Pith. sign in

REVIEW 3 major objections 4 minor 44 references

Approximating non-Gaussian Bayesian partitions with normalising flows: statistics, inference and application to cosmology

T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Normalising flows compute Bayesian partition functions by differentiating a learned map at one point.

desk verdict A clean entropy and mean/variance pipeline for normalising flows, but the advertised non-Gaussian moments do not survive contact with the paper's own tables. read the letter →

arxiv 2501.04791 v3 pith:F4OIEEJI submitted 2025-01-08 astro-ph.CO astro-ph.IM

Tobias Röspel , Adrian Schlosser , Björn Malte Schäfer This is my paper · ORCID
classification astro-ph.COastro-ph.IM PACS 98.80.-k02.50.-r
keywords normalisingflowsBayesianpartitionfunctionevidencenon-GaussianposteriorinformationentropysupernovacosmologycumulantsMCMC
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

This paper argues that normalising flows—invertible, differentiable maps learned between a standard Gaussian and a non-Gaussian posterior—can serve as an analytical toolkit for Bayesian inference, not just a sampler. The key move is to rewrite the Bayesian partition function (the evidence and its temperature/source generalisations) as a Gaussian integral over the flow, then apply a derivative-operator identity that evaluates the integral by differentiating the flow map at a single point, $\alpha=0$. This yields series expressions for posterior moments and a simple formula for the information entropy in terms of the Gaussian base density and the flow's Jacobian determinant. Applied to supernova constraints on the matter density $\Omega_m$ and the dark-energy equation-of-state parameter $w_0$, the method reproduces the mean, variance, covariance and entropy obtained by MCMC; the same expansion currently fails for skewness and kurtosis because higher-order derivatives of the learned network become unreliable. If smoother flows can be trained, the approach would give parameter-free analytical access to non-Gaussian posterior statistics without additional sampling.

What carries the argument

The load-bearing object is the trained normalising flow $f(\alpha)=\theta(\alpha)$, an invertible differentiable map with Jacobian $Df$ that Gaussianises the posterior. The mechanistic identity is the flow expansion of Eq. (15), which uses the operator identity $\exp\!\left(\frac{T}{2}\delta^{\rho\sigma}\partial_\rho\partial_\sigma\right)$ to convert the Gaussian integral over $\alpha$ into derivatives of $g(\alpha)=|\det Df(\alpha)|^{1-1/T}\exp(J_\gamma\theta^\gamma(\alpha)/T)$ evaluated at $\alpha=0$. Differentiating this expression with respect to the source $J$ generates the moment series, and the change-of-variables formula supplies the entropy formula; the same learned flow therefore carries the partition function, moments, and entropy from a single mapping.

What would settle it

Train a normalising flow on a smooth, heavy-tailed one-dimensional distribution whose skewness and kurtosis are known analytically, then compute these moments from the flow expansion terminated at $k=4$; if the expansion values differ from the true values by more than the sampling noise, the claim that posterior moments follow from local derivatives of the flow is refuted for that class of distributions. The paper's Tables 4 and 5 already indicate such a failure for the supernova posterior kurtosis.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central discovery is that the normalising flow map $f(\alpha)=\theta(\alpha)$, trained to send a standard normal $p(\alpha)$ to the posterior $p(\theta|y)$, can be differentiated through the partition function. After the change of variables, the Bayesian partition function becomes $$Z[T,J]=\frac{(2\pi T)^{n/2}}{N(T)}\,\exp\!\left(\frac{T}{2}\$delta^{{\rho\sigma}}$\partial_\rho\partial_\$\sigma$\right)g(\$\alpha$)\big|_{\$\alpha$=0}$$ with $g(\alpha)=|\det Df(\alpha)|^{1-1/T}\exp(J_\gamma\theta^\gamma(\alpha)/T)$; the paper calls this the flow expansion. Posterior moments follow as $$\langle \$theta^{{\gamma_1}}$\cdots\$theta^{{\gamma_m}}$\rangle=\sum_{k=0}^\infty \frac{1}{2^k k!}\bigl(\$delta^{{\rho\sigma}}$\partial_\rho\partial_\$\sigma$\bigr)^k\, \$theta^{{\gamma_1}}$(\$\alpha$)\cdots\$theta^{{\gamma_m}}$(\$\alpha$)\big|_{\$\alpha$=0},$$ while the entropy is computed as a sample average of $-\ln p(\alpha)-\ln|\det Df(\alpha)|$ over the Gaussian base. The paper verifies the expansion on a Gaussian toy model and on the Union2.1 supernova posterior for $(\Omega_m,w_0)$, where mean, variance, covariance and entropy match the emcee ground truth within errors. Its own tables show that skewness and kurtosis from the expansion are not reliable, diverging from the sampled values by up to an order of magnitude, which the paper attributes to unstable higher-order derivatives of the network.

Load-bearing premise

The flow map is smooth enough that its Taylor expansion around $\alpha=0$ converges quickly, so truncating the moment series at fourth order gives accurate statistics; the paper's own tables show this breaks for skewness and kurtosis.

Editorial extensions

If this is right

  • A single trained flow yields the Bayesian evidence, the temperature-extended partition function $Z[T,J]$, and posterior moments and entropy from one object, so evidence and parameter estimation no longer require separate sampling runs.
  • The dependence of the partition function on temperature $T$ and sources $J$ can be mapped out cheaply by sampling the Gaussian base and transforming through the flow, opening a thermodynamic view of posterior inference.
  • Because the free energy is available as a function of $(T,J_1,J_2)$, one can in principle compensate a change in temperature with a change in the sources while keeping the free energy fixed, which the paper suggests may improve sampling.
  • For smooth-enough flows, the expansion would produce cumulants of arbitrary order analytically, making skewness, kurtosis and related statistics direct outputs of the network rather than noisy sample estimates; the present bottleneck is network smoothness, not the formalism.

Reading between the lines

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

  • A natural stress test is to apply the flow expansion to a one-dimensional mixture of Gaussians with known skewness and kurtosis: the single-point Taylor character means accuracy should degrade as the posterior becomes multimodal or heavy-tailed, and the order at which the series breaks down can be measured directly.
  • If smoother flow architectures deliver reliable higher derivatives, the same derivative machinery could compute the Fisher information and the surprise statistic (the Kullback-Leibler divergence between prior and posterior) without extra sampling, since both are expectation values over the posterior.
  • The free-energy compensation curves in the paper's Figs. 5 and 6 hint at a practical sampling strategy: choose $(T,J)$ along a constant-free-energy contour to anneal the posterior, which could be tested on a multimodal benchmark.
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 / 4 minor

Summary. The paper proposes combining normalising flows with Bayesian partition functions to compute evidence, entropy, and posterior moments analytically. The central device is a Gaussian integral identity (Eq. 15) that expresses the partition function as a differential operator acting on the flow map evaluated at alpha=0, leading to a series expansion (Eq. 18) for posterior moments. The authors apply the method to a supernova cosmology example (Omega_m and w0) and compare flow-based entropy, mean, variance, skewness, and kurtosis against histogram, KDE, and emcee posterior samples. They also present a Gaussian toy model for verification.

Significance. If the central claim were fully established, the paper would offer a useful alternative route to Bayesian evidence, information entropy, and posterior moments without additional Monte Carlo sampling. The formal Gaussian-integral identity in Eq. (15) is correct, and the entropy computation via Eq. (24) is clean and agrees with independent estimates. The toy-model validation with an analytic Gaussian target is a strength, and the code is made available. However, the headline promise of 'analytical expressions for posterior distributions beyond the Gaussian limit' is only partially supported: the flow expansion is truncated at k=4 and produces kurtosis values that are wrong by an order of magnitude in the cosmological application. The paper is honest about this limitation in its summary, but the abstract overstates what is demonstrated.

major comments (3)
  1. [§2.4, Eq. (18) and Tables 4/5] The central claim of analytical posterior moments beyond the Gaussian limit is not supported by the numerical results. In Tables 4 and 5, the flow expansion truncated at k=4 gives kurtosis values 6.2±1.5 versus 0.55±0.05 for Omega_m, and 5.2±1.3 versus 0.21±0.04 for w0, i.e. errors of roughly an order of magnitude. Kurtosis is a fourth-order quantity and is exactly the kind of non-Gaussian statistic the abstract promises. The paper should either restrict its claims to mean, variance, and covariance, or demonstrate that the expansion can be made accurate at higher order.
  2. [§2.4, Eq. (18) and Appendix A] The expansion in Eq. (18) computes moments from derivatives of the learned flow map f at alpha=0, but the training loss in Eq. (5) only constrains the pushforward density on training samples; it does not constrain the Taylor coefficients at alpha=0. The appendix itself notes that higher derivatives diverge with the FrEIA architecture because of permutations and other non-smooth components. Consequently, the truncation at k=4 is not controlled by the training objective, and the formal identity in Eq. (18) does not guarantee accurate moments for a learned flow. The authors should provide convergence evidence (e.g. comparison of truncation orders k=2,3,4) or use a provably smooth flow and show that the expansion converges.
  3. [§3, Tables 2, 4, 5] The validation is partially circular. The normalising flow is trained on emcee posterior samples, and the same emcee posterior is later treated as ground truth in Tables 4 and 5. Since the expansion moments are deterministic functions of the trained flow, agreement with emcee values is partly built into the training procedure. The toy model in Table 7 provides an independent check only for a Gaussian target. The supernova validation should be repeated with an independent sampler, or the comparison should split the emcee samples into disjoint training and test sets so that the ground truth is not the same data used for training.
minor comments (4)
  1. [Tables 1-7] Several table captions are formatted inconsistently, e.g. 'T able 1', 'T able 4'; these should be corrected to 'Table 1', 'Table 4', etc.
  2. [§3, near Eq. (22)] The text says 'c.p. Foreman-Mackey et al. 2013'; this should read 'cf.' as a standard abbreviation for 'confer'.
  3. [§2.3, Eq. (23)] In Eq. (23), the sample entropy estimate uses ln p(theta_i|y), but the paper does not specify how p(theta_i|y) is evaluated for the histogram and KDE estimates; a brief remark would improve reproducibility.
  4. [§2.2, near Eq. (6)] The statement that the Fisher information matrix F is the identity is tied to the standard normal choice; if a non-unit covariance is used, the resulting linear transform should be spelled out for clarity.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the flow-expansion derivation is a self-contained Gaussian integral identity, and the paper's empirical comparisons are honest validation rather than disguised fitting.

full rationale

The paper's central derivation is the 'flow expansion' (Eq. 15/18), which is a standard Gaussian integral identity: integrating exp(-alpha^2/2) g(alpha) yields exp(1/2 d^2/dalpha^2) g(alpha) at alpha=0, valid for sufficiently smooth g. This identity is taken from external calculus (with a pointer to Rota and Doubilet) and is not derived from the trained flow. The moment formula in Eq. 18 follows by differentiating that identity, and the entropy formula in Eq. 24 is a direct change-of-variables rewriting. None of these steps presupposes the posterior moments or entropy being computed. The normalising flow is trained on emcee posterior samples, so the flow-based moments are surrogate-model evaluations rather than independent predictions; but the paper does not hide this and explicitly compares the expansion against both flow sampling and emcee ground truth, reporting large kurtosis discrepancies (Tables 4 and 5). That is an honest failure of Taylor-series convergence, not a definitional equivalence. External grounding is provided by the Gaussian toy model (Tables 6 and 7), where the flow entropy matches the analytic value and the Gaussian kurtosis is recovered. Self-citations, such as Schosser et al. (2024) for the constant-w0 model choice and Kuntz et al. (2024) for comparative entropy values, are contextual or comparative and are not load-bearing in the derivation. Consequently there is no circular step that reduces the paper's claims to its own inputs.

Assumptions & free parameters 4 free parameters · 7 assumptions · 0 invented entities

The central quantities are all expressed through the trained flow; the flow weights and the truncation order are the main fitted inputs. No new physical entities are introduced.

free parameters (4)
  • Flow network weights (trained) = not reported; optimized by Adam
    The flow mapping f(alpha) is fit to posterior samples; all subsequent moment, entropy, and partition calculations depend on these fitted weights.
  • Flow expansion truncation order k = 4
    The series in Equation 18 is truncated at k=4; the paper states higher derivatives diverge, and the truncation directly causes the inaccurate kurtosis values.
  • Flow architecture hyperparameters = 1 layer width 64 (toy), 2 layers width 128 (supernova)
    Chosen by hand for the two examples; not derived from first principles.
  • Uniform prior bounds for Omega_m and w0 = not stated
    The prior range affects the posterior and is not specified in the text, hindering exact reproduction.
assumptions (7)
  • standard math Change-of-variables formula for probability densities
    Used in Equation 8 to relate p(alpha) and p(theta) via the Jacobian of the flow.
  • standard math Gaussian integral identity exp(1/2 nabla^2) h(0) = integral N(0,I) h dalpha
    Basis of the flow expansion in Equation 15.
  • standard math Bayes theorem and definition of evidence
    Equations 1 and 2 define the posterior and evidence.
  • domain assumption Flat FLRW cosmology with constant dark energy equation of state w0
    Assumed for the supernova likelihood in Section 3.
  • domain assumption Gaussian likelihood with known errors and uniform priors
    Defines the posterior for the supernova example, Equation 22.
  • domain assumption The trained flow accurately approximates the true posterior p(theta|y)
    The method's accuracy is bounded by training quality; no convergence guarantee is given.
  • ad hoc to paper Taylor expansion of the flow map converges and can be truncated at k=4
    The paper's own Tables 4 and 5 show this fails for kurtosis, so this is a load-bearing ad hoc assumption.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Approximating non-Gaussian Bayesian partitions with normalising flows: statistics, inference and application to cosmology." pith.science (2026). https://pith.science/paper/F4OIEEJI

@misc{pith2026250104791,
  author       = {Pith},
  title        = {Pith review of: Approximating non-Gaussian Bayesian partitions with normalising flows: statistics, inference and application to cosmology},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/F4OIEEJI}},
  note         = {Machine review of arXiv:2501.04791}
}
abstract

Subject of this paper is the simplification of Markov chain Monte Carlo sampling as used in Bayesian statistical inference by means of normalising flows, a machine learning method which is able to construct an invertible and differentiable transformation between Gaussian and non-Gaussian random distributions. We use normalising flows to compute Bayesian partition functions for non-Gaussian distributions and show how normalising flows can be employed in finding analytical expressions for posterior distributions beyond the Gaussian limit. Flows offer advantages for the numerical evaluation of the partition function itself, as well as for cumulants and for the information entropy. We demonstrate how normalising flows in conjunction with Bayes partitions can be used in inference problems in cosmology and apply them to the posterior distribution for the matter density $\Omega_m$ and a dark energy equation of state parameter $w_0$ on the basis of supernova data.

Figures

Figures reproduced from arXiv: 2501.04791 by the authors.

Figure 1
Figure 1. Schematic overview of the normalising flow learning the transformation f −1 (θ) from the posterior distribution p(θ| y) to a standard normal distribution p(α) for given data y during training. Sampling from a standard normal distribution in α and using the inverted normalising flow f (α) allows to reconstruct the original posterior distribution p(θ| y). We use derivatives of also higher orders of the normalising flo… view at source ↗
Figure 2
Figure 2. The normalising flow (orange) reproduces the posterior samples (blue) of the supernova Ia example, thus the two distribution as well as their marginals match. The Union2.1 data set (Suzuki et al. 2012; Kowalski et al. 2008; Amanullah et al. 2010) of supernovae of type Ia is used and contains 580 measurements. The distance modulus y is defined by the difference between the apparent magnitude m and the absolute magnit… view at source ↗
Figure 3
Figure 3. Geometric visualisation of the transformation induced by the normalising flow - zoom in on the maximum a posteriori region. The specific cell marked in red is analysed in more detail. The shading in the background represents the posterior probability. Calculating the coefficients introduced in Equation 25 for this matrix and rounding up to four digits, yields κ = −0.0509, γ1 = 0.0111, γ2 = −0.0564, and ω = −0.0780. … view at source ↗
Figures from the paper (3 more)
Figure 6
Figure 6. Figure 6: Helmholtz free energy (Equation 14) as a function of J1 and J2 at temperature T = 1 - normalised to its value at Jγ = 0 for γ ∈ {1, 2}. 6. SUMMARY AND DISCUSSION The subject of this paper was a hybrid approach to Bayesian inference with non-Gaussian distributions, com￾…
Figure 5
Figure 5. Figure 5: Helmholtz free energy (Equation 14) as a function of tempera￾ture T and J1 while J2 = 0, normalised to its value at T = 1 and Jγ = 0 for γ ∈ {1, 2}. In both plots, the isocontours reveal that there exist certain choices of {T, J1 } and {J1 , J2 } such that the Helmholt…
Figure 7
Figure 7. Figure 7: The inverted normalising flow (orange) reproduces nicely the posterior samples (blue) of the toy model. The toy model offers the possibility to calculate the infor￾mation entropy analytically. For a two-dimensional normal [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 26 canonical work pages

  1. [1]

    Trotta , journal Contemporary Physics volume 49 , pages 71 ( year 2008 ), ISSN issn 0010-7514, 1366-5812 , 0803.4089 , ://arxiv.org/abs/0803.4089

    author R. Trotta , journal Contemporary Physics volume 49 , pages 71 ( year 2008 ), ISSN issn 0010-7514, 1366-5812 , 0803.4089 , ://arxiv.org/abs/0803.4089

    arXiv 2008

  2. [2]

    Trotta , journal arXiv e-prints 1701.01467 ( year 2017 ), 1701.01467 , ://arxiv.org/abs/1701.01467

    author R. Trotta , journal arXiv e-prints 1701.01467 ( year 2017 ), 1701.01467 , ://arxiv.org/abs/1701.01467

    arXiv 2017

  3. [3]

    author C. R. Jenkins and author J. A. Peacock , journal MNRAS volume 413 , pages 2895 ( year 2011 ), ISSN issn 00358711 , 1101.4822 , ://arxiv.org/abs/1101.4822

    work page Pith review arXiv 2011

  4. [4]

    Handley and author P

    author W. Handley and author P. Lemos , journal arXiv e-prints 1903.06682 ( year 2019 ), 1903.06682 , ://arxiv.org/abs/1903.06682

    arXiv 1903

  5. [5]

    Trotta , volume 378 , pages 819 ( year 2007 ), ISSN issn 0035-8711, 1365-2966 , ://mnras.oxfordjournals.org/cgi/doi/10.1111/j.1365-2966.2007.11861.x

    author R. Trotta , volume 378 , pages 819 ( year 2007 ), ISSN issn 0035-8711, 1365-2966 , ://mnras.oxfordjournals.org/cgi/doi/10.1111/j.1365-2966.2007.11861.x

    arXiv 2007

  6. [6]

    author A. R. Liddle , author P. Mukherjee , author D. Parkinson , and author Y. Wang , journal PRD volume 74 ( year 2006 ), ISSN issn 1550-7998, 1550-2368 , astro-ph/0610126 , ://arxiv.org/abs/astro-ph/0610126

    arXiv 2006

  7. [7]

    Kerscher and author J

    author M. Kerscher and author J. Weller , journal SciPost p. pages 9 ( year 2019 ), ISSN issn 2590-1990 , 1901.07726 , ://arxiv.org/abs/1901.07726

    arXiv 2019

  8. [8]

    author K. H. Knuth , author M. Habeck , author N. K. Malakar , author A. M. Mubeen , and author B. Placek , journal Digital Signal Processing volume 47 , pages 50 ( year 2015 ), ISSN issn 1051-2004 , ://www.sciencedirect.com/science/article/pii/S1051200415001980

    work page 2015

Show all 44 references
  1. [9]

    Schosser , author T

    author B. Schosser , author T. R \"o spel , and author B. M. Schaefer ( year 2024 ), 2407.06259 , ://arxiv.org/abs/2407.06259

    2024

  2. [10]

    o ver , author H. von Campe , author M. P. Herzog , author R. M. Kuntz , and author B. M. Sch \

    author L. R \"o ver , author H. von Campe , author M. P. Herzog , author R. M. Kuntz , and author B. M. Sch \"a fer , journal Monthly Notices of the Royal Astronomical Society volume 526 , pages 473 ( year 2023 ), ISSN issn 0035-8711 , https://academic.oup.com/mnras/article-pd...

    2023 doi

  3. [11]

    o ver , and author B. M. Sch \

    author R. M. Kuntz , author M. P. Herzog , author H. von Campe , author L. R \"o ver , and author B. M. Sch \"a fer , title Partition function approach to non-gaussian likelihoods: partitions for the inference of functions and the fisher-functional ( year 2023 ), 2306.17224

    2023 arXiv

  4. [12]

    o ver , and author B. M. Sch \

    author M. P. Herzog , author H. von Campe , author R. M. Kuntz , author L. R \"o ver , and author B. M. Sch \"a fer , title Partition function approach to non-gaussian likelihoods: macrocanonical partitions and replicating markov-chains ( year 2023 ), 2311.16218

    2023 arXiv

  5. [13]

    author R. M. Kuntz , author H. von Campe , author T. Röspel , author M. P. Herzog , and author B. M. Schäfer , title Partition function approach to non-gaussian likelihoods: information theory and state variables for bayesian inference ( year 2024 ), 2411.13625 , ://arxiv.org/...

    2024 arXiv

  6. [14]

    o ver , author L. C. Bartels , and author B. M. Sch \

    author L. R \"o ver , author L. C. Bartels , and author B. M. Sch \"a fer , journal volume 523 , pages 2027 ( year 2023 ), 2210.03138

    2027 arXiv

  7. [15]

    Skilling , journal Bayesian Analysis volume 1 , pages 833 ( year 2006 ), ://doi.org/10.1214/06-BA127

    author J. Skilling , journal Bayesian Analysis volume 1 , pages 833 ( year 2006 ), ://doi.org/10.1214/06-BA127

    2006 doi

  8. [16]

    Ashton , author N

    author G. Ashton , author N. Bernstein , author J. Buchner , author X. Chen , author G. Cs \'a nyi , author A. Fowlie , author F. Feroz , author M. Griffiths , author W. Handley , author M. Habeck , et al. , journal Nature Reviews Methods Primers volume 2 , pages 39 ( year 202...

    2022 doi

  9. [17]

    Feroz , author M

    author F. Feroz , author M. Hobson , and author M. Bridges , journal Monthly Notices of the Royal Astronomical Society volume 398 , pages 1601 ( year 2009 )

    2009

  10. [18]

    author J. S. Speagle , journal Monthly Notices of the Royal Astronomical Society volume 493 , pages 3132 ( year 2020 ), https://doi.org/10.1093

    2020

  11. [19]

    Kilbinger , author D

    author M. Kilbinger , author D. Wraith , author C. P. Robert , author K. Benabed , author O. Capp \'e , author J.-F. Cardoso , author G. Fort , author S. Prunet , and author F. R. Bouchet , volume 405 , pages 2381 ( year 2010 )

    2010

  12. [20]

    Polanska , author M

    author A. Polanska , author M. A. Price , author D. Piras , author A. Spurio Mancini , and author J. D. McEwen , journal arXiv e-prints eid arXiv:2405.05969 ( year 2024 ), 2405.05969

    2024

  13. [21]

    Papamakarios , author E

    author G. Papamakarios , author E. Nalisnick , author D. Jimenez Rezende , author S. Mohamed , and author B. Lakshminarayanan , journal arXiv e-prints eid arXiv:1912.02762 ( year 2019 ), 1912.02762

    1912 arXiv

  14. [22]

    Cabezas , author L

    author A. Cabezas , author L. Sharrock , and author C. Nemeth , journal arXiv e-prints eid arXiv:2405.14392 ( year 2024 ), 2405.14392

    2024 arXiv

  15. [23]

    Srinivasan , author M

    author R. Srinivasan , author M. Crisostomi , author R. Trotta , author E. Barausse , and author M. Breschi , journal arXiv e-prints eid arXiv:2404.12294 ( year 2024 ), 2404.12294

    2024 arXiv

  16. [24]

    Gabri \'e , author G

    author M. Gabri \'e , author G. M. Rotskoff , and author E. Vanden-Eijnden , journal arXiv e-prints eid arXiv:2107.08001 ( year 2021 ), 2107.08001

    2021 arXiv

  17. [25]

    Raveri , author C

    author M. Raveri , author C. Doux , and author S. Pandey , journal arXiv preprint arXiv:2409.09101 ( year 2024 )

    2024 arXiv

  18. [26]

    Mootoovaloo , author C

    author A. Mootoovaloo , author C. Garc \' a-Garc \' a , author D. Alonso , and author J. Ruiz-Zapatero , journal arXiv preprint arXiv:2409.01407 ( year 2024 )

    2024 arXiv

  19. [27]

    Prathaban , author H

    author M. Prathaban , author H. Bevins , and author W. Handley , journal arXiv preprint arXiv:2411.17663 ( year 2024 )

    2024

  20. [28]

    Bevins , author W

    author H. Bevins , author W. Handley , author P. Lemos , author P. Sims , author E. de Lera Acedo , and author A. Fialkov , in booktitle Physical Sciences Forum ( organization MDPI , year 2022 ), vol. volume 5 , p. pages 1

    2022

  21. [29]

    author H. T. Bevins , author W. J. Handley , author P. Lemos , author P. H. Sims , author E. de Lera Acedo , author A. Fialkov , and author J. Alsing , journal Monthly Notices of the Royal Astronomical Society volume 526 , pages 4613 ( year 2023 )

    2023

  22. [30]

    Rezende and author S

    author D. Rezende and author S. Mohamed , in booktitle International conference on machine learning ( organization PMLR , year 2015 ), pp. pages 1530--1538

    2015

  23. [31]

    author J. C. Baez and author T. Fritz , journal arXiv e-prints 1402.3067 ( year 2014 ), 1402.3067 , ://arxiv.org/abs/1402.3067

    2014 arXiv

  24. [32]

    Ardizzone , author T

    author L. Ardizzone , author T. Bungert , author F. Draxler , author U. K \"o the , author J. Kruse , author R. Schmier , and author P. Sorrenson ( year 2018-2022 ), ://github.com/vislearn/FrEIA

    2018

  25. [33]

    Rota and author P

    author G.-C. Rota and author P. Doubilet , title Finite operator calculus ( publisher Academic Press, New York , year 1975 )

    1975

  26. [34]

    author A. G. Riess , author A. V. Filippenko , author P. Challis , author A. Clocchiatti , author A. Diercks , author P. M. Garnavich , author R. L. Gilliland , author C. J. Hogan , author S. Jha , author R. P. Kirshner , et al. , journal The astronomical journal volume 116 , ...

    1998

  27. [35]

    Suzuki , author D

    author N. Suzuki , author D. Rubin , author C. Lidman , author G. Aldering , author R. Amanullah , author K. Barbary , author L. Barrientos , author J. Botyanszki , author M. Brodwin , author N. Connolly , et al. , journal The Astrophysical Journal volume 746 , pages 85 ( year 2012 )

    2012

  28. [36]

    Kowalski , author D

    author M. Kowalski , author D. Rubin , author G. Aldering , author R. Agostinho , author A. Amadon , author R. Amanullah , author C. Balland , author K. Barbary , author G. Blanc , author P. J. Challis , et al. , journal The Astrophysical Journal volume 686 , pages 749 ( year 2008 )

    2008

  29. [37]

    Amanullah , author C

    author R. Amanullah , author C. Lidman , author D. Rubin , author G. Aldering , author P. Astier , author K. Barbary , author M. Burns , author A. Conley , author K. Dawson , author S. Deustua , et al. , journal The Astrophysical Journal volume 716 , pages 712 ( year 2010 )

    2010

  30. [38]

    Arutjunjan , author B

    author R. Arutjunjan , author B. M. Sch \"a fer , and author C. Kreutz , journal to be submitted to JRSSB ( year 2022 )

    2022

  31. [39]

    Buchner , author A

    author J. Buchner , author A. Georgakakis , author K. Nandra , author L. Hsu , author C. Rangel , author M. Brightman , author A. Merloni , author M. Salvato , author J. Donley , and author D. Kocevski , journal volume 564 , eid A125 ( year 2014 ), 1402.0004

    2014 arXiv

  32. [40]

    Foreman-Mackey , author D

    author D. Foreman-Mackey , author D. W. Hogg , author D. Lang , and author J. Goodman , journal volume 125 , pages 306 ( year 2013 ), 1202.3665

    2013 arXiv

  33. [41]

    Ao and author J

    author Z. Ao and author J. Li , in booktitle Proceedings of the AAAI Conference on Artificial Intelligence ( year 2022 ), vol. volume 36 , pp. pages 9990--9998

    2022

  34. [42]

    author B. M. Sch \"a fer and author R. Reischke , journal MNRAS volume 460 , pages 3398 ( year 2016 ), ISSN issn 0035-8711, 1365-2966 , 1603.03626 , ://arxiv.org/abs/1603.03626

    2016 arXiv

  35. [43]

    Ardizzone , author J

    author L. Ardizzone , author J. Kruse , author S. Wirkert , author D. Rahner , author E. W. Pellegrini , author R. S. Klessen , author L. Maier-Hein , author C. Rother , and author U. K \"o the , journal arXiv preprint arXiv:1808.04730 ( year 2018 )

    2018 arXiv

  36. [44]

    o hler , author A. Kr \

    author J. K \"o hler , author A. Kr \"a mer , and author F. No \'e , journal Advances in Neural Information Processing Systems volume 34 , pages 2796 ( year 2021 )

    2021

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.