possible to compare propagation–population correlations even though the GP has no single parameter analogous to ; in the limit where the low- and high-redshift averages reduce to and , respectively, this summary becomes in the parameterization.
a principal-component analysis of the event-binned covariance, which shows how the two models distribute posterior uncertainty across independent redshift-dependent modes.
These diagnostics reveal that the propagation–rate degeneracy remains present in the flexible reconstruction. The difference is instead in the geometry of the allowed propagation histories: the posterior is restricted to a single coherent mode, whereas the GP retains additional independent directions in function space. We expect this type of comparison to be useful more generally when assessing whether constraints obtained with a fixed functional ansatz are driven by the data or by the structure imposed by the parameterization.
The applicability of the present analysis is explicitly restricted to luminal, massless, source-free tensor propagation, with anomalous friction as its only modification and standard GR binary waveform generation. This restriction belongs to the adopted likelihood: the standard spectral-siren mass conversion does not hold for generic homogeneous time-dependent . Functional flexibility in alone therefore does not extend the analysis to those theories. This restriction applies to any constraint on the GW-to-EM distance ratio obtained with current spectral-siren techniques. A nonluminal extension requires joint modelling of mass redshifting and amplitude propagation. This would provide a possible next step, even though deviations from luminality are disfavoured, at least at low redshift, by the stringent GW170817 constraint. Within the adopted luminal sector, our reconstruction remains much more general than parametric approaches in the redshift dependence of physical GW friction.
A further extension would be to relax the assumption that is monotonic. Monotonicity is required here because the likelihood is evaluated through the inverse map . More general likelihood constructions can avoid this inverse map by drawing proposal samples directly in redshift space, as in [73], and could therefore in principle accommodate non-monotonic distance–redshift relations. The treatment of selection effects however becomes less straightforward. Standard injection campaigns provide finite sets of signals selected according to their detectability in detector-frame parameter space, where luminosity distance rather than redshift determines detectability. They are therefore not immediately adapted to a likelihood formulated entirely through redshift-space proposals when the distance–redshift map is not invertible. A possible route would be to evaluate the detection probability continuously and on the fly for proposed sources, using selection-function emulators such as that of Ref. [74]. Such an extension would permit substantially more general propagation histories, but would require a dedicated likelihood and selection-function design as well as significantly greater computational development.
The increasing size and precision of GW catalogs will make these aspects progressively more important. Flexible reconstructions provide a way to test GW propagation without committing to a specific redshift dependence, but their interpretation requires explicit control of the function-space prior and quantitative diagnostics of how the posterior departs from and contracts relative to that prior. In this work we have introduced and applied several concrete tools for this purpose. Such prior–posterior checks are essential for distinguishing information supplied by the data from structure imposed by a non-parametric reconstruction, while retaining physical interpretability.
Data availability
The population inference code used for this work is available at Ref. [56]. The version used for this paper will be archived on Zenodo. Data products for events in the GWTC-5.0 catalog are available at [65, 66, 67]. Sensitivity estimates are available at [69].
Acknowledgments
We thank the LVK Tests of CDM group members, Simone Mastrogiovanni, and Grégoire Pierra for useful discussions. We also thank Elena Colangeli for internal LVK review. This work received support from the French government under the France 2030 investment plan, as part of the Initiative d’Excellence d’Aix-Marseille Université – A*MIDEX AMX-22-CEI-02. We acknowledge use of computational resources of the MesoCentre of Aix-Marseille University through the allocation b435-GWCROSS.
This research has made use of data or software obtained from the Gravitational Wave Open Science Center (gwosc.org), a service of the LIGO Scientific Collaboration, the Virgo Collaboration, and KAGRA. This material is based upon work supported by NSF’s LIGO Laboratory which is a major facility fully funded by the National Science Foundation, as well as the Science and Technology Facilities Council (STFC) of the United Kingdom, the Max-Planck-Society (MPS), and the State of Niedersachsen/Germany for support of the construction of Advanced LIGO and construction and operation of the GEO600 detector. Additional support for Advanced LIGO was provided by the Australian Research Council. Virgo is funded, through the European Gravitational Observatory (EGO), by the French Centre National de Recherche Scientifique (CNRS), the Italian Istituto Nazionale di Fisica Nucleare (INFN) and the Dutch Nikhef, with contributions by institutions from Belgium, Germany, Greece, Hungary, Ireland, Japan, Monaco, Poland, Portugal, Spain. KAGRA is supported by Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan Society for the Promotion of Science (JSPS) in Japan; National Research Foundation (NRF) and Ministry of Science and ICT (MSIT) in Korea; Academia Sinica (AS) and National Science and Technology Council (NSTC) in Taiwan.
Population-model definitions
This appendix gives the explicit functional form of the population model. The model follows the default GWTC-4.0 BBH population prescription of Ref. [47] for the primary-mass mixture structure, with the sole difference that we apply the smoothstep low-mass taper defined in Sec. A 0 c.
We write the source-population density as
where collects source-frame parameters (masses, spins, redshift), and .
a. Primary-mass model: broken power law with two Gaussian peaks The primary-mass spectrum is a mixture of a broken power law and two truncated Gaussians:
defined on , where and . The normalization constant is chosen such that
The broken power law has slopes and around a break mass :
with support restricted to . The Gaussian components are
each truncated to and renormalized on that interval.
b. Secondary-mass model The secondary mass is modeled conditionally on as a power law in with low-mass taper and the physical constraint :
where is the Heaviside step function and the normalization factor
ensures for each .
c. Low-mass tapering (smoothstep) To avoid hard cutoffs and ensure differentiability, we apply a smoothstep taper at the low-mass edge of both and . For any mass variable with low-mass bound and taper width , we define
We apply with and with . This smoothstep prescription differs from the low-mass smoothing used in the GWTC-4.0 population analysis [47] but leaves the remainder of the model unchanged.
d. Auxiliary parametrization for low-mass bounds The low-mass bound of the primary mass is sampled via an auxiliary variable :
so that while preserving a smooth unconstrained sampling geometry. The secondary low-mass bound is constructed conditionally using :
which enforces by construction.
e. Redshift distribution (Madau–Dickinson) The population redshift distribution is
where is the differential comoving volume element in the assumed cosmology and accounts for time dilation. We use a Madau–Dickinson merger-rate density
Statistical framework
This appendix details the statistical framework used to infer the hyperparameters describing the BBH population and the cosmological model. The method follows the hierarchical Bayesian approach introduced in Ref. [50]. Single-event parameter estimation is performed in terms of detector-frame parameters . The data for each event , denoted , are represented through posterior samples drawn from obtained using a reference prior .
For population inference we instead work with source-frame parameters (component masses, spins, and redshift). These are related to the detector-frame quantities through a deterministic mapping which depends on the assumed cosmology via the distance–redshift relation and the conversion between detector- and source-frame masses. Changing variables introduces the Jacobian .
a. Catalog likelihood Let denote the population model for the source parameters. Assuming statistical independence of events, the hierarchical posterior for a catalog of detections is
where denotes the hyperprior. The overall merger-rate normalization has been analytically marginalized assuming a scale-invariant prior. Waveform parameters not explicitly modeled in are effectively assumed to follow the same prior adopted in the single-event parameter estimation.
We sample the hierarchical posterior in Eq. (58) directly in the joint parameter space [57]. In this approach the event-level posterior is approximated with a Gaussian Mixture Model, and the number of mixture components is determined by minimizing the Bayesian Information Criterion, yielding a smooth representation of the posterior density suitable for hierarchical sampling. Further implementation details are provided in Ref. [49, 57]. A blinded mock data challenge demonstrating the equivalence of this strategy with alternative likelihood-evaluation procedures is presented in Ref. [75].
b. Likelihood marginalization In many applications it is convenient to marginalize Eq. (58) over the event-level parameters. The resulting likelihood for the hyperparameters alone becomes
The integrals over are evaluated using posterior samples from the single-event analyses. If denotes the sample set for event , the integral can be approximated through Monte Carlo reweighting as
We adopt this approach as a further validation test. We retain 9500 posterior samples per event to evaluate these integrals.
c. Selection effects The factor represents the detection efficiency, namely the fraction of sources drawn from the population model that would be identified by the search pipeline. We estimate this quantity using simulated injections [52, 53]. If injections are drawn from an injection distribution , the Monte Carlo estimator takes the form
The sum runs over injections recovered by the search pipeline.
d. Monte Carlo uncertainty Both the event-level integrals and the selection-function estimator are computed via Monte Carlo sums, which introduce stochastic noise in the likelihood estimate [53, 54, 55, 76]. Following Refs. [54, 55], the variance of the log-likelihood can be approximated as
Here denotes the estimator of the th event likelihood in Eq. (60), with variance , while and correspond to the injection-based selection estimator and its variance. When the full likelihood of Eq. (58) is sampled directly, only the second term contributes.
To avoid biases from excessive Monte Carlo fluctuations, we check that the commonly adopted requirement [47, 54, 55] remains respected.
Prior distributions
We summarize in Table IV the adopted prior distributions. For compactness, we introduce the notation for the transformed prior
| Parameter | Prior | Parameter | Prior |
| Cosmology and GW propagation | |||
| Merger-rate distribution | |||
| Marginalized | |||
| Primary-mass distribution | |||
| Dirichlet | |||
| Secondary-mass distribution | |||
Table IV. Prior distributions for the GP and – models, using the notation defined in the text. The GP amplitude and smoothness priors are specified in Sec. II B. The mass and merger-rate entries describe the reparametrized sampling branches.
Similarly, denotes
These priors have bounded support but are not uniform.
We denote by a Gaussian distribution whose central 95% interval is . The notation represents a shifted log-normal distribution with lower bound , median , and 95th percentile . Finally, denotes a uniform prior.
For the primary-mass slopes we use , where . The auxiliary variables and follow independent distributions. For the upper-mass cutoff, . This gives a prior median of and a 95th percentile of for .
References
- [1]B. F. Schutz, Nature 323, 310 (1986).
- [2]D. E. Holz and S. A. Hughes, Astrophys. J. 629, 15 (2005), arXiv:astro-ph/0504616 [astro-ph].DOI
- [3]S. R. Taylor, J. R. Gair, and I. Mandel, Phys. Rev. D85, 023535 (2012), arXiv:1108.5161 [gr-qc].DOI
- [4]M. W. Farr, M. Fishbach, J. Ye, and D. Holz, Astrophys. J. Lett. 88, L42 (2019), arXiv:1908.09084 [astro-ph.CO].DOI
- [5]S. Mastrogiovanni, K. Leyde, C. Karathanasis, E. Chassande-Mottin, D. A. Steer, J. Gair, A. Ghosh, R. Gray, S. Mukherjee, and S. Rinaldi, Phys. Rev. D 104, 062009 (2021), arXiv:2103.14663 [gr-qc].DOI
- [6]J. M. Ezquiaga and D. E. Holz, Phys. Rev. Lett. 129, 061102 (2022), arXiv:2202.08240 [astro-ph.CO].DOI
- [7]S. Mastrogiovanni, D. Laghi, R. Gray, G. C. Santoro, A. Ghosh, C. Karathanasis, K. Leyde, D. A. Steer, S. Perries, and G. Pierra, Phys. Rev. D 108, 042002 (2023), arXiv:2305.10488 [astro-ph.CO].arxiv.org/abs/2305.10488
- [8]R. Gray et al., JCAP 12, 023, arXiv:2308.02281 [astro-ph.CO].arxiv.org/abs/2308.02281
- [9]N. Borghi, M. Mancarella, M. Moresco, M. Tagliacozzi, F. Iacovelli, A. Cimatti, and M. Maggiore, Astrophys. J. 964, 191 (2024), arXiv:2312.05302 [astro-ph.CO].DOI
- [10]I. D. Saltas, I. Sawicki, L. Amendola, and M. Kunz, Phys. Rev. Lett. 113, 191101 (2014), arXiv:1406.7139 [astro-ph.CO].DOI
- [11]V. Pettorino and L. Amendola, Phys. Lett. B 742, 353 (2015), arXiv:1408.2224 [astro-ph-CO].
- [12]A. Nishizawa, Phys. Rev. D 97, 104037 (2018), arXiv:1710.04825 [gr-qc].DOI
- [13]L. Amendola, I. Sawicki, M. Kunz, and I. D. Saltas, JCAP 08, 030, arXiv:1712.08623 [astro-ph.CO].DOI
- [14]E. Belgacem, Y. Dirian, S. Foffa, and M. Maggiore, Phys. Rev. D 97, 104066 (2018), arXiv:1712.08108 [astro-ph.CO].DOI
- [15]E. Belgacem, Y. Dirian, S. Foffa, and M. Maggiore, Phys. Rev. D 98, 023510 (2018), arXiv:1805.08731 [gr-qc].DOI
- [16]M. Lagos, M. Fishbach, P. Landry, and D. E. Holz, Phys. Rev. D 99, 083504 (2019), arXiv:1901.03321 [astro-ph.CO].DOI
- [17]M. Ishak et al., JCAP 09, 053, arXiv:2411.12026 [astro-ph.CO].DOI
- [18]E. Belgacem et al. (LISA Cosmology Working Group), JCAP 07, 024, arXiv:1906.01593 [astro-ph.CO].DOI
- [19]A. Finke, S. Foffa, F. Iacovelli, M. Maggiore, and M. Mancarella, JCAP 08, 026, arXiv:2101.12660 [astro-ph.CO].DOI
- [20]J. M. Ezquiaga, Phys. Lett. B 822, 136665 (2021), arXiv:2104.05139 [astro-ph.CO].DOI
- [21]M. Mancarella, E. Genoud-Prachex, and M. Maggiore, Phys. Rev. D 105, 064030 (2022), arXiv:2112.05728 [gr-qc].DOI
- [22]K. Leyde, S. Mastrogiovanni, D. A. Steer, E. Chassande-Mottin, and C. Karathanasis, JCAP 09, 012, arXiv:2202.00025 [gr-qc].DOI
- [23]A. Chen, R. Gray, and T. Baker, JCAP 02, 035, arXiv:2309.03833 [gr-qc].DOI
- [24]A. G. Abac et al. (LIGO Scientific, Virgo, KAGRA), Astrophys. J. Lett. 1007, L17 (2026), arXiv:2509.04348 [astro-ph.CO].
- [25]D. Langlois and K. Noui, JCAP 02, 034, arXiv:1510.06930 [gr-qc].DOI
- [26]G. R. Dvali, G. Gabadadze, and M. Porrati, Phys. Lett. B 485, 208 (2000), arXiv:hep-th/0005016.DOI
- [27]S. F. Hassan and R. A. Rosen, JHEP 02, 126, arXiv:1109.3515 [hep-th].DOI
- [28]E. Bellini and I. Sawicki, JCAP 07, 050, arXiv:1404.3713 [astro-ph.CO].
- [29]J. Gleyzes, D. Langlois, and F. Vernizzi, Int. J. Mod. Phys. D 23, 1443010 (2015), arXiv:1411.3712 [hep-th].
- [30]J. Noller and A. Nicola, Phys. Rev. D 99, 103502 (2019), arXiv:1811.12928 [astro-ph.CO].DOI
- [31]T. Baker and I. Harrison, JCAP 01, 068, arXiv:2007.13791 [astro-ph.CO].DOI
- [32]E. Seraille, J. Noller, and B. D. Sherwin, Phys. Rev. D 110, 123525 (2024), arXiv:2401.06221 [astro-ph.CO].
- [33](2026), arXiv:2605.27227 [astro-ph.CO].arxiv.org/abs/2605.27227
- [34]E. V. Linder, G. Sengör, and S. Watson, JCAP 05, 053, arXiv:1512.06180 [astro-ph.CO].
- [35]E. V. Linder, Phys. Rev. D 95, 023518 (2017), arXiv:1607.03113 [astro-ph.CO].DOI
- [36]M. D. Denssennya and E. V. Linder, JCAP 11, 010, arXiv:1808.00013 [astro-ph.CO].
- [37]J. Gleyzes, Phys. Rev. D 96, 063516 (2017), arXiv:1705.04714 [astro-ph.CO].DOI
- [38]G. Pierra, A. Colombo, and S. Mastrogiovanni, Class. Quant. Grav. 43, 155015 (2026), arXiv:2511.11795 [astro-ph.CO].
- [39]E. Belgacem, S. Foffa, M. Maggiore, and T. Yang, Phys. Rev. D 101, 063505 (2020), arXiv:1911.11497 [astro-ph.CO].DOI
- [40]E. Colangeli, K. Leyde, T. Baker, and A. Chen, (2026), arXiv:2608.18310 [gr-qc].arxiv.org/abs/2608.18310
- [41]N. Abac et al. (LIGO Scientific, VIRGO, KAGRA), (2026), arXiv:2605.27223 [gr-qc].arxiv.org/abs/2605.27223
- [42]J. M. Ezquiaga and M. Zumalacárregui, Front. Astron. Space Sci. 5, 44 (2018), arXiv:1807.09241 [astro-ph.CO].
- [43]B. P. Abbott et al., Phys. Rev. Lett. 119, 161101 (2017), arXiv:1710.05832 [gr-qc].
- [44]B. P. Abbott et al. (LIGO Scientific, Virgo, Fermi-GBM, INTEGRAL), Astrophys. J. Lett. 848, L13 (2017), arXiv:1710.05834 [astro-ph.HE].
- [45]D. Simpson, H. Rue, A. Riebler, T. G. Martins, and S. H. Sørbye, Statist. Sci. 32, 1 (2017), arXiv:1403.4630 [stat.ME].
- [46]G.-A. Fuglstad, D. Simpson, F. Lindgren, and H. Rue, J. Amer. Statist. Assoc. 114, 445 (2019), arXiv:1503.00256 [stat.ME].
- [47]A. G. Abac et al. (LIGO Scientific, VIRGO, KAGRA), (2025), arXiv:2508.18083 [astro-ph.HE].DOI
- [48](2026), arXiv:2605.27226 [astro-ph.HE].arxiv.org/abs/2605.27226
- [49]A. Agapito, V. De Renzis, and M. Mancarella, (2026), arXiv:2605.20112 [astro-ph.CO].DOI
- [50]I. Mandel, W. M. Farr, and J. R. Gair, Mon. Not. R. Astron. Soc. 486, 1086 (2019), arXiv:1809.02063 [physics.data-an].
- [51]S. Vitale, D. Gerosa, W. M. Farr, and S. R. Taylor 10.1007/978-981-15-4702-7_45-1 (2020), arXiv:2007.05579 [astro-ph.IM].DOI
- [52]V. Tiwari, Class. Quantum Grav. 35, 145009 (2018), arXiv:1712.00482 [astro-ph.HE].
- [53]W. M. Farr, Research Notes of the AAS 3, 66 (2019), arXiv:1904.10879 [astro-ph.IM].
- [54]C. Talbot and J. Golomb, Mon. Not. Roy. Astron. Soc. 526, 3495 (2023), arXiv:2304.06138 [astro-ph.IM].
- [55]J. Heinzel and S. Vitale, (2025), arXiv:2509.07221 [astro-ph.HE].arxiv.org/abs/2509.07221
- [56]M. Mancarella, github.com/CosmoStatGW/pymcpop-gw (2025).
- [57]M. Mancarella and D. Gerosa, Phys. Rev. D 111, 103012 (2025), arXiv:2502.12156 [gr-qc].DOI
- [58]M. D. Hoffmann and A. Gelman, J. Mach. Learn. Res. 15, 1593 (2014), arXiv:1111.4246 [stat.CO].
- [59]S. Brooks, A. Gelman, G. Jones, and X.-L. Meng, Handbook of Markov Chain Monte Carlo (Chapman and Hall/CRC, 2011).
- [60]D. Phan, N. Pradhan, and M. Jankowiak, NeurIPS 2019 (2019), arXiv:1912.11554 [stat].
- [61]E. Bingham, J. P. Chen, M. Jankowiak, F. Obermeyer, N. Pradhan, T. Karaletsos, R. Singh, P. A. Szerlip, P. Horsfall, and N. D. Goodman, J. Mach. Learn. Res. 20, 28 (2019), arXiv:1810.09538.
- [62]A. Gelman and D. B. Rubin, Stat. Sci. 7, 457 (1992).
- [63]A. M. Farah, B. Edelman, M. Zevin, M. Fishbach, J. M. Ezquiaga, B. Farr, and D. E. Holz, Astrophys. J. 955, 107 (2023), arXiv:2301.00834 [astro-ph.HE].
- [64](2026), arXiv:2605.27225 [gr-qc].arxiv.org/abs/2605.27225
- [65]L. S. Collaboration, V. Collaboration, and K. Collaboration, 10.5281/zenodo.17014085 (2025).
- [66]L. S. Collaboration, V. Collaboration, and K. Collaboration, 10.5281/zenodo.8177023 (2023).
- [67]L. S. Collaboration and V. Collaboration, 10.5281/zenodo.6513631 (2022).
- [68]L. S. Collaboration, V. Collaboration, and K. Collaboration, 10.5281/zenodo.20348005 (2026).
- [69]L. S. Collaboration, V. Collaboration, and K. Collaboration, 10.5281/zenodo.19500052 (2026).
- [70]R. Essick et al., Phys. Rev. D 112, 102001 (2025), arXiv:2508.10638 [gr-qc].
- [71]R. Essick, Phys. Rev. D 108, 043011 (2023), arXiv:2307.02765 [gr-qc].
- [72]R. Essick, Phys. Rev. D 108, 043011 (2023).
- [73]A. Agapito and M. Mancarella, (2026), arXiv:2609.38110 [astro-ph.CO].arxiv.org/abs/2609.38110
- [74]T. A. Callister, R. Essick, and D. E. Holz, Phys. Rev. D 110, 123041 (2024), arXiv:2408.16828 [astro-ph.HE].DOI
- [75]M. Tagliazucchi, M. Moresco, A. Agapito, M. Mancarella, S. Ferraiolo, S. Mastrogiovanni, N. Borghi, F. Pannarale, and D. Bonacorsi, Astron. Astrophys. 711, A120 (2026), arXiv:2602.17756 [astro-ph.CO].DOI
- [76]R. Essick and W. Farr, (2022), arXiv:2204.00461 [astro-ph.IM].arxiv.org/abs/2204.00461




