Pith. sign in

REVIEW 3 major objections 1 minor 52 references

Neural Galerkin Normalizing Flows for Bayesian Inference of Diffusions with Inaccessible Boundaries

T0 review · 3 major / 1 minor · reviewed 2026-06-28 · grok-4.3

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

Pith's one-line read Neural Galerkin normalizing flows approximate transition densities of diffusions with inaccessible boundaries to enable efficient Bayesian inference via MCMC.

desk verdict The paper trains a Neural Galerkin normalizing flow offline over initial data and coefficients to approximate transition densities for diffusions with vanishing diffusion at inaccessible boundaries, but the accuracy and stability of that approximation under the singular Dirac initial condition still need concrete checks. read the letter →

arxiv 2606.04324 v1 pith:JTNSXGA4 submitted 2026-06-03 cs.LG stat.ML

Riccardo Saporiti , Fabio Nobile This is my paper · ORCID
classification cs.LGstat.ML
keywords normalizingflowsfokker-planckbayesianinferencediffusionprocessesstochasticvolatilitymcmcneuralgalerkinmethods
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

Bayesian inference for diffusion parameters from discrete observations requires the transition density between times, which is usually unavailable analytically. The paper develops a normalizing flow architecture within a Neural Galerkin framework that learns to solve the associated Fokker-Planck equation starting from a Dirac mass initial condition. Training occurs over a distribution of initial conditions and diffusion coefficients, with special handling for cases where the diffusion matrix vanishes at inaccessible boundaries such as in Feller-satisfying stochastic volatility models. After training, the product of these densities along an observed path approximates the likelihood, supporting fast posterior sampling with MCMC without solving PDEs repeatedly or using simulation-based alternatives.

What carries the argument

Neural Galerkin Normalizing Flow architecture for solving the Fokker-Planck equation with Dirac initial condition.

What would settle it

A direct comparison showing that the MCMC posterior obtained using the approximated likelihood differs substantially from the true posterior computed using exact transition densities or highly accurate numerical methods would falsify the claim.

Watch

Extended reading notes

Core claim

By solving the Fokker-Planck equation in a Neural Galerkin framework using a new normalizing flow architecture, the method learns the transition density function for diffusions between observation times, with the product of these densities serving as an approximation to the likelihood function that permits efficient Markov chain Monte Carlo sampling for posterior inference.

Load-bearing premise

The proposed architecture accurately solves the Fokker-Planck equation with Dirac initial condition for the considered diffusions across the specified training distribution.

Editorial extensions

If this is right

  • Enables cheap posterior sampling via MCMC after offline training.
  • Avoids real-time solution of the FP equation for each MCMC proposal.
  • Avoids repeated simulation of diffusion bridges required in other likelihood-free methods.
  • Handles diffusions where the diffusion matrix vanishes at inaccessible boundaries.

Reading between the lines

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

  • The approach may generalize to other stochastic processes beyond the trained distribution if the architecture captures the underlying dynamics well.
  • It could facilitate parameter estimation in high-frequency data scenarios where many transitions are observed.
  • Integration with other inference techniques like variational methods might further speed up computations.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 1 minor

Summary. The paper proposes a Neural Galerkin Normalizing Flow architecture to solve the Fokker-Planck equation with Dirac initial condition over a training distribution of initial data and diffusion coefficients. This is applied to diffusions with vanishing diffusion matrix at inaccessible boundaries (e.g., Feller-condition stochastic volatility models). The resulting transition densities are multiplied along observed trajectories to approximate the likelihood, enabling efficient MCMC posterior sampling without real-time PDE solves or repeated simulations.

Significance. If the approximation is accurate with controlled per-step error and generalizes beyond the training distribution, the offline-training approach would provide a computationally efficient route to Bayesian inference for parameter-dependent diffusions where analytic transitions are unavailable. The focus on degenerate FP operators is a targeted strength, but the lack of any reported validation leaves the practical significance unestablished.

major comments (3)
  1. [Abstract] Abstract: the central claim that the product of transition densities 'approximates the likelihood function' and enables 'cheap posterior sampling via MCMC' is load-bearing, yet the abstract supplies no numerical validation, error metrics, baseline comparisons, or tests on vanishing-diffusion cases, so the claim cannot be assessed.
  2. [Method description] Method (Neural Galerkin framework): no stability analysis of the residual for the degenerate FP operator or a priori error bounds for the Dirac-initial-condition solve are provided; without these, it is unclear whether the learned density remains positive and integrates to one when coefficients lie outside the training support.
  3. [Training] Training and generalization: the architecture is trained over a distribution of initial data and coefficients, but no out-of-distribution tests or accumulation-of-error analysis for the product along trajectories are described; coherent O(Δt) errors per step would bias the MCMC posterior even if individual solves appear plausible.
minor comments (1)
  1. Notation for the transition density p(x_{t+Δt}|x_t, θ) and the precise form of the Neural Galerkin loss could be clarified to make the connection to the standard Fokker-Planck operator explicit.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We are grateful to the referee for the constructive feedback. We address each major comment below and will revise the manuscript accordingly.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim that the product of transition densities 'approximates the likelihood function' and enables 'cheap posterior sampling via MCMC' is load-bearing, yet the abstract supplies no numerical validation, error metrics, baseline comparisons, or tests on vanishing-diffusion cases, so the claim cannot be assessed.

    Authors: We agree that the abstract, as currently written, does not reference supporting numerical evidence. In the revision we will add a concise statement summarizing the key validation results (error metrics, baseline comparisons, and performance on vanishing-diffusion models) to make the central claim assessable from the abstract alone. revision: yes

  2. Referee: [Method description] Method (Neural Galerkin framework): no stability analysis of the residual for the degenerate FP operator or a priori error bounds for the Dirac-initial-condition solve are provided; without these, it is unclear whether the learned density remains positive and integrates to one when coefficients lie outside the training support.

    Authors: The normalizing-flow parameterization guarantees positivity and unit integral by construction for any coefficient values. We acknowledge that the manuscript currently lacks a dedicated stability analysis of the residual for the degenerate operator and a priori bounds. We will add a short subsection providing a stability argument for the Neural Galerkin residual on degenerate Fokker-Planck operators together with preliminary a priori considerations. revision: yes

  3. Referee: [Training] Training and generalization: the architecture is trained over a distribution of initial data and coefficients, but no out-of-distribution tests or accumulation-of-error analysis for the product along trajectories are described; coherent O(Δt) errors per step would bias the MCMC posterior even if individual solves appear plausible.

    Authors: We agree that the current manuscript does not report explicit out-of-distribution tests or a dedicated accumulation-of-error analysis along trajectories. We will include both in the revised version: additional numerical experiments on parameter/initial-condition values outside the training distribution, and a quantitative study of per-step error propagation and its effect on the MCMC posterior. revision: yes

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; derivation is a standard numerical approximation technique.

full rationale

The paper trains a Neural Galerkin Normalizing Flow to approximate solutions of the Fokker-Planck equation over a distribution of initial data and coefficients, then forms the likelihood as the product of the resulting transition densities along a trajectory. This follows directly from the Markov property of the diffusion and does not reduce any prediction to a fitted input by construction, invoke load-bearing self-citations for uniqueness, or smuggle ansatzes. No quoted steps exhibit self-definition or renaming of known results; the method remains an independent forward solver whose accuracy is an empirical question separate from the derivation chain.

Assumptions & free parameters 0 free parameters · 2 assumptions · 0 invented entities

The central claim rests on the standard Fokker-Planck equation governing the transition density and the representational power of normalizing flows within the Neural Galerkin framework; no free parameters or invented entities are explicitly introduced in the abstract.

assumptions (2)
  • domain assumption The Fokker-Planck equation describes the time evolution of the probability density for the diffusion process.
    Invoked implicitly as the equation solved by the Neural Galerkin method.
  • domain assumption Normalizing flows can represent solutions to the Fokker-Planck PDE with Dirac initial conditions over the training distribution.
    Core assumption enabling the architecture to learn the transition density.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Neural Galerkin Normalizing Flows for Bayesian Inference of Diffusions with Inaccessible Boundaries." pith.science (2026). https://pith.science/paper/JTNSXGA4

@misc{pith2026260604324,
  author       = {Pith},
  title        = {Pith review of: Neural Galerkin Normalizing Flows for Bayesian Inference of Diffusions with Inaccessible Boundaries},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JTNSXGA4}},
  note         = {Machine review of arXiv:2606.04324}
}
read the original abstract

One of the primary challenges in Bayesian inference on the parameters of a diffusion model from discrete observations is the unavailability of an analytical expression for the transition density function between consecutive observation times, which is needed to derive the likelihood function. Extending previous studies that solve Fokker-Planck (FP) type partial differential equations with Normalizing Flows, we propose a new Normalizing Flow architecture to learn the transition density function of the diffusion process between two observation times. We do so by solving in a Neural Galerkin framework the associated FP equation with a Dirac mass as initial condition, over a specified training distribution of the initial datum and the coefficients of the diffusion. We specifically focus on processes whose diffusion matrix vanishes in certain inaccessible boundary regions, such as Stochastic Volatility models that satisfy a Feller condition. The product of the obtained transition densities evaluated along the observed trajectory approximates the likelihood function, thereby enabling cheap posterior sampling via Markov chain Monte Carlo (MCMC). After the offline training phase, inference becomes significantly more efficient, as it avoids the need to solve the FP equation in real time for each parameter proposed by the MCMC sampler or to rely on other likelihood-free methods for Bayesian inference that involve repeated simulation of diffusion bridges.

Figures

Figures reproduced from arXiv: 2606.04324 by the authors.

Figure 1
Figure 1. Log-likelihood on Heston-generated data, with error bars showing the maximum relative [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Top: log-lh slices for SVCEV data, with bars of the maximum relative-error of log-lh and [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Transformation executed by the Normalizing Flow [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Log-likelihood profiles for the Heston model [PITH_FULL_IMAGE:figures/full_fig_p018_4.png]
Figure 5
Figure 5. Figure 5: Log-likelihood profiles for the Heston model [PITH_FULL_IMAGE:figures/full_fig_p019_5.png]
Figure 6
Figure 6. Figure 6: Trace plots and autocorrelation function for the parameters [PITH_FULL_IMAGE:figures/full_fig_p019_6.png]
Figure 7
Figure 7. Figure 7: Posterior marginal distributions for the parameters of the Heston model (19). [PITH_FULL_IMAGE:figures/full_fig_p020_7.png]
Figure 8
Figure 8. Figure 8: Samples drawn from the posterior distribution of the Heston model. In the left figure, the [PITH_FULL_IMAGE:figures/full_fig_p020_8.png]
Figure 9
Figure 9. Figure 9: Log-likelihood evaluated on data generated from the BDFS model [PITH_FULL_IMAGE:figures/full_fig_p025_9.png]
Figure 10
Figure 10. Figure 10: Posterior marginal distributions for the parameters of the BDFS model [PITH_FULL_IMAGE:figures/full_fig_p025_10.png]
Figure 11
Figure 11. Figure 11: Q-Q (quantile-quantile) plot comparing the reference distribution of the conditional density [PITH_FULL_IMAGE:figures/full_fig_p027_11.png]
Figure 12
Figure 12. Figure 12: Q-Q (quantile-quantile) plot comparing the reference distribution of the conditional density [PITH_FULL_IMAGE:figures/full_fig_p027_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 4 canonical work pages

  1. [1]

    Roberts, and Paul Fearnhead

    Alexandros Beskos, Omiros Papaspiliopoulos, Gareth O. Roberts, and Paul Fearnhead. Exact and Computationally Efficient Likelihood-Based Estimation for Discretely Observed Diffusion Processes (with Discussion).Journal of the Royal Statistical Society: Series B (Statistical Methodology), 68(3):333–382, 2006

    2006

  2. [2]

    Roberts, Alexandros Beskos, and Paul Fearnhead

    Giorgos Sermaidis, Omiros Papaspiliopoulos, Gareth O. Roberts, Alexandros Beskos, and Paul Fearnhead. Markov Chain Monte Carlo for Exact Inference for Diffusions.Scandinavian Journal of Statistics, 40(2):294–321, 2013

    2013

  3. [3]

    MCMC Analysis of Diffusion Models with Application to Finance.Journal of Business & Economic Statistics, 19(2):177–191, 2001

    Bjorn Eraker. MCMC Analysis of Diffusion Models with Application to Finance.Journal of Business & Economic Statistics, 19(2):177–191, 2001

    2001

  4. [4]

    Likelihood Inference for Discretely Observed Nonlinear Diffusions.Econometrica, 69(4):959–993, 2001

    Omar Elerian, Siddhartha Chib, and Neil Shephard. Likelihood Inference for Discretely Observed Nonlinear Diffusions.Econometrica, 69(4):959–993, 2001

    2001

  5. [5]

    Christophe Andrieu and Gareth O. Roberts. The Pseudo-Marginal Approach for Efficient Monte Carlo Computations.The Annals of Statistics, 37(2):697–725, 2009

    2009

  6. [6]

    Particle Markov chain Monte Carlo methods.Journal of the Royal Statistical Society: Series B (Statistical Methodology), 72(3):269–342, 2010

    Christophe Andrieu, Arnaud Doucet, and Roman Holenstein. Particle Markov chain Monte Carlo methods.Journal of the Royal Statistical Society: Series B (Statistical Methodology), 72(3):269–342, 2010

    2010

  7. [7]

    Bayesian Inference for Diffusions with Low-Frequency Observations

    Christiane Fuchs. Bayesian Inference for Diffusions with Low-Frequency Observations. In Inference for Diffusion Processes: With Applications in Life Sciences, pages 171–278. Springer, 2013

    2013

  8. [8]

    Roberts and Osnat Stramer

    Gareth O. Roberts and Osnat Stramer. On Inference for Non-linear Diffusion Models Using Metropolis-Hastings Algorithms.Biometrika, 88(3):603–621, 2001

    2001

Show all 52 references
  1. [9]

    Springer-Verlag, New York, Heidelberg, 1972

    Iosif Ilyich Gikhman and Anatoli Vladimirovich Skorokhod.Stochastic Differential Equations. Springer-Verlag, New York, Heidelberg, 1972

    1972

  2. [10]

    Maziar Raissi, Paris Perdikaris, and George Em Karniadakis. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations.Journal of Computational Physics, 378:686–707, 2019

    2019

  3. [11]

    DGM: A deep learning algorithm for solving partial differential equations.Journal of Computational Physics, 375:1339–1364, 2018

    Justin Sirignano and Konstantinos Spiliopoulos. DGM: A deep learning algorithm for solving partial differential equations.Journal of Computational Physics, 375:1339–1364, 2018

    2018

  4. [12]

    Solving high-dimensional partial differential equations using deep learning.Proceedings of the National Academy of Sciences, 115(34):8505– 8510, 2018

    Jiequn Han, Arnulf Jentzen, and Weinan E. Solving high-dimensional partial differential equations using deep learning.Proceedings of the National Academy of Sciences, 115(34):8505– 8510, 2018

    2018

  5. [13]

    Kevrekidis, Lu Lu, Paris Perdikaris, Sifan Wang, and Liu Yang

    George Em Karniadakis, Ioannis G. Kevrekidis, Lu Lu, Paris Perdikaris, Sifan Wang, and Liu Yang. Physics-informed machine learning.Nature Reviews Physics, 3(6):422–440, 2021

    2021

  6. [14]

    Alfio Quarteroni, Paola Gervasio, and Francesco Regazzoni. Combining physics-based and data-driven models: advancing the frontiers of research with scientific machine learning.Math- ematical Models and Methods in Applied Sciences, page 1–167, March 2025

    2025

  7. [15]

    Haozhe Su, M. V . Tretyakov, and David P. Newton. Deep Learning of Transition Probability Densities for Stochastic Asset Models with Applications in Option Pricing.Management Science, 70(7):4428–4449, 2024

    2024

  8. [16]

    Learning the temporal evolution of multivariate densities via normalizing flows.Chaos: An Interdisciplinary Journal of Nonlinear Science, 32(3), 2022

    Yubin Lu, Romit Maulik, Ting Gao, Felix Dietrich, Ioannis G Kevrekidis, and Jinqiao Duan. Learning the temporal evolution of multivariate densities via normalizing flows.Chaos: An Interdisciplinary Journal of Nonlinear Science, 32(3), 2022

    2022

  9. [17]

    Adaptive deep density approximation for Fokker-Planck equations.Journal of Computational Physics, 457:111080, 2022

    Kejun Tang, Xiaoliang Wan, and Qifeng Liao. Adaptive deep density approximation for Fokker-Planck equations.Journal of Computational Physics, 457:111080, 2022. 10

    2022

  10. [18]

    Solving Time Dependent Fokker-Planck Equations via Temporal Normalizing Flow.Communications in Computational Physics, 32(2):401–423, January 2022

    Xiaodong Feng, Tao Zhou, and Li Zeng. Solving Time Dependent Fokker-Planck Equations via Temporal Normalizing Flow.Communications in Computational Physics, 32(2):401–423, January 2022

    2022

  11. [19]

    Characterizing possible failure modes in physics-informed neural networks.Advances in neural information processing systems, 34:26548–26560, 2021

    Aditi Krishnapriyan, Amir Gholami, Shandian Zhe, Robert Kirby, and Michael W Mahoney. Characterizing possible failure modes in physics-informed neural networks.Advances in neural information processing systems, 34:26548–26560, 2021

    2021

  12. [20]

    Mitigating Propagation Failures in Physics-informed Neural Networks using Retain-Resample-Release (R3) Sampling

    Arka Daw, Jie Bu, Sifan Wang, Paris Perdikaris, and Anuj Karpatne. Mitigating Propagation Failures in Physics-informed Neural Networks using Retain-Resample-Release (R3) Sampling. In Andreas Krause, Emma Brunskill, Kyunghyun Cho, Barbara Engelhardt, Sivan Sabato, and Jonathan ...

    2023

  13. [21]

    Neural Galerkin Normalizing Flow for Transition Probabil- ity Density Functions of Diffusion Models.arXiv preprint arXiv:2603.18907, 2026

    Riccardo Saporiti and Fabio Nobile. Neural Galerkin Normalizing Flow for Transition Probabil- ity Density Functions of Diffusion Models.arXiv preprint arXiv:2603.18907, 2026

    2026

  14. [22]

    Neural Galerkin schemes with active learning for high-dimensional evolution equations.Journal of Computational Physics, 496:112588, 2024

    Joan Bruna, Benjamin Peherstorfer, and Eric Vanden-Eijnden. Neural Galerkin schemes with active learning for high-dimensional evolution equations.Journal of Computational Physics, 496:112588, 2024

    2024

  15. [23]

    Graham, Alexandre H

    Matthew M. Graham, Alexandre H. Thiery, and Alexandros Beskos. Manifold Markov Chain Monte Carlo Methods for Bayesian Inference in Diffusion Models.Journal of the Royal Statistical Society Series B: Statistical Methodology, 84(4):1229–1256, 04 2022

    2022

  16. [24]

    Augmented pseudo-marginal Metropolis–Hastings for partially observed diffusion processes.Statistics and Computing, 32(1):21, Feb 2022

    Andrew Golightly and Chris Sherlock. Augmented pseudo-marginal Metropolis–Hastings for partially observed diffusion processes.Statistics and Computing, 32(1):21, Feb 2022

    2022

  17. [25]

    Bayesian inference for non-synchronously observed diffusions.SIAM/ASA Journal on Uncertainty Quantification, 14(2):369–393, 2026

    Ajay Jasra, Kengo Kamatani, and Amin Wu. Bayesian inference for non-synchronously observed diffusions.SIAM/ASA Journal on Uncertainty Quantification, 14(2):369–393, 2026

    2026

  18. [26]

    Bayesian inference from time se- ries of allele frequency data using exact simulation techniques.arXiv preprint arXiv:2502.12279, 2025

    Jaromir Sant, Paul A Jenkins, Jere Koskela, and Dario Spano. Bayesian inference from time se- ries of allele frequency data using exact simulation techniques.arXiv preprint arXiv:2502.12279, 2025

    2025

  19. [27]

    Unbiased Likelihood Estimation of Wright–Fisher Diffusion Processes

    Celia García-Pareja and Fabio Nobile. Unbiased Likelihood Estimation of Wright–Fisher Diffusion Processes. InInternational Conference on Monte Carlo and Quasi-Monte Carlo Methods in Scientific Computing, pages 259–275. Springer, 2022

    2022

  20. [28]

    Inference for SDE Models via Approximate Bayesian Computation.Journal of Computational and Graphical Statistics, 23(4):1080–1100, 2014

    Umberto Picchini. Inference for SDE Models via Approximate Bayesian Computation.Journal of Computational and Graphical Statistics, 23(4):1080–1100, 2014

    2014

  21. [29]

    Towards Data-Conditional Simula- tion for ABC Inference in Stochastic Differential Equations.Bayesian Analysis, 21(1):309 – 339, 2026

    Petar Jovanovski, Andrew Golightly, and Umberto Picchini. Towards Data-Conditional Simula- tion for ABC Inference in Stochastic Differential Equations.Bayesian Analysis, 21(1):309 – 339, 2026

    2026

  22. [30]

    Black-box variational inference for stochastic differential equations

    Tom Ryder, Andrew Golightly, A Stephen McGough, and Dennis Prangle. Black-box variational inference for stochastic differential equations. InInternational Conference on Machine Learning, pages 4423–4432. PMLR, 2018

    2018

  23. [31]

    Closed-form likelihood expansions for multivariate diffusions.The Annals of Statistics, 36(2):906–937, 2008

    Yacine Aït-Sahalia. Closed-form likelihood expansions for multivariate diffusions.The Annals of Statistics, 36(2):906–937, 2008

    2008

  24. [32]

    Explicit form of approximate transition probability density functions of diffusion processes.Journal of Econometrics, 187(1):57–73, 2015

    Seungmoon Choi. Explicit form of approximate transition probability density functions of diffusion processes.Journal of Econometrics, 187(1):57–73, 2015

    2015

  25. [33]

    A new delta expansion for multivariate diffusions via the Itô-Taylor expansion.Journal of Econometrics, 209(2):256–288, 2019

    Nian Yang, Nan Chen, and Xiangwei Wan. A new delta expansion for multivariate diffusions via the Itô-Taylor expansion.Journal of Econometrics, 209(2):256–288, 2019

    2019

  26. [34]

    A closed-form transition density expansion for elliptic and hypo-elliptic SDEs.arXiv preprint arXiv:2502.07047, 2025

    Yuga Iguchi and Alexandros Beskos. A closed-form transition density expansion for elliptic and hypo-elliptic SDEs.arXiv preprint arXiv:2502.07047, 2025

    2025

  27. [35]

    Pavliotis.Stochastic Processes and Applications

    Grigorios A. Pavliotis.Stochastic Processes and Applications. Springer New York, NY , 2014. 11

    2014

  28. [36]

    Boundary conditions for computing densities in hybrid models via PDE methods.Stochastics, 84(5-6):705–718, 2012

    Vladimir Lucic. Boundary conditions for computing densities in hybrid models via PDE methods.Stochastics, 84(5-6):705–718, 2012

    2012

  29. [37]

    Calibration of the Heston stochastic local volatility model: A finite volume scheme.International Journal of Financial Engineering, 08(01):2050048, 2021

    Bernd Engelmann, Frank Koster, and Daniel Oeltz. Calibration of the Heston stochastic local volatility model: A finite volume scheme.International Journal of Financial Engineering, 08(01):2050048, 2021

    2021

  30. [38]

    LSMR: An Iterative Algorithm for Sparse Least-Squares Problems.SIAM Journal on Scientific Computing, 33(5):2950–2971, 2011

    David Chin-Lung Fong and Michael Saunders. LSMR: An Iterative Algorithm for Sparse Least-Squares Problems.SIAM Journal on Scientific Computing, 33(5):2950–2971, 2011

    2011

  31. [39]

    Johnson, Samuel Kotz, and Narayanaswamy Balakrishnan.Continuous Univariate Distributions, Volume 2

    Norman L. Johnson, Samuel Kotz, and Narayanaswamy Balakrishnan.Continuous Univariate Distributions, Volume 2. John Wiley & Sons, 1995

    1995

  32. [40]

    Maximum likelihood estimation of stochastic volatility models.Journal of Financial Economics, 83(2):413–452, 2007

    Yacine Aït-Sahalia and Robert Kimmel. Maximum likelihood estimation of stochastic volatility models.Journal of Financial Economics, 83(2):413–452, 2007

    2007

  33. [41]

    Cox, Jonathan E

    John C. Cox, Jonathan E. Ingersoll, and Stephen A. Ross. A Theory of the Term Structure of Interest Rates.Econometrica, 53(2):385–407, 1985

    1985

  34. [42]

    Lamoureux and A

    C. Lamoureux and A. Paseka. Information in Option Prices and the Underlying Asset Dynamics. Working paper, Eller School of Business, University of Arizona, 2005

    2005

  35. [43]

    Radford M. Neal. Slice sampling.The Annals of Statistics, 31(3):705–767, 2003

    2003

  36. [44]

    Springer, Berlin, Heidelberg, 2009

    Cédric Villani.Optimal Transport: Old and New, volume 338 ofGrundlehren der Mathematis- chen Wissenschaften. Springer, Berlin, Heidelberg, 2009

    2009

  37. [45]

    Ronald Gallant

    Garland B Durham and A. Ronald Gallant. Numerical Techniques for Maximum Likelihood Estimation of Continuous-Time Diffusion Processes.Journal of Business & Economic Statistics, 20(3):297–338, 2002

    2002

  38. [46]

    Bayesian Inference for Irreducible Diffusion Processes Using the Pseudo-Marginal Approach.Bayesian Analysis, 6(2):231–258, 2011

    Osnat Stramer and Matthew Bognar. Bayesian Inference for Irreducible Diffusion Processes Using the Pseudo-Marginal Approach.Bayesian Analysis, 6(2):231–258, 2011

    2011

  39. [47]

    Ilya I. Gikhman. A Short Remark on Feller’s Square Root Condition. SSRN Working Paper, 2011

    2011

  40. [48]

    Pytorch: An imperative style, high-performance deep learning library.Advances in neural information processing systems, 32, 2019

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An imperative style, high-performance deep learning library.Advances in neural information processing systems, 32, 2019

    2019

  41. [49]

    Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows

    George Papamakarios, David Sterratt, and Iain Murray. Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows. In Kamalika Chaudhuri and Masashi Sugiyama, editors,Proceedings of the Twenty-Second International Conference on Artificial Intelligenc...

    2019

  42. [50]

    Das, Silverio Foresi, and Rangarajan K

    Pierluigi Balduzzi, Sanjiv R. Das, Silverio Foresi, and Rangarajan K. Sundaram. A Simple Approach to Three-Factor Affine Term Structure Models.Journal of Fixed Income, 6(1):43–53, 1996

    1996

  43. [51]

    H. E. Daniels. Saddlepoint Approximations in Statistics.The Annals of Mathematical Statistics, 25(4):631–650, 1954

    1954

  44. [52]

    ∂n2,m θl (τ)(x1 m) ∂x1m ≃δ x0((x0)m)

    Yaniv Leviathan, Matan Kalman, and Yossi Matias. Fast inference from transformers via speculative decoding. InInternational Conference on Machine Learning, pages 19274–19286. PMLR, 2023. 12 A Structure of the Normalizing Flow The multi-layer schematic of the bounded Normalizin...

    2023

Pith tools

Reviewed June 28, 2026 · model on record in the stance chip above.