Pith. sign in

REVIEW 3 major objections 7 minor 1 cited by

Deep Neural Emulation of the Supermassive Black-hole Binary Population

T0 review · 3 major / 7 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A normalizing-flow emulator trained on the full gravitational-wave-background strain distribution beats Gaussian-process emulators in fidelity and training cost, reproducing tails and frequency covariances that GPs miss.

desk verdict Flow-based emulator beats GP on marginal Hellinger distances in held-out tests, but the covariance-capture claim is asserted, not demonstrated. read the letter →

arxiv 2411.10519 v1 pith:D7UY6Y6O submitted 2024-11-15 astro-ph.IM gr-qc

Nima Laal , Stephen R. Taylor , Luke Zoltan Kelley , Joseph Simon , Kayhan Gultekin , David Wright , Bence Becsy , J. Andrew Casey-Clyde
show 8 more authors
Siyuan Chen Alexander Cingoranelli Daniel J. D'Orazio Emiko C. Gardiner William G. Lamb Cayenne Matt Magdalena S. Siwek Jeremy M. Wachter
This is my paper · ORCID
classification astro-ph.IMgr-qc
keywords supermassiveblack-holebinariesgravitationalwavebackgroundpulsartimingarraysnormalizingflowsGaussianprocessemulatorspopulationsynthesisBayesianinferencecharacteristicstrain
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper sets out to replace Gaussian-process (GP) emulators, which map binary-evolution parameters to the mean and standard deviation of the gravitational-wave-background (GWB) strain spectrum, with a normalizing-flow emulator that learns the entire conditional strain distribution over simulated universes. The motivation is that pulsar-timing-array measurements of a nanohertz background are the most direct probe of supermassive-black-hole-binary demographics, and inference about those demographics is only as good as the emulator's fidelity to the full spectrum, including its tails, non-Gaussian features, and inter-frequency covariances. By training an autoregressive-coupling rational-quadratic-spline flow on the full ensemble of characteristic-strain samples from a population-synthesis library, the paper achieves a median Hellinger distance of $0.08^{+0.05}_{-0.02}$ between emulated and library strain distributions, versus $0.20^{+0.10}_{-0.05}$ for the GP baseline, while training in about thirty minutes on a single GPU. The larger point is that simulation-based inference for the gravitational-wave background can now use complete distributional likelihoods rather than summary statistics.

What carries the argument

The load-bearing object is the ACRQS normalizing flow: an invertible, differentiable transformation made of rational-quadratic spline couplings parameterized by a masked autoencoder, which maps a uniform base distribution to the target density while making the change-of-variables Jacobian cheap and exact. Because the flow is trained to maximize the log-likelihood of samples drawn from the full library distribution of $\log_{10} h_c$ for each context vector $\theta_{\rm evo}$, it learns the joint five-bin distribution rather than independent per-frequency summaries, which is what lets it represent tails, multimodality, and cross-frequency covariances. The same exact-likelihood property lets the flow act as a surrogate likelihood inside a short Metropolis-Hastings chain that draws a power-spectral-density vector from the data posterior and then proposes $\theta_{\rm evo}$, a procedure the paper uses for its posterior-recovery comparisons.

What would settle it

Compute Hellinger distances on the full thirty-bin spectrum and, in addition, evaluate both emulators on a joint multivariate distance that is sensitive to cross-frequency correlation, such as the energy distance or maximum mean discrepancy between the emulated and library strain vectors; if the GP matches or beats the flow on the omitted bins or on the joint metric, the paper's central claim that the flow captures the full distribution and frequency covariances more faithfully would be contradicted.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that a normalizing flow of the autoregressive-coupling rational-quadratic-spline type, trained on the complete conditional distribution $p(\log_{10} h_c \mid \theta_{\rm evo})$ of the GWB characteristic strain across five jointly modeled frequency bins, reproduces the library's strain ensemble distributions markedly better than the per-frequency GP trained on medians and standard deviations. The headline quantitative evidence is the Hellinger distance computed over 2000 independent test distributions: $0.08^{+0.05}_{-0.02}$ for the best-trained flow, compared with $0.20^{+0.10}_{-0.05}$ for the GP. The flow also tracks the upper and lower quartiles of the strain distributions more accurately, and its multivariate treatment of the five bins preserves frequency covariances that the GP cannot represent. The paper is careful to report where GP still wins: GP median estimates are closer to the library medians, and neither emulator recovers the underlying binary-evolution parameters well in full six-dimensional inference, a failure the paper attributes to degeneracies and to the uniform distribution of parameters in the training library.

Load-bearing premise

The paper's headline comparison uses only five of the thirty frequency bins and judges each frequency separately, so the claimed advantage rests on those bins standing in for the whole spectrum and on per-bin comparisons capturing the flow's joint-distribution benefits.

Editorial extensions

If this is right

  • PTA-based inference can move from summary-statistic surrogate likelihoods to full distributional likelihoods, so the non-Gaussian width and shape of the strain ensemble enter parameter estimation directly.
  • The same flow architecture scales to all thirty frequency bins of the library in about nine hours on the same GPU class, a regime the paper says the GP approach cannot handle, opening higher-dimensional emulation of the binary population.
  • Emulating the tails of the strain distribution makes it possible to quantify the probability of unusually loud or quiet background realizations, which matters for interpreting single-universe measurements such as the current PTA background.
  • Because training completes in tens of minutes on a single GPU, emulators can be retrained quickly as new population-synthesis libraries or PTA data releases arrive.
  • The paper's sequential MCMC procedure gives a template for using flow emulators in hierarchical inference where the likelihood is a multivariate conditional density rather than a product of one-dimensional integrals.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper's five-bin tests, I would expect flow-based likelihoods to sharpen constraints on the amplitude and slope of the background spectrum and on binary-environment parameters, because those parameters shape exactly the covariances and tails that GPs discard.
  • The paper's own diagnosis that uniform and degenerate training parameters defeat posterior recovery suggests a direct follow-up test: train the flow on a library with a non-uniform, physically motivated prior over $\theta_{\rm evo}$ and check whether the inferred posteriors sharpen; the paper does not perform this test.
  • A hybrid emulator that uses the GP for the central value and the flow for distributional shape would combine the strengths the paper documents; this is my suggestion, not theirs.
  • The tail-fidelity result suggests a practical use in forecasting: flow emulators could estimate, for a given population model, the probability that a future PTA observes a background above a detection threshold, a quantity that median-based emulators cannot provide.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 7 minor

Summary. The paper presents a conditional normalizing-flow emulator (ACRQS) for the ensemble distribution of the gravitational-wave background characteristic strain h_c(f) produced by the holodeck population-synthesis library. The flow is conditioned on six supermassive black-hole binary evolution parameters θevo and is trained on five of the 30 available frequency bins, with the stated goal of placing it on equal footing with the Gaussian-process emulator used in Agazie et al. The authors compare the NF against the GP by per-frequency Hellinger distances on held-out holodeck samples, including tail-restricted versions; by median prediction accuracy; and by posterior recovery of θevo in a likelihood-based MCMC pipeline. They report a best-trained NF with Hellinger distance 0.08 (+0.05/-0.02) versus 0.20 (+0.10/-0.05) for the GP, while noting that the GP predicts medians more accurately and that neither emulator clearly outperforms the other in recovering θevo. The conclusion claims that the NF is faster, easier to train, and more faithful to the full strain distribution, including frequency covariances, tails, non-Gaussianities, and multimodalities.

Significance. If the central comparison is fully supported, the paper would be a useful methodological contribution to pulsar-timing-array inference: it demonstrates a surrogate that can generate full distributional samples at scale, with a clearly specified architecture and hyperparameters in Table I. The benchmark against an independent holodeck instance is a strength, and the authors are commendably explicit that the GP performs better at point statistics and that the NF does not improve θevo recovery. However, the significance is currently bounded by the fact that the headline fidelity comparison rests on marginal, five-bin evaluations; the multivariate and full-spectrum claims, which are the main advertised advantages over GPs, are not yet empirically established.

major comments (3)
  1. [V, Fig. 3, Eq. (8)] The headline fidelity claim is supported only by per-frequency marginal Hellinger distances. The Hellinger distance in Eq. (8) is computed on 1D histograms of log10 h_c for each frequency bin, so it is invariant to the joint dependency structure across frequencies; a flow that modeled the five marginals independently would receive the same score. The statement in §IV B that the NF 'retains frequency covariance information' is an architectural assertion, not a demonstrated result. To support the abstract and §VI claims about frequency covariances, the authors should add a multivariate comparison, for example empirical covariance matrices, a joint Hellinger distance on the 5D distribution, a maximum-mean-discrepancy test, or a joint two-sample test, and ideally perform the same evaluation on all 30 frequency bins.
  2. [IV B, V] The restriction to five of 30 frequency bins is not adequately justified. The text says this choice puts the NF on 'equal footing' with the GP, but the GP described in §III is trained independently per frequency, so it is unclear why the comparison could not be performed on all 30 bins or on an explicitly justified representative subset. Without evidence that the selected five bins are representative of the full spectrum—for example, by showing that held-out bins have similar Hellinger behavior or by spanning the frequency range—the generalization of the 0.08 versus 0.20 result to the full GWB spectrum is not established.
  3. [IV B, Fig. 3] The numerical comparison does not account for finite-sample noise in the reference histograms. The library supplies 2000 realizations per parameter vector, while the emulators generate 10^6 samples per θevo; with 15 histogram bins, sampling noise in the reference histogram alone contributes to the estimated Hellinger distance and enters the distributions shown in Fig. 3. The authors should quantify this effect, for instance by bootstrap resampling the library samples at fixed θevo or by evaluating both emulators and the library at matched sample sizes, to confirm that the reported NF advantage is not inflated by asymmetric sample-size noise.
minor comments (7)
  1. [III] The word 'keratosis' should be 'kurtosis'.
  2. [VI] The phrase 'inform inform strategies' contains a duplicated word.
  3. [Appendix A] The appendix heading reads 'Details of Construction of ARQS' but the acronym used throughout is ACRQS.
  4. [Eq. (A1)] The coefficients α, β, γ, a, b, and c are said to be functions of the knot values, but the explicit rational-quadratic spline formula is not given; a pointer to the relevant equations in Durkan et al. [28] would improve reproducibility.
  5. [V A] The statement that NF tails 'never exceed the bounds set by the entire training-set' should be explicitly tied to the chosen normalization bounds and base-distribution range described in §IV B.
  6. [V B] The sentence 'a MCMC simulation uses the the same kernel-density-estimates' contains a doubled article.
  7. [VI] The claim that 'GPs cannot handle 30 frequency-bins at all' is stronger than demonstrated, since separate per-frequency GPs could in principle be applied to 30 bins; the authors should soften this to refer to the GP architecture used in this comparison.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the NF-vs-GP fidelity comparison is a held-out empirical evaluation; the unsupported covariance claim is an evidence gap, not a circular step.

full rationale

The paper's derivation chain contains no load-bearing step that reduces to its own inputs. The NF is trained on the holodeck phenomenological library through the likelihood loss of Eq. 7, and its fidelity is then evaluated with Eq. 8 Hellinger distances between generated samples and an independent test-set instance of the same library that was not used in training: 'For the test-set, we use an instance of holodeck's library that is independent of the training-set.' Improvement of NF over GP on held-out θevo points is therefore an empirical generalization result, not resubstitution of training data. The GP baseline is imported from Agazie et al. [19], which shares authors, but it is a separately trained regression model built with the George library; no equation or fitted value from that citation is used to force the NF comparison, so the self-citation is contextual rather than load-bearing. The claim that NF captures frequency covariances, tails, and multimodalities goes beyond what the marginal per-bin Hellinger metric of Eq. 8 can demonstrate, but that is a correctness or evidence concern, not circularity: the claim is not true by construction, and the paper even reports a negative result ('Surprisingly, NF does not outperform the GP in inferring the right θevo'). No ansatz is smuggled in by citation; ACRQS is a standard published architecture adopted with cited references. Overall, the central fidelity comparison is self-contained with respect to the training procedure and rests on independent held-out samples.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

This paper introduces no new physical entities or fitted astrophysical constants. The scientific assumptions are inherited from the holodeck library, and the hand-tuned quantities are ML hyperparameters and evaluation choices. The central claim rests on the validity of holodeck as the ground truth and on the adequacy of the chosen fidelity metric.

free parameters (5)
  • NF architecture hyperparameters = bin_count=16; neurons=128; layers=4; learning_rate=1e-4; decay_rate=0.96; batch_size=1000
    Chosen empirically through many trials (Table I); the reported Hellinger comparison depends on this configuration.
  • Number of GWB frequency bins = 5
    Chosen to put the NF on equal footing with the GP from Agazie et al.; the headline fidelity scores apply only to this reduced set, not the full 30-bin spectrum.
  • Hellinger histogram bins = 15
    Number of bins used to discretize log10 hc for the Hellinger distance; the choice affects the magnitude of the reported distances.
  • NF training steps = 40000
    Stopping criterion chosen after validation and training losses stabilize; Figure 3 shows little improvement after 28000 steps.
  • Normalization bounds and base distribution = inputs scaled to [-5,5]; base U(-6,6)
    Arbitrary constants set in Section IV B to avoid edge effects; part of the training recipe.
assumptions (5)
  • domain assumption Holodeck's phenomenological binary evolution library faithfully generates GWB characteristic-strain distributions for SMBH binaries.
    All training, validation, and test distributions come from holodeck; the emulator fidelity claim is measured against this simulator, not real PTA data. Invoked in Section II.
  • domain assumption The six evolution parameters theta_evo = (phi0, m_phi0, mu, epsilon_mu, tau_f, nu_inner) are sufficient and independent enough to span the relevant simulation variety.
    Adopted from Agazie et al. [19]; the emulator conditions only on these parameters, and the theta_evo recovery test shows degeneracies. Invoked in Section II and Section V B.
  • standard math ACRQS normalizing flows with the selected architecture can represent the target conditional distributions of log10 hc.
    Training assumes sufficient flexibility of the flow family; only validated empirically on the held-out test set. Invoked in Section IV.
  • ad hoc to paper Hellinger distance with 15 histogram bins per frequency bin is an adequate measure of full-distribution fidelity, including tails and shape.
    The paper claims covariance and multimodality advantages but quantifies fidelity with marginal per-frequency histograms. This metric choice is made by the authors in Section IV B / V A and is not independently justified.
  • domain assumption The held-out test set, an independent instance of holodeck, is representative of the training distribution.
    If the test set has parameter regions with insufficient support, both emulators fail (as stated in Figure 3 caption); the overall Hellinger comparison averages over this. Invoked in Section IV B.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Deep Neural Emulation of the Supermassive Black-hole Binary Population." pith.science (2026). https://pith.science/paper/D7UY6Y6O

@misc{pith2026241110519,
  author       = {Pith},
  title        = {Pith review of: Deep Neural Emulation of the Supermassive Black-hole Binary Population},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/D7UY6Y6O}},
  note         = {Machine review of arXiv:2411.10519}
}
read the original abstract

While supermassive black-hole (SMBH)-binaries are not the only viable source for the low-frequency gravitational wave background (GWB) signal evidenced by the most recent pulsar timing array (PTA) data sets, they are expected to be the most likely. Thus, connecting the measured PTA GWB spectrum and the underlying physics governing the demographics and dynamics of SMBH-binaries is extremely important. Previously, Gaussian processes (GPs) and dense neural networks have been used to make such a connection by being built as conditional emulators; their input is some selected evolution or environmental SMBH-binary parameters and their output is the emulated mean and standard deviation of the GWB strain ensemble distribution over many Universes. In this paper, we use a normalizing flow (NF) emulator that is trained on the entirety of the GWB strain ensemble distribution, rather than only mean and standard deviation. As a result, we can predict strain distributions that mirror underlying simulations very closely while also capturing frequency covariances in the strain distributions as well as statistical complexities such as tails, non-Gaussianities, and multimodalities that are otherwise not learnable by existing techniques. In particular, we feature various comparisons between the NF-based emulator and the GP approach used extensively in past efforts. Our analyses conclude that the NF-based emulator not only outperforms GPs in the ease and computational cost of training but also outperforms in the fidelity of the emulated GWB strain ensemble distributions.

Figures

Figures reproduced from arXiv: 2411.10519 by the authors.

Figure 1
Figure 1. FIG. 1: Training (blue) and validation (orange) loss [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2: (a) A series of density plots comparing the characteristic-strain samples generated using NF (solid purple) [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3: Distributions illustrating the Hellinger distance [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (2 more)
Figure 5
Figure 5. Figure 5: FIG. 5: A series of distributions comparing the [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 7
Figure 7. Figure 7: FIG. 7: A series of density plots comparing the [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Summary statistic for pulsar timing arrays

    astro-ph.CO 2026-08 conditional novelty 7.0 of 10

    A PTA likelihood expressed in terms of low-order spherical harmonics of the Earth term and pulsar-term variance retains roughly 95% of the information about a stochastic background, and ell_max=3 plus the pulsar-term ...

Reference graph

Works this paper leans on

55 extracted references · 19 canonical work pages · cited by 1 Pith paper

  1. [1]

    Rajagopal and R

    M. Rajagopal and R. W. Romani, Ultra–Low-Frequency Gravitational Radiation from Massive Black Hole Bi- naries, ApJ 446, 543 (1995), arXiv:astro-ph/9412038 [astro-ph]

    arXiv 1995

  2. [2]

    A. H. Jaffe and D. C. Backer, Gravitational Waves Probe the Coalescence Rate of Massive Black Hole Binaries, ApJ 583, 616 (2003), arXiv:astro-ph/0210148 [astro-ph]

    arXiv 2003

  3. [3]

    J. S. B. Wyithe and A. Loeb, Low-Frequency Gravita- tional Waves from Massive Black Hole Binaries: Predic- tions for LISA and Pulsar Timing Arrays, ApJ 590, 691 (2003), arXiv:astro-ph/0211556 [astro-ph]

    arXiv 2003

  4. [4]

    Sesana, F

    A. Sesana, F. Haardt, P. Madau, and M. Volonteri, Low- Frequency Gravitational Radiation from Coalescing Mas- sive Black Hole Binaries in Hierarchical Cosmologies, ApJ 611, 623 (2004), arXiv:astro-ph/0401543 [astro-ph]

    arXiv 2004

  5. [5]

    Burke-Spolaor, S

    S. Burke-Spolaor, S. R. Taylor, M. Charisi, T. Dolch, J. S. Hazboun, A. M. Holgado, L. Z. Kelley, T. J. W. Lazio, D. R. Madison, N. McMann, C. M. F. Min- garelli, A. Rasskazov, X. Siemens, J. J. Simon, and T. L. Smith, The astrophysics of nanohertz gravitational waves, A&A Rev. 27, 5 (2019), arXiv:1811.08826 [astro- ph.HE]

    arXiv 2019

  6. [6]

    M. V. Sazhin, Opportunities for detecting ultralong grav- itational waves, Soviet Physics Journal 22, 36 (1978)

    work page 1978

  7. [7]

    Detweiler, Pulsar timing measurements and the search for gravitational waves, ApJ 234, 1100 (1979)

    S. Detweiler, Pulsar timing measurements and the search for gravitational waves, ApJ 234, 1100 (1979)

    1979

  8. [8]

    R. S. Foster and D. C. Backer, Constructing a Pulsar Timing Array, ApJ 361, 300 (1990)

    1990

Show all 55 references
  1. [9]

    S. R. Taylor, Nanohertz gravitational wave astronomy (CRC Press, 2021)

    2021

  2. [10]

    The training procedure for our NF-based astro- emulator goes as follows

    for the fairness of comparisons made in §V. The training procedure for our NF-based astro- emulator goes as follows. First, we transform the li- brary’s content ( θevo as well as log 10 hc) into values bounded between arbitrary constants −5 and 5. This step is required in orde...

  3. [11]

    Antoniadis, P

    J. Antoniadis, P. Arumugam, S. Arumugam, S. Babak, M. Bagchi, A. S. Bak Nielsen, C. G. Bassa, A. Bathula, A. Berthereau, M. Bonetti, E. Bortolas, P. R. Brook, M. Burgay, R. N. Caballero, A. Chalumeau, D. J. Cham- pion, S. Chanlaridis, S. Chen, I. Cognard, S. Danda- pat, D. Deb...

    2023 arXiv

  4. [12]

    D. J. Reardon, A. Zic, R. M. Shannon, G. B. Hobbs, M. Bailes, V. Di Marco, A. Kapur, A. F. Rogers, E. Thrane, J. Askew, N. D. R. Bhat, A. Cameron, M. Cury lo, W. A. Coles, S. Dai, B. Goncharov, M. Kerr, A. Kulkarni, Y. Levin, M. E. Lower, R. N. Manchester, R. Mandow, M. T. Mil...

    2023 arXiv

  5. [13]

    Agazie, A

    G. Agazie, A. Anumarlapudi, A. M. Archibald, Z. Ar- zoumanian, P. T. Baker, B. B´ ecsy, L. Blecha, A. Brazier, P. R. Brook, S. Burke-Spolaor, R. Burnette, R. Case, M. Charisi, S. Chatterjee, K. Chatziioannou, B. D. Cheeseboro, S. Chen, T. Cohen, J. M. Cordes, N. J. Cornish, F....

    2023 arXiv

  6. [14]

    Bi, Y.-M

    Y.-C. Bi, Y.-M. Wu, Z.-C. Chen, and Q.-G. Huang, Implications for the supermassive black hole binaries from the NANOGrav 15-year data set, Science China Physics, Mechanics, and Astronomy 66, 120402 (2023), arXiv:2307.00722 [astro-ph.CO]

    2023 arXiv

  7. [15]

    Ellis, M

    J. Ellis, M. Fairbairn, G. H¨ utsi, J. Raidal, J. Urru- tia, V. Vaskonen, and H. Veerm¨ ae, Gravitational waves from supermassive black hole binaries in light of the NANOGrav 15-year data, Phys. Rev. D 109, L021302 (2024), arXiv:2306.17021 [astro-ph.CO]

    2024 arXiv

  8. [16]

    R. W. Hellings and G. S. Downs, Upper limits on the isotropic gravitational radiation background from pulsar timing analysis, ApJ 265, L39 (1983)

    1983

  9. [17]

    Sato-Polito, M

    G. Sato-Polito, M. Zaldarriaga, and E. Quataert, Where are the supermassive black holes measured by PTAs?, Phys. Rev. D 110, 063020 (2024)

    2024

  10. [18]

    Goncharov, S

    B. Goncharov, S. Sardana, A. Sesana, J. Antoniadis, A. Chalumeau, D. Champion, S. Chen, E. F. Keane, G. Shaifullah, and L. Speri, Fewer supermassive binary black holes in pulsar timing array observations, arXiv e-prints , arXiv:2409.03627 (2024), arXiv:2409.03627 [astro-ph.HE]

    2024

  11. [19]

    [20] analyze NANOGrav’s latest data set in search of astrophysical or cosmological models that can explain the origin of the GWB signal measured by NANOGrav

    and Afzal et al. [20] analyze NANOGrav’s latest data set in search of astrophysical or cosmological models that can explain the origin of the GWB signal measured by NANOGrav. Even though the findings are consistent with astrophysical expectations and a selected few cos- mologi...

    2000

  12. [20]

    E. R. Liepold and C.-P. Ma, Big Galaxies and Big Black Holes: The Massive Ends of the Local Stellar and Black Hole Mass Functions and the Implications for 13 Nanohertz Gravitational Waves, ApJ 971, L29 (2024), arXiv:2407.14595 [astro-ph.GA]

    2024 arXiv

  13. [21]

    S. R. Taylor, J. Simon, and L. Sampson, Con- straints on the Dynamical Environments of Supermas- sive Black-Hole Binaries Using Pulsar-Timing Arrays, Phys. Rev. Lett. 118, 181102 (2017), arXiv:1612.02817 [astro-ph.GA]

    2017 arXiv

  14. [22]

    Agazie, A

    G. Agazie, A. Anumarlapudi, A. M. Archibald, P. T. Baker, B. B´ ecsy, L. Blecha, A. Bonilla, A. Brazier, P. R. Brook, S. Burke-Spolaor, R. Burnette, R. Case, J. A. Casey-Clyde, M. Charisi, S. Chatterjee, K. Chatziioan- nou, B. D. Cheeseboro, S. Chen, T. Cohen, J. M. Cordes, N....

    2023 arXiv

  15. [23]

    In §V we put this NF technique in use to learn the connection 12 between SMBHs’ binary evolution parameters and their GWB characteristic-strain

    which is ideal for our proposes in this work. In §V we put this NF technique in use to learn the connection 12 between SMBHs’ binary evolution parameters and their GWB characteristic-strain

  16. [24]

    Afzal, G

    A. Afzal, G. Agazie, A. Anumarlapudi, A. M. Archibald, Z. Arzoumanian, P. T. Baker, B. B´ ecsy, J. J. Blanco- Pillado, L. Blecha, K. K. Boddy, A. Brazier, P. R. Brook, S. Burke-Spolaor, R. Burnette, R. Case, M. Charisi, S. Chatterjee, K. Chatziioannou, B. D. Cheeseboro, S. Che...

    2023

  17. [25]

    K. W. K. Wong and D. Gerosa, Machine-learning in- terpolation of population-synthesis simulations to in- terpret gravitational-wave observations: A case study, Phys. Rev. D 100, 083015 (2019), arXiv:1909.06373 [astro-ph.HE]

    2019 arXiv

  18. [26]

    Coccaro, M

    A. Coccaro, M. Letizia, H. Reyes-Gonzalez, and R. Torre, Comparative Study of Coupling and Autoregressive Flows through Robust Statistical Tests, arXiv e-prints , arXiv:2302.12024 (2023), arXiv:2302.12024 [stat.ML]

    2023 arXiv

  19. [27]

    J. F. Crenshaw, J. Bryce Kalmbach, A. Gagliano, Z. Yan, A. J. Connolly, A. I. Malz, S. J. Schmidt, and The LSST Dark Energy Science Collaboration, Probabilis- tic Forward Modeling of Galaxy Catalogs with Normal- izing Flows, arXiv e-prints , arXiv:2405.04740 (2024), arXiv:2405...

    2024 arXiv

  20. [28]

    K. W. K. Wong, G. Contardo, and S. Ho, Gravitational- wave population inference with deep flow-based gen- erative network, Phys. Rev. D 101, 123005 (2020), arXiv:2002.09491 [astro-ph.IM]

    2020 arXiv

  21. [29]

    Bonetti, A

    M. Bonetti, A. Franchini, B. G. Galuzzi, and A. Sesana, Neural networks unveiling the properties of gravitational wave background from supermassive black hole binaries, A&A 687, A42 (2024), arXiv:2311.04276 [astro-ph.HE]

    2024 arXiv

  22. [30]

    Delbourgo and J

    R. Delbourgo and J. A. Gregory, Rational quadratic spline interpolation to monotonic data. (1982)

    1982

  23. [31]

    Durkan, A

    C. Durkan, A. Bekasov, I. Murray, and G. Papamakarios, Neural spline flows (2019), arXiv:1906.04032 [stat.ML]

    2019 arXiv

  24. [32]

    E. S. Phinney, A Practical Theorem on Gravitational Wave Backgrounds, ArXiv Astrophysics e-prints (2001), astro-ph/0108028

    2001 arXiv

  25. [33]

    L. S. Finn and K. S. Thorne, Gravitational waves from a compact star in a circular, inspiral orbit, in the equato- rial plane of a massive, spinning black hole, as observed by LISA, Phys. Rev. D 62, 124021 (2000), arXiv:gr- qc/0007074 [gr-qc]

    2000

  26. [34]

    S. Chen, A. Sesana, and C. J. Conselice, Constraining as- trophysical observables of galaxy and supermassive black hole binary mergers using pulsar timing arrays, MNRAS 488, 401 (2019), arXiv:1810.04184 [astro-ph.GA]

    2019 arXiv

  27. [35]

    Kormendy and L

    J. Kormendy and L. C. Ho, Coevolution (Or Not) of Su- permassive Black Holes and Host Galaxies, ARA&A 51, 511 (2013), arXiv:1304.7762 [astro-ph.CO]

    2013 arXiv

  28. [36]

    Hinshaw, D

    G. Hinshaw, D. Larson, E. Komatsu, D. N. Spergel, C. L. Bennett, J. Dunkley, M. R. Nolta, M. Halpern, R. S. Hill, N. Odegard, L. Page, K. M. Smith, J. L. Weiland, B. Gold, N. Jarosik, A. Kogut, M. Limon, S. S. Meyer, 14 G. S. Tucker, E. Wollack, and E. L. Wright, Nine-year Wil...

    2013 arXiv

  29. [37]

    Aigrain and D

    S. Aigrain and D. Foreman-Mackey, Gaussian Process Regression for Astronomical Time Series, ARA&A 61, 329 (2023), arXiv:2209.08940 [astro-ph.IM]

    2023 arXiv

  30. [38]

    W. G. Lamb and S. R. Taylor, Spectral Variance in a Stochastic Gravitational-wave Background from a Binary Population, ApJ 971, L10 (2024), arXiv:2407.06270 [gr- qc]

    2024 arXiv

  31. [39]

    J. S. Hazboun, J. Simon, S. R. Taylor, M. T. Lam, S. J. Vigeland, K. Islo, J. S. Key, Z. Arzoumanian, P. T. Baker, A. Brazier, P. R. Brook, S. Burke-Spolaor, S. Chatterjee, J. M. Cordes, N. J. Cornish, F. Craw- ford, K. Crowter, H. T. Cromartie, M. DeCesar, P. B. Demorest, T. ...

    2020 arXiv

  32. [40]

    Ambikasaran, D

    S. Ambikasaran, D. Foreman-Mackey, L. Greengard, D. W. Hogg, and M. O’Neil, Fast direct methods for gaus- sian processes, IEEE Transactions on Pattern Analysis and Machine Intelligence 38, 252 (2016)

    2016

  33. [41]

    S. R. Taylor and D. Gerosa, Mining gravitational-wave catalogs to understand binary stellar evolution: A new hierarchical Bayesian framework, Phys. Rev. D 98, 083017 (2018), arXiv:1806.08365 [astro-ph.HE]

    2018 arXiv

  34. [42]

    J. P. Agnelli, M. Cadeiras, E. G. Tabak, C. V. Turner, and E. Vanden-Eijnden, Clustering and clas- sification through normalizing flows in feature space, Multiscale Modeling & Simulation 8, 1784 (2010), https://doi.org/10.1137/100783522

    2010 doi

  35. [43]

    E. G. Tabak and C. V. Turner, A family of nonpara- metric density estimation algorithms, Communications on Pure and Applied Mathematics 66, 145 (2013), https://onlinelibrary.wiley.com/doi/pdf/10.1002/cpa.21423

    2013 doi

  36. [44]

    Kobyzev, S

    I. Kobyzev, S. J. Prince, and M. A. Brubaker, Normaliz- ing flows: An introduction and review of current meth- ods, IEEE Transactions on Pattern Analysis and Machine Intelligence 43, 3964 (2021)

    2021

  37. [45]

    D. P. Kingma and J. Ba, Adam: A method for stochastic optimization, CoRR abs/1412.6980 (2014)

    2014 arXiv

  38. [46]

    Kullback and R

    S. Kullback and R. A. Leibler, On information and suf- ficiency, The Annals of Mathematical Statistics 22, 79 (1951)

    1951

  39. [47]

    Germain, K

    M. Germain, K. Gregor, I. Murray, and H. Larochelle, Made: Masked autoencoder for distribution estimation, in Proceedings of the 32nd International Conference on Machine Learning, Proceedings of Machine Learning Re- search, Vol. 37, edited by F. Bach and D. Blei (PMLR, Lille, ...

    2015

  40. [48]

    Hellinger, Neue begr¨ undung der theorie quadratischer formen von unendlichvielen ver¨ anderlichen., Journal f¨ ur die reine und angewandte Mathematik 1909, 210 (1909)

    E. Hellinger, Neue begr¨ undung der theorie quadratischer formen von unendlichvielen ver¨ anderlichen., Journal f¨ ur die reine und angewandte Mathematik 1909, 210 (1909)

    1909

  41. [49]

    Bingham, J

    E. Bingham, J. P. Chen, M. Jankowiak, F. Obermeyer, N. Pradhan, T. Karaletsos, R. Singh, P. A. Szerlip, P. Horsfall, and N. D. Goodman, Pyro: Deep univer- sal probabilistic programming, J. Mach. Learn. Res. 20, 28:1 (2019)

    2019

  42. [50]

    W. G. Lamb, S. R. Taylor, and R. van Haasteren, The Need For Speed: Rapid Refitting Techniques for Bayesian Spectral Characterization of the Gravita- tional Wave Background Using PTAs, arXiv e-prints , arXiv:2303.15442 (2023), arXiv:2303.15442 [astro- ph.HE]

    2023 arXiv

  43. [51]

    N. Laal, W. G. Lamb, J. D. Romano, X. Siemens, S. R. Taylor, and R. van Haasteren, Exploring the capabilities of Gibbs sampling in pulsar timing arrays, Phys. Rev. D 108, 063008 (2023), arXiv:2305.12285 [astro-ph.IM]

    2023 arXiv

  44. [52]

    Kumar, C

    R. Kumar, C. Carroll, A. Hartikainen, and O. Mar- tin, Arviz a unified library for exploratory analysis of bayesian models in python, Journal of Open Source Soft- ware 4, 1143 (2019)

    2019

  45. [53]

    Ellis and R

    J. Ellis and R. van Haasteren, jellis18/ptmcmcsampler: Official release (2017)

    2017

  46. [54]

    Ansel, E

    J. Ansel, E. Yang, H. He, N. Gimelshein, A. Jain, M. Voznesensky, B. Bao, P. Bell, D. Berard, E. Burovski, G. Chauhan, A. Chourdia, W. Constable, A. Desmaison, Z. DeVito, E. Ellison, W. Feng, J. Gong, M. Gschwind, B. Hirsh, S. Huang, K. Kalambarkar, L. Kirsch, M. La- zos, M. L...

    2024

  47. [55]

    J. D. Hunter, Matplotlib: A 2d graphics environment, Computing in Science & Engineering 9, 90 (2007)

    2007

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.