Pith. sign in

REVIEW 4 major objections 5 minor 85 references

Discrete Spatial Diffusion: Intensity-Preserving Diffusion Modeling

T0 review · 4 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read Diffusion models can conserve total image intensity exactly by treating each intensity unit as a particle that hops across the pixel lattice, and the approach extends to rock and battery-electrode microstructures.

desk verdict DSD's spatial random-walk diffusion with exact intensity conservation is a real contribution, but the no-exclusion dynamics can produce overlapping phases, and the scientific metrics are computed without showing the outputs are valid single-phase images. read the letter →

arxiv 2505.01917 v2 pith:IMNWDSVK submitted 2025-05-03 cs.GR cond-mat.mtrl-scics.LGeess.IV

Javier E. Santos , Agnese Marcato , Roman Colman , Nicholas Lubbers , Yen Ting Lin This is my paper · ORCID
classification cs.GRcond-mat.mtrl-scics.LGeess.IV
keywords discrete-statediffusionintensityconservationjumpprocesslatticerandomwalkmassmicrostructuregenerationimageinpaintinggenerativemodeling
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 establish that generative diffusion models can obey a hard conservation law: the total number of intensity units (particles) in each color channel is preserved exactly, in every sample, through both the forward corruption and the reverse generation process. Earlier discrete diffusion models randomize each pixel's intensity independently, so global quantities such as mass can at best be matched on average or forced afterward by projection; DSD instead lets every unit of intensity perform a continuous-time random walk on the spatial lattice, making total intensity conserved by construction while a neural network learns the reverse hopping rates. The paper validates the approach on MNIST, CIFAR-10, and CelebA, including inpainting, class conditioning, and exact total-intensity conditioning, and on two scientific applications: porous rock microstructures and lithium-ion battery electrodes, where generated samples hit target porosity or phase volume fractions exactly and reproduce transport-relevant statistics. If correct, this gives scientists a generative model whose outputs respect conservation laws automatically rather than approximately.

What carries the argument

The central object is a Markov jump process on a discrete lattice in which each unit of intensity is an independent particle hopping to one of four nearest neighbors at rate $r$ (Eq. 1); under periodic boundaries its transition probabilities are computed exactly by exponentiating the Markov matrix, which the discrete Fourier transform diagonalizes (Eq. 22). The load-bearing identity is Eq. 4: the reverse-time rate for the first of $n$ particles to leave a pixel is the sum, over those $n$ particles, of the per-particle reverse rates, each proportional to the forward transition probability ratio, a result derived from survival analysis of the independent-particle system. Around this identity the paper builds a training scheme — an NCSN++ network with four directional outputs and a SoftPlus nonlinearity, trained with either rate-matching or process-likelihood loss — and a sampler based on binomial $\tau$-leaping with an adaptive step size capped by a Courant–Friedrichs–Lewy condition, so that no transition probability exceeds a fixed tolerance at any step.

What would settle it

Measure the conservation claim directly by summing per-channel intensity before and after full reverse-time sampling; it should match the conditioning value to machine precision, since exactness holds by construction. To test the independence premise that carries the method, build a training set whose defining feature is an interaction between intensity units that a single-particle random walk cannot express — for example, binary images with a hard rule that occupied pixels never share a diagonal neighbor — train DSD on it, and compare the generated samples' two-point correlation with the training set's; a measurable mismatch while total intensity is exactly right would identify the independence assumption as the limiting factor.

Watch

Extended reading notes

Core claim

At the center of the paper is the claim that a diffusion model built on a continuous-time, discrete-state jump process — each unit of image intensity is a particle hopping to a nearest neighbor on the pixel lattice at a constant rate $r$ — preserves total intensity per color channel exactly in both the forward and reverse phases, and that this exactly constrained process still learns realistic structure. The reverse-time transition rate for a pixel holding $n$ particles is the sum of per-particle rates, each proportional to a ratio of forward transition probabilities (Eq. 4), obtained by survival analysis of the independent-particle system, and a neural network is trained to predict these rates from corrupted images. On MNIST, CIFAR-10, and CelebA the paper reports coherent samples, with the conserved total intensity acting as a semantic control: raising the allowed particle count changes a thin digit into a bold one. On three rock types and NMC battery cathodes, DSD-generated microstructures reproduce two-point correlations, pore size distributions, interface length, triple-phase boundary, and relative diffusivity of the training data while exactly hitting porosity or phase-fraction targets.

Load-bearing premise

The load-bearing premise is that each unit of intensity moves independently of all the others, so that the reverse-time rate for a pixel holding many particles (Eq. 4) is just a sum of single-particle rates; any interaction structure in the data that this independence cannot express must be learned only approximately by the network.

Editorial extensions

If this is right

  • Total intensity per color channel is conserved exactly in every generated sample, making quantities such as porosity, mass fraction, or digit stroke area hard constraints that hold identically, not merely on average.
  • Conditioning on total intensity is exact by construction at every intensity value, including far outside the typical training range, where the paper shows Gaussian-diffusion conditioning fails at the distribution tails.
  • The same machinery transfers from natural images to scientific microstructures: generated rocks and battery electrodes match training-set two-point correlations, pore size distributions, and electrochemical transport metrics while hitting exact porosity or phase volume fractions.
  • Computational cost scales linearly with the total intensity of the image, which makes the method inexpensive for binary and low-bit-depth scientific data and increasingly costly for high-intensity natural images.
  • Because the forward process is a plain lattice random walk, the authors state that extension to 3D requires only implementation changes, and they argue the framework opens the door to noise processes that exploit other conservation laws and symmetries.

Reading between the lines

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

  • Beyond the paper: the independence assumption predicts a concrete failure mode — training data whose structure lives in interactions between intensity units (for instance, a rule against adjacent occupied pixels) should be reproduced only approximately, since the forward random walk erases exactly that structure; an experiment comparing two-point statistics of real and generated samples on such da
  • Beyond the paper: the exact conservation law covers per-channel particle count only, so cross-channel invariants such as stoichiometric ratios or color balance are not pinned down; extending the construction to other additive conserved quantities is the natural next step that the paper's closing remark gestures toward.
  • Beyond the paper: in the fully corrupted limit the particle configuration is the maximally spread layout with fixed total count, so DSD is most at home when the conserved quantity is the physically dominant constraint — the reported CIFAR-10 FID near 20 (after filtering and classifier screening) plausibly reflects the price of hard conservation for texture-rich natural images.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes Discrete Spatial Diffusion (DSD), a continuous-time discrete-state diffusion framework in which intensity units are treated as particles performing nearest-neighbor random walks on the image lattice. The forward process conserves total particle count per color channel, and the reverse process is defined through per-particle reverse transition rates that are aggregated via survival analysis in Eq. (4). The authors train a neural network to predict reverse rates using either a rate-matching or a likelihood loss, and sample via a binomial tau-leaping scheme with an adaptive CFL-based step size. Experiments cover MNIST inpainting and class conditioning, unconditional CIFAR-10 and CelebA generation, and two scientific applications: binary rock microstructure generation and two-channel lithium-ion battery electrode generation. The central claim is that DSD exactly preserves total intensity by construction, enabling constrained generation for scientific microstructure modeling.

Significance. If the construction holds, DSD is a useful contribution to discrete-state diffusion modeling: it introduces spatial correlations into the noise process while preserving an exact global invariant, which is rare among diffusion-based generative models. The reverse-rate derivation in Appendix A via survival analysis is a clean way to aggregate independent-particle rates, and the spectral solution for periodic boundaries in Appendix C is elegant. The paper also demonstrates the method on scientifically relevant datasets, which is valuable for the microstructure community. The main caveats are that the scientific validation assumes generated images are valid single-phase configurations, an assumption not enforced by the sampling algorithm, and that several implementation details in the pseudocode and appendices need correction.

major comments (4)
  1. [Sec. 4.2/4.3 and Table 3] The forward process in Eq. (1) and the sampling in Algorithm 2 do not enforce single-phase occupancy, so generated binary rock images can contain pixels with intensity greater than 1, and generated electrode images can have both the carbon-binder and active-material channels active in the same pixel. Because no thresholding or exclusion post-processing is described for the scientific datasets, the PoreSpy and TauFactor metrics in Figures 16 and 19 may be computed on physically invalid phase configurations. This undermines the claim that the generated microstructures are realistic under strict mass conservation, since total intensity per channel only equals phase volume fraction when the image is a valid phase map. The Table 3 note that the CFL 'tolerance avoids overlapping particles' is not supported by Algorithm 2, because the CFL condition only bounds the per-step jump probability and does not prevent a pixel from receiving particles from multiple neighbors or from multiple channels.
  2. [Algorithm 2] In Algorithm 2, the binomial sampling step draws n_Sigma from Binom([I_t]_{x,y,c}, sum_nu r_NN_nu), where r_NN_nu are rates with units of inverse time. As written, the second argument is not a probability: it omits the factor tau that was just computed by the CFL rule, and it is not normalized by the number of particles at the pixel. A correct binomial tau-leaping step would use a probability of the order of (tau / [I_t]_{x,y,c}) times the total rate, for example p = 1 - exp(-tau * sum_nu r_NN_nu / [I_t]_{x,y,c}). This is the central sampling step, so the pseudocode must be corrected and the implementation checked against it.
  3. [Appendix C, Eqs. (14), (21), (22)] The normalization of the discrete Fourier pair is inconsistent. With the unitary convention in Eq. (14), a single particle at the origin gives k_{m,n}(0) = 1/sqrt(N), not 1/N as stated in Eq. (21). Consequently, Eq. (22) does not satisfy p_{0,0}(0) = 1 with the stated prefactor 1/sqrt(N^3). Since these transition probabilities are used in Algorithm 1 to corrupt images and to compute the rates in Eq. (4), the normalization error must be fixed. The prefactor cancels in the rate ratios of Eq. (4), but it matters when sampling forward trajectories from pt.
  4. [Table 1 and Sec. D.2] The FID of 20.6 reported for the 'CFL epsilon = 0.01 + filtering + classifier thresholding' row is computed on only 21,780 of the 50,000 generated images, after discarding samples with classifier confidence below 0.99. This is not comparable to standard FID scores computed on all generated samples, and the paper should either report the metric on the full set of 50,000 samples under the same filtering rule or clearly label the retained-subset score as an oracle-style upper bound. The same concern applies to the sFID value in that row.
minor comments (5)
  1. [Sec. 3.2, Eq. (2)] The schedule equation is difficult to parse: the left-hand side uses t_k but t_k is defined only implicitly by the equation, and the notation 'Phi(e^{-tau2 tk})' should be written as a function of t_k. Also, the definition of Phi should be Phi(p) = log(p/(1-p)), with parentheses, to avoid ambiguity.
  2. [Sec. 3.4, Eqs. (6)-(7)] The sign of the likelihood loss appears inconsistent: Eq. (6) defines -log L = -E[integral(...)], while Eq. (7) states log L = E[(t_k - t_{k-1}) sum(...)]. Please check the overall sign so that the loss is nonnegative and the equations are consistent.
  3. [Algorithm 2] The step size line 'tau <- min{t, epsilon min_{...} (r_NN)^{-1}}' is correct as epsilon divided by the maximum rate, but the use of 'min' over the inverse rates is easy to misread. Writing tau = min{t, epsilon / max_{...} r_NN} would be clearer.
  4. [Sec. F.2] The comparison of the rock FID of 0.9 with a literature value of 18.1 from Lee and Yun is acknowledged not to be one-to-one; this comparison should be moved to the appendix or marked as indicative rather than quantitative.
  5. [General] No code repository or seed information is provided. Given the number of algorithm choices and the pseudocode issues above, a code release with the exact sampling and post-processing steps would substantially improve reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the intensity-conservation claim is a theorem of the defined jump process, and the empirical evaluations do not reduce to fitted inputs or self-citations.

full rationale

The paper's central claim is that the DSD forward and reverse processes exactly preserve total intensity per color channel. This is not a fitted prediction: Eq. (1) defines each intensity unit as a particle that hops to a nearest neighbor without changing its color channel and without creation or destruction, so the total count per channel is conserved by the process definition. The reverse-time rate in Eq. (4) is derived in Appendix A via survival analysis from the independent-particle assumption and the forward transition probabilities, rather than being set equal to the neural network output or to the measured quantities. The neural network is trained to match these derived rates through the rate-matching or likelihood losses in Eqs. (5) and (6), so the reported generations are not forced by a parameter fitted to the same evaluation metric. The paper does cite the same authors' prior work for the continuous-time discrete-state formalism and for the noise-schedule ansatz, but those citations are not load-bearing for the conservation result: the schedule in Eq. (2) is explicitly described as arbitrary and heuristic, and the reverse-time derivation is reproduced in the paper itself. The external scientific metrics (PoreSpy for rocks, TauFactor for electrodes) are computed with independent tools, and the paper honestly reports FID values that are not state-of-the-art, which is inconsistent with any circular inflation of results. The Table 3 note that a CFL tolerance 'avoids overlapping particles' is unexplained and raises a genuine validity concern about phase-volume interpretation on multi-occupancy pixels, but that is a correctness risk rather than a circularity: no equation or fitted constant is being disguised as an independent prediction. Overall, no step was found where the claimed result reduces by construction to its own input.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The method rests on standard Markov-process theory, an explicit independence assumption for the forward jumps, and standard approximation assumptions for neural network training and tau-leaping. No new physical entities are introduced; the 'particles' are a bookkeeping device for integer intensity units. Free parameters are limited to the noise schedule and sampling tolerance.

free parameters (3)
  • forward transition rate r = 120 (MNIST), 85 (inpainting), 160 (CIFAR-10), 200 (CelebA), 200x5 (electrodes), 250x4 (rocks)
    The hopping rate is chosen per dataset to ensure full corruption within the time window; tuned on a subset of training samples together with the schedule. Listed in Table 3.
  • schedule parameters tau1, tau2 = 7.5 and 2.5
    Parameters of the logit-based noise schedule (Eq. 2), chosen so SSIM degrades roughly evenly across training times. The paper states the functional form is arbitrary and without theoretical foundation.
  • CFL tolerance epsilon = 0.01, 0.05, 0.10, 0.15, 0.2 (per dataset)
    Adaptive time-step parameter in Algorithm 2; set empirically per experiment and affects sample quality and fidelity.
assumptions (4)
  • standard math Existence and form of reverse-time rates for continuous-time Markov jump processes (Anderson's theorem)
    Used to write the reverse-time transition (Eq. 3) and the per-particle rates (Eq. 8); standard result cited to [2, 11, 63].
  • domain assumption Independence of particle motion in the forward process
    Each intensity unit independently performs a nearest-neighbor random walk at rate r (Sec. 3.1, Eq. 1); the many-particle reverse rate (Eq. 4) is derived from this independence via survival analysis in Appendix A.
  • domain assumption Neural network can learn the conditional expectation of reverse rates from the corrupted configuration only
    The training target depends on latent initial positions; the implicit assumption is that the network's prediction converges to the true marginal reverse rate given the current image, analogous to score matching in continuous diffusion.
  • domain assumption Total per-channel intensity is the only conserved quantity that needs to be enforced
    The method conserves particle counts by construction; it does not account for other physical constraints (e.g., interactions, momentum, connectivity) that may matter in specific scientific data.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Discrete Spatial Diffusion: Intensity-Preserving Diffusion Modeling." pith.science (2026). https://pith.science/paper/IMNWDSVK

@misc{pith2026250501917,
  author       = {Pith},
  title        = {Pith review of: Discrete Spatial Diffusion: Intensity-Preserving Diffusion Modeling},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IMNWDSVK}},
  note         = {Machine review of arXiv:2505.01917}
}
read the original abstract

Generative diffusion models have achieved remarkable success in producing high-quality images. However, these models typically operate in continuous intensity spaces, diffusing independently across pixels and color channels. As a result, they are fundamentally ill-suited for applications involving inherently discrete quantities-such as particle counts or material units-that are constrained by strict conservation laws like mass conservation, limiting their applicability in scientific workflows. To address this limitation, we propose Discrete Spatial Diffusion (DSD), a framework based on a continuous-time, discrete-state jump stochastic process that operates directly in discrete spatial domains while strictly preserving particle counts in both forward and reverse diffusion processes. By using spatial diffusion to achieve particle conservation, we introduce stochasticity naturally through a discrete formulation. We demonstrate the expressive flexibility of DSD by performing image synthesis, class conditioning, and image inpainting across standard image benchmarks, while exactly conditioning total image intensity. We validate DSD on two challenging scientific applications: porous rock microstructures and lithium-ion battery electrodes, demonstrating its ability to generate structurally realistic samples under strict mass conservation constraints, with quantitative evaluation using state-of-the-art metrics for transport and electrochemical performance.

Figures

Figures reproduced from arXiv: 2505.01917 by the authors.

Figure 1
Figure 1. Schematic diagrams illustrating how intensity is modeled in different diffusion frameworks. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. (a) The forward processes for Gaussian Diffusion [36], Inverse Heat Dissipation Model [61], and Discrete Spatial Diffusion (ours) applied on an image, sampled at discrete times. (b) Percentage change in intensity relative to the original image under the forward process. 3.3 Reverse-time process Following the process formalism from [11, 63], there exists a reverse-time process that evolves in opposite time and whose … view at source ↗
Figure 3
Figure 3. (a) Inpainting realizations on MNIST; 15% difference of conditioning intensity between consecutive rows. (b) Conditioned MNIST generations across different intensities and classes. (c) Unconditional CIFAR-10 generations. (d) Unconditional CelebA generations. 4.2 Subsurface rock microstructures The microstructure of subsurface rocks governs a wide range of physical processes, including fluid transport, electrical res… view at source ↗
Figures from the paper (16 more)
Figure 4
Figure 4. Figure 4: Distributions of porosity and generated samples for three rock classes. Each plot shows [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Generated cathode microstructures with varying phase volume fractions. The carbon binder domain appears in black, active material particles in gray, and electrolyte-filled pore space in white. The active material is expensive, creating a strong economic incentive to un…
Figure 6
Figure 6. Figure 6: Structural Similarity Index Metric between the original and corrupted MNIST and CelebA [PITH_FULL_IMAGE:figures/full_fig_p018_6.png]
Figure 7
Figure 7. Figure 7: Unconditional MNIST generations. Then, in our additional MNIST experiments, we explored class-conditional and inpainting generation. These experiments are particularly notable due to their interactions with the intensity-preserving property of DSD. For class-conditioni…
Figure 8
Figure 8. Figure 8: Class-conditional MNIST generations for digits 0 through 4. Each row corresponds to a [PITH_FULL_IMAGE:figures/full_fig_p020_8.png]
Figure 9
Figure 9. Figure 9: Unconditional MNIST inpainting with progressively increasing total intensity in the masked [PITH_FULL_IMAGE:figures/full_fig_p021_9.png]
Figure 10
Figure 10. Figure 10: Unconditional CIFAR-10 generations produced by DSD with CFL tolerance [PITH_FULL_IMAGE:figures/full_fig_p022_10.png]
Figure 11
Figure 11. Figure 11: FID scores computed on 20,000 CIFAR-10 generations under different sampling configu [PITH_FULL_IMAGE:figures/full_fig_p023_11.png]
Figure 12
Figure 12. Figure 12: CIFAR-10 generations after bilateral filtering. These images correspond directly to [PITH_FULL_IMAGE:figures/full_fig_p023_12.png]
Figure 13
Figure 13. Figure 13: Unconditional generations of the CelebA 64 [PITH_FULL_IMAGE:figures/full_fig_p024_13.png]
Figure 14
Figure 14. Figure 14: Sample-wise mean intensity across 1,000 realizations of mnist digits for each target [PITH_FULL_IMAGE:figures/full_fig_p025_14.png]
Figure 15
Figure 15. Figure 15: Schematic representation of three rock types: Berea Sandstone, Savonnières Carbonate, [PITH_FULL_IMAGE:figures/full_fig_p027_15.png]
Figure 16
Figure 16. Figure 16: Quantitative comparison between training and generated rock samples. [PITH_FULL_IMAGE:figures/full_fig_p027_16.png]
Figure 17
Figure 17. Figure 17: Realizations of Discrete Spatial Diffusion (DSD) on 1,000 [PITH_FULL_IMAGE:figures/full_fig_p028_17.png]
Figure 18
Figure 18. Figure 18: Microstructural characterization metrics for 80 training samples and 80 generated samples. [PITH_FULL_IMAGE:figures/full_fig_p029_18.png]
Figure 19
Figure 19. Figure 19: (Top) Eight randomly picked samples from the NMC cathodes dataset. (Bottom) Random unconditional realizations of our model [PITH_FULL_IMAGE:figures/full_fig_p030_19.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

85 extracted references · 57 canonical work pages

  1. [1]

    score_models: A repository to store state of the art score model architectures

    Alexandre Adam. score_models: A repository to store state of the art score model architectures. https://github.com/AlexandreAdam/score_models, 2023. Accessed: 2025-05-13

    2023

  2. [2]

    Anderson

    Brian D.O. Anderson. Reverse-time diffusion equation models. Stochastic Processes and their Applications, 12(3):313–326, 1982. ISSN 0304-4149. doi: https://doi.org/10.1016/ 0304-4149(82)90051-5

    1982

  3. [3]

    Johnson, Jonathan Ho, Daniel Tarlow, and Rianne van den Berg

    Jacob Austin, Daniel D. Johnson, Jonathan Ho, Daniel Tarlow, and Rianne van den Berg. Structured denoising diffusion models in discrete state-spaces. In Marc’Aurelio Ranzato, Alina Beygelzimer, Yann N. Dauphin, Percy Liang, and Jennifer Wortman Vaughan, editors,Advances in Neural Information Processing Systems 34: Annual Conference on Neural Information P...

    2021

  4. [4]

    Predictive microstructure image gener- ation using denoising diffusion probabilistic models

    Erfan Azqadan, Hamid Jahed, and Arash Arami. Predictive microstructure image gener- ation using denoising diffusion probabilistic models. Acta Materialia, 261:119406, 2023. ISSN 1359-6454. doi: https://doi.org/10.1016/j.actamat.2023.119406. URL https://www. sciencedirect.com/science/article/pii/S135964542300736X

    arXiv 2023

  5. [5]

    Li, Hamid Kazemi, Furong Huang, Micah Goldblum, Jonas Geiping, and Tom Goldstein

    Arpit Bansal, Eitan Borgnia, Hong-Min Chu, Jie S. Li, Hamid Kazemi, Furong Huang, Micah Goldblum, Jonas Geiping, and Tom Goldstein. Cold diffusion: Inverting arbitrary image transforms without noise, 2022

    2022

  6. [6]

    Dynamics of fluids in porous media

    Jacob Bear. Dynamics of fluids in porous media. Courier Corporation, 2013

    2013

  7. [7]

    Pore-scale imaging and modelling

    Martin J Blunt, Branko Bijeljic, Hu Dong, Oussama Gharbi, Stefan Iglauer, Peyman Mostaghimi, Adriana Paluszny, and Christopher Pentland. Pore-scale imaging and modelling. Advances in Water resources, 51:197–216, 2013

    2013

  8. [8]

    3D mapping of water in oolithic limestone at atmospheric and vacuum saturation using x-ray micro-ct differential imaging

    MA Boone. 3D mapping of water in oolithic limestone at atmospheric and vacuum saturation using x-ray micro-ct differential imaging. Materials Characterization, 97:150–160, 2014

    work page 2014

Show all 85 references
  1. [9]

    Investigating the relative permeability behavior of microporosity-rich carbonates and tight sandstones with multiscale pore network models

    Tom Bultreys, Jeroen Van Stappen, Tim De Kock, Wesley De Boever, Marijn A Boone, Luc Van Hoorebeke, and Veerle Cnudde. Investigating the relative permeability behavior of microporosity-rich carbonates and tight sandstones with multiscale pore network models. Journal of Geophys...

    2016

  2. [10]

    Deformation and fracture mechanisms in fine-and ultrafine-grained ferrite/martensite dual-phase steels and the effect of aging

    Marion Calcagnotto, Yoshitaka Adachi, Dirk Ponge, and Dierk Raabe. Deformation and fracture mechanisms in fine-and ultrafine-grained ferrite/martensite dual-phase steels and the effect of aging. Acta Materialia, 59(2):658–670, 2011

    2011

  3. [11]

    A continuous time framework for discrete denoising models

    Andrew Campbell, Joe Benton, Valentin De Bortoli, Thomas Rainforth, George Deligiannidis, and Arnaud Doucet. A continuous time framework for discrete denoising models. In S. Koyejo, S. Mohamed, A. Agarwal, D. Belgrave, K. Cho, and A. Oh, editors, Advances in Neural Information...

    2022

  4. [12]

    Gillespie, and Linda R

    Yang Cao, Daniel T. Gillespie, and Linda R. Petzold. Avoiding negative populations in explicit Poisson tau-leaping. The Journal of Chemical Physics, 123(5):054104, August 2005. ISSN 0021-9606. doi: 10.1063/1.1992473

    2005 doi

  5. [13]

    Gillespie, and Linda R

    Yang Cao, Daniel T. Gillespie, and Linda R. Petzold. Efficient step size selection for the tau-leaping simulation method. The Journal of Chemical Physics, 124(4):044109, January 2006. ISSN 0021-9606. doi: 10.1063/1.2159468. 10

    2006 doi

  6. [14]

    Vlachos, and Markos A

    Abhijit Chatterjee, Dionisios G. Vlachos, and Markos A. Katsoulakis. Binomial distribution based τ-leap accelerated stochastic simulation. The Journal of Chemical Physics , 122(2), January 2005. ISSN 0021-9606. doi: 10.1063/1.1833357

    2005 doi

  7. [15]

    Score-based diffusion models for accelerated MRI.Medical Image Analysis, 80:102479, August 2022

    Hyungjin Chung and Jong Chul Ye. Score-based diffusion models for accelerated MRI.Medical Image Analysis, 80:102479, August 2022. ISSN 1361-8415. doi: 10.1016/j.media.2022.102479

    2022

  8. [16]

    Diffusion Posterior Sampling for General Noisy Inverse Problems

    Hyungjin Chung, Jeongsol Kim, Michael Thompson McCann, Marc Louis Klasky, and Jong Chul Ye. Diffusion Posterior Sampling for General Noisy Inverse Problems. In The Eleventh International Conference on Learning Representations, September 2022

    2022

  9. [17]

    Improving Diffusion Models for Inverse Problems using Manifold Constraints

    Hyungjin Chung, Byeongsu Sim, Dohoon Ryu, and Jong Chul Ye. Improving Diffusion Models for Inverse Problems using Manifold Constraints. InAdvances in Neural Information Processing Systems, October 2022

    2022

  10. [18]

    Come-Closer-Diffuse-Faster: Accelerating Conditional Diffusion Models for Inverse Problems through Stochastic Contraction

    Hyungjin Chung, Byeongsu Sim, and Jong Chul Ye. Come-Closer-Diffuse-Faster: Accelerating Conditional Diffusion Models for Inverse Problems through Stochastic Contraction. In 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 12403– 12412, June 20...

    2022

  11. [19]

    High-resolution x-ray computed tomography in geosciences: A review of the current technology and applications

    Veerle Cnudde and Matthieu Nicolaas Boone. High-resolution x-ray computed tomography in geosciences: A review of the current technology and applications. Earth-Science Reviews, 123: 1–17, 2013

    2013

  12. [20]

    Alice Corbella, Simon E. F. Spencer, and Gareth O. Roberts. Automatic Zig-Zag sampling in practice. Statistics and Computing , 32(6):107, November 2022. ISSN 1573-1375. doi: 10.1007/s11222-022-10142-x

    2022 doi

  13. [21]

    Courant, K

    R. Courant, K. Friedrichs, and H. Lewy. Über die partiellen Differenzengleichungen der mathematischen Physik. Mathematische Annalen, 100(1):32–74, December 1928. ISSN 1432-1807. doi: 10.1007/BF01448839

    1928 doi

  14. [22]

    Recent progress on advanced imaging techniques for lithium-ion batteries

    Zhe Deng, Xing Lin, Zhenyu Huang, Jintao Meng, Yun Zhong, Guangting Ma, Yu Zhou, Yue Shen, Han Ding, and Yunhui Huang. Recent progress on advanced imaging techniques for lithium-ion batteries. Advanced Energy Materials, 11(2):2000806, 2021

    2021

  15. [23]

    CoNFiLD: Conditional Neural Field Latent Diffusion Model Generating Spatiotemporal Turbulence, March 2024

    Pan Du, Meet Hemant Parikh, Xiantao Fan, Xin-Yang Liu, and Jian-Xun Wang. CoNFiLD: Conditional Neural Field Latent Diffusion Model Generating Spatiotemporal Turbulence, March 2024

    2024

  16. [24]

    Machine learning-assisted multi-objective optimization of battery manufacturing from synthetic data generated by physics-based simulations

    Marc Duquesnoy, Chaoyue Liu, Diana Zapata Dominguez, Vishank Kumar, Elixabete Ayerbe, and Alejandro A Franco. Machine learning-assisted multi-objective optimization of battery manufacturing from synthetic data generated by physics-based simulations. Energy Storage Materials, 5...

    2023

  17. [25]

    Conditional diffusion-based microstructure reconstruction

    Christian Düreth, Paul Seibert, Dennis Rücker, Stephanie Handford, Markus Kästner, and Maik Gude. Conditional diffusion-based microstructure reconstruction. Materials Today Communica- tions, 35:105608, 2023. ISSN 2352-4928. doi: https://doi.org/10.1016/j.mtcomm.2023.105608. UR...

    2023

  18. [26]

    Über die von der molekularkinetischen Theorie der Wärme geforderte Bewe- gung von in ruhenden Flüssigkeiten suspendierten Teilchen

    Albert Einstein. Über die von der molekularkinetischen Theorie der Wärme geforderte Bewe- gung von in ruhenden Flüssigkeiten suspendierten Teilchen. Annalen der Physik, vol. 4, t. 17, 1905

    1905

  19. [27]

    User-defined Event Sampling and Uncertainty Quantification in Diffusion Models for Physical Dynamical Systems

    Marc Anton Finzi, Anudhyan Boral, Andrew Gordon Wilson, Fei Sha, and Leonardo Zepeda- Nunez. User-defined Event Sampling and Uncertainty Quantification in Diffusion Models for Physical Dynamical Systems. In Proceedings of the 40th International Conference on Machine Learning, ...

    2023

  20. [28]

    Gardiner

    Crispin W. Gardiner. Stochastic Methods: A Handbook for the Natural and Social Sciences . Number 13 in Springer Series in Synergetics. Springer, Berlin Heidelberg, 4th ed edition, 2009. ISBN 978-3-642-08962-6 978-3-540-70712-7. 11

    2009

  21. [29]

    Pores for thought: generative adversarial networks for stochastic reconstruction of 3d multi-phase electrode microstructures with periodic boundaries

    Andrea Gayon-Lombardo, Lukas Mosser, Nigel P Brandon, and Samuel J Cooper. Pores for thought: generative adversarial networks for stochastic reconstruction of 3d multi-phase electrode microstructures with periodic boundaries. npj Computational Materials , 6(1):82, 2020

    2020

  22. [30]

    A general method for numerically simulating the stochastic time evolution of coupled chemical reactions

    Daniel T Gillespie. A general method for numerically simulating the stochastic time evolution of coupled chemical reactions. Journal of Computational Physics, 22(4):403–434, December

  23. [31]

    Gillespie

    Daniel T. Gillespie. Approximate accelerated stochastic simulation of chemically reacting systems. The Journal of Chemical Physics, 115(4):1716–1733, 2001. doi: https://doi.org/10. 1063/1.1378322

    2001

  24. [32]

    Gillespie and Linda R

    Daniel T. Gillespie and Linda R. Petzold. Improved leap-size selection for accelerated stochastic simulation. The Journal of Chemical Physics, 119(16):8229–8234, October 2003. ISSN 0021-

    2003

  25. [33]

    Exact Spatiotemporal Dynamics of Confined Lattice Random Walks in Arbitrary Dimensions: A Century after Smoluchowski and P\’olya

    Luca Giuggioli. Exact Spatiotemporal Dynamics of Confined Lattice Random Walks in Arbitrary Dimensions: A Century after Smoluchowski and P\’olya. Physical Review X, 10(2):021045, May 2020. doi: 10.1103/PhysRevX.10.021045

    2020 doi

  26. [34]

    Porespy: A python toolkit for quantitative analysis of porous media images

    Jeff T Gostick, Zohaib A Khan, Thomas G Tranter, Matthew DR Kok, Mehrez Agnaou, Mohammadamin Sadeghi, and Rhodri Jervis. Porespy: A python toolkit for quantitative analysis of porous media images. Journal of Open Source Software, 4(37):1296, 2019

    2019

  27. [35]

    Bioinspired hierarchical composite design using machine learning: simulation, additive manufacturing, and experiment

    Grace X Gu, Chun-Teh Chen, Deon J Richmond, and Markus J Buehler. Bioinspired hierarchical composite design using machine learning: simulation, additive manufacturing, and experiment. Materials horizons., 5(5):939–945, 2018. ISSN 2051-6355 (electronic)

    2018

  28. [36]

    Denoising diffusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models. In H. Larochelle, M. Ranzato, R. Hadsell, M.F. Balcan, and H. Lin, editors, Advances in Neural Information Processing Systems, volume 33, pages 6840–6851. Curran Associates, Inc., 2020

    2020

  29. [37]

    Grainpaint: A multi-scale diffusion-based generative model for microstructure reconstruction of large-scale objects

    Nathan Hoffman, Cashen Diniz, Dehao Liu, Theron Rodgers, Anh Tran, and Mark Fuge. Grainpaint: A multi-scale diffusion-based generative model for microstructure reconstruction of large-scale objects. Acta Materialia, 288:120784, 2025. ISSN 1359-6454. doi: https://doi. org/10.10...

    2025

  30. [38]

    Blurring Diffusion Models

    Emiel Hoogeboom and Tim Salimans. Blurring Diffusion Models. InThe Eleventh International Conference on Learning Representations, September 2022

    2022

  31. [39]

    Argmax flows and multinomial diffusion: Learning categorical distributions

    Emiel Hoogeboom, Didrik Nielsen, Priyank Jaini, Patrick Forré, and Max Welling. Argmax flows and multinomial diffusion: Learning categorical distributions. In M. Ranzato, A. Beygelz- imer, Y . Dauphin, P.S. Liang, and J. Wortman Vaughan, editors,Advances in Neural Information ...

    2021

  32. [40]

    Tenenbaum, and Sergey Levine

    Michael Janner, Yilun Du, Joshua B. Tenenbaum, and Sergey Levine. Planning with diffusion for flexible behavior synthesis, 2022. URL https://arxiv.org/abs/2205.09991

    2022 arXiv

  33. [41]

    Taufactor 2: A gpu accelerated python tool for microstructural analysis

    Steve Kench, Isaac Squires, and Samuel Cooper. Taufactor 2: A gpu accelerated python tool for microstructural analysis. Journal of Open Source Software, 8(88):5358, 2023

    2023

  34. [42]

    Kingma and Max Welling

    Diederik P. Kingma and Max Welling. Auto-Encoding Variational Bayes. In 2nd International Conference on Learning Representations, ICLR 2014, Banff, AB, Canada, April 14-16, 2014, Conference Track Proceedings, 2014

    2014

  35. [43]

    Learning multiple layers of features from tiny images

    Alex Krizhevsky, Geoffrey Hinton, et al. Learning multiple layers of features from tiny images. 2009

    2009

  36. [44]

    Gregory F. Lawler. Random Walk and the Heat Equation . Number volume 55 in Student Mathematical Library. American Mathematical Society, Providence, R.I, 2010. ISBN 978-0- 8218-4829-6. 12

    2010

  37. [45]

    Yann LeCun, Corinna Cortes, and Christopher J.C. Burges. Mnist handwritten digit database,

  38. [46]

    Microstructure reconstruction using diffusion-based genera- tive models

    Kang-Hyun Lee and Gun Jin Yun. Microstructure reconstruction using diffusion-based genera- tive models. Mechanics of Advanced Materials and Structures, 31(18):4443–4461, 2024

    2024

  39. [47]

    A data-driven framework for designing microstructure of multifunctional composites with deep-learned diffusion-based generative models

    Kang-Hyun Lee, Hyoung Jun Lim, and Gun Jin Yun. A data-driven framework for designing microstructure of multifunctional composites with deep-learned diffusion-based generative models. Engineering Applications of Artificial Intelligence, 129:107590, 2024. ISSN 0952-1976. doi: h...

    2024

  40. [48]

    Simplified porosity measurements

    RW Leonard. Simplified porosity measurements. The Journal of the Acoustical Society of America, 20(1):39–41, 1948

    1948

  41. [49]

    Yaron Lipman, Ricky T. Q. Chen, Heli Ben-Hamu, Maximilian Nickel, and Matthew Le. Flow Matching for Generative Modeling. In The Eleventh International Conference on Learning Representations, September 2022

    2022

  42. [50]

    Deep learning face attributes in the wild

    Ziwei Liu, Ping Luo, Xiaogang Wang, and Xiaoou Tang. Deep learning face attributes in the wild. Proceedings of the IEEE International Conference on Computer Vision (ICCV), pages 3730–3738, 2015

    2015

  43. [51]

    Discrete diffusion modeling by estimating the ratios of the data distribution

    Aaron Lou, Chenlin Meng, and Stefano Ermon. Discrete diffusion modeling by estimating the ratios of the data distribution. In Ruslan Salakhutdinov, Zico Kolter, Katherine Heller, Adrian Weller, Nuria Oliver, Jonathan Scarlett, and Felix Berkenkamp, editors, Proceedings of the ...

    2024

  44. [52]

    Microstructure reconstruction of 2d/3d random materials via diffusion-based deep generative models

    Xianrui Lyu and Xiaodan Ren. Microstructure reconstruction of 2d/3d random materials via diffusion-based deep generative models. Scientific Reports, 14(1):5041, 2024

    2024

  45. [53]

    Gianforcaro, Santiago Orrego, Maobin Yang, and Peter I

    Carolina Montoya, Yu Du, Anthony L. Gianforcaro, Santiago Orrego, Maobin Yang, and Peter I. Lelkes. On the road to smart biomaterials for bone research: definitions, concepts, advances, and outlook. Bone Research, 9(1):12, February 2021. ISSN 2095-6231. doi: 10.1038/s41413-020...

    2021 doi

  46. [54]

    High accuracy capillary network representation in digital rock reveals permeability scaling functions

    Rodrigo F Neumann, Mariane Barsi-Andreeta, Everton Lucas-Oliveira, Hugo Barbalho, Willian A Trevizan, Tito J Bonagamba, and Mathias B Steiner. High accuracy capillary network representation in digital rock reveals permeability scaling functions. Scientific reports, 11(1):11370, 2021

    2021

  47. [55]

    Improved denoising diffusion probabilistic models

    Alexander Quinn Nichol and Prafulla Dhariwal. Improved denoising diffusion probabilistic models. In Marina Meila and Tong Zhang, editors, Proceedings of the 38th International Conference on Machine Learning, volume 139 of Proceedings of Machine Learning Research, pages 8162–81...

    2021

  48. [56]

    Process based reconstruction of sandstones and prediction of transport properties

    Pål-Eric Øren and Stig Bakke. Process based reconstruction of sandstones and prediction of transport properties. Transport in porous media, 46(2):311–343, 2002

    2002

  49. [57]

    A practical model for organic richness from porosity and resistivity logs

    QR Passey, S Creaney, JB Kulla, FJ Moretti, and JD Stroud. A practical model for organic richness from porosity and resistivity logs. AAPG bulletin, 74(12):1777–1794, 1990

    1990

  50. [58]

    Microstructural influence on electrochemical devices

    Peeyush Phogat, Shreya Sharma, Ranjana Jha, and Sukhvir Singh. Microstructural influence on electrochemical devices. In Electrochemical Devices: Principles to Applications , pages 257–306. Springer, 2024

    2024

  51. [59]

    Rotskoff, and Lexing Ying

    Yinuo Ren, Haoxuan Chen, Grant M. Rotskoff, and Lexing Ying. How Discrete and Continuous Diffusion Meet: Comprehensive Analysis of Discrete Diffusion Models via a Stochastic Integral Framework, October 2024

    2024

  52. [60]

    The Fokker-Planck Equation

    Hermann Risken. The Fokker-Planck Equation. Springer-Verlag Berlin Heidelberg, 1984. ISBN 978-3-642-96809-9. 13

    1984

  53. [61]

    Generative modelling with inverse heat dissipation

    Severi Rissanen, Markus Heinonen, and Arno Solin. Generative modelling with inverse heat dissipation. In The Eleventh International Conference on Learning Representations, 2023. doi: 10.48550/arxiv.2206.13397

    2023 doi arXiv

  54. [62]

    Santos and Yen Ting Lin

    Javier E. Santos and Yen Ting Lin. Understanding Denoising Diffusion Probabilistic Models and their Noise Schedules via the Ornstein–Uhlenbeck Process. In NeurIPS 2023 Workshop on Diffusion Models, October 2023

    2023

  55. [63]

    Santos, Zachary R

    Javier E. Santos, Zachary R. Fox, Nicholas Lubbers, and Yen Ting Lin. Blackout diffusion: Generative diffusion models in discrete-state spaces. In Andreas Krause, Emma Brunskill, Kyunghyun Cho, Barbara Engelhardt, Sivan Sabato, and Jonathan Scarlett, editors, Proceedings of th...

    2023

  56. [64]

    Simpli- fied and generalized masked diffusion for discrete data

    Jiaxin Shi, Kehang Han, Zhe Wang, Arnaud Doucet, and Michalis Titsias. Simpli- fied and generalized masked diffusion for discrete data. In A. Globerson, L. Mackey, D. Belgrave, A. Fan, U. Paquet, J. Tomczak, and C. Zhang, editors, Advances in Neu- ral Information Processing Sy...

    2024

  57. [65]

    Materials for electrochemical capacitors

    Patrice Simon and Yury Gogotsi. Materials for electrochemical capacitors. Nature Materials, 7 (11):845–854, 2008. ISSN 1476-4660. doi: 10.1038/nmat2297. URL https://doi.org/10. 1038/nmat2297

    2008 doi

  58. [66]

    Deep unsu- pervised learning using nonequilibrium thermodynamics

    Jascha Sohl-Dickstein, Eric Weiss, Niru Maheswaranathan, and Surya Ganguli. Deep unsu- pervised learning using nonequilibrium thermodynamics. In Francis Bach and David Blei, editors, Proceedings of the 32nd International Conference on Machine Learning, volume 37 of Proceedings...

    2015

  59. [67]

    Denoising diffusion implicit models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denoising diffusion implicit models. In International Conference on Learning Representations, 2021

    2021

  60. [68]

    Generative modeling by estimating gradients of the data distribution

    Yang Song and Stefano Ermon. Generative modeling by estimating gradients of the data distribution. In Advances in Neural Information Processing Systems, pages 11895–11907, 2019

    2019

  61. [69]

    Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole

    Yang Song, Jascha Sohl-Dickstein, Diederik P. Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score-Based Generative Modeling through Stochastic Differential Equations, February 2021

    2021

  62. [70]

    Score-based Continuous- time Discrete Diffusion Models

    Haoran Sun, Lijun Yu, Bo Dai, Dale Schuurmans, and Hanjun Dai. Score-based Continuous- time Discrete Diffusion Models. In The Eleventh International Conference on Learning Representations, September 2022

    2022

  63. [71]

    Binomial leap methods for simulating stochastic chemical kinetics

    Tianhai Tian and Kevin Burrage. Binomial leap methods for simulating stochastic chemical kinetics. The Journal of Chemical Physics , 121(21):10356–10364, December 2004. ISSN 0021-9606. doi: 10.1063/1.1810475

    2004 doi

  64. [72]

    Random heterogeneous materials: microstructure and macroscopic proper- ties

    Salvatore Torquato. Random heterogeneous materials: microstructure and macroscopic proper- ties. Springer, 2002

    2002

  65. [73]

    G. E. Uhlenbeck and L. S. Ornstein. On the theory of the brownian motion. Phys. Rev., 36: 823–841, Sep 1930. doi: 10.1103/PhysRev.36.823. URL https://link.aps.org/doi/10. 1103/PhysRev.36.823

    1930 doi

  66. [74]

    Resolving the discrepancy in tortuosity factor estimation for li-ion battery electrodes through micro-macro modeling and experiment

    Francois LE Usseglio-Viretta, Andrew Colclasure, Aashutosh N Mistry, Koffi Pierre Yao Claver, Fezzeh Pouraghajan, Donal P Finegan, Thomas MM Heenan, Daniel Abraham, Partha P Mukherjee, Dean Wheeler, et al. Resolving the discrepancy in tortuosity factor estimation for li-ion ba...

    2018

  67. [75]

    Stochastic Processes in Physics and Chemistry

    N G Van Kampen. Stochastic Processes in Physics and Chemistry . Elsevier Science B.V ., Amsterdam, 2007

    2007

  68. [76]

    Bovik, H.R

    Zhou Wang, A.C. Bovik, H.R. Sheikh, and E.P. Simoncelli. Image quality assessment: from error visibility to structural similarity. IEEE Transactions on Image Processing, 13(4):600–612,

  69. [77]

    Weber and Erwin Frey

    Markus F. Weber and Erwin Frey. Master equations and the theory of stochastic path integrals. Reports on Progress in Physics, 80(4), 2017. ISSN 00344885. doi: 10.1088/1361-6633/aa5ae2

    2017 doi

  70. [78]

    Bridging discrete and continuous state spaces: exploring the ehrenfest process in time-continuous diffusion models

    Ludwig Winkler, Lorenz Richter, and Manfred Opper. Bridging discrete and continuous state spaces: exploring the ehrenfest process in time-continuous diffusion models. In Proceedings of the 41st International Conference on Machine Learning, ICML’24. JMLR.org, 2024

    2024

  71. [79]

    Phase equilibria and thermodynamic properties in the fe-cr system

    Wei Xiong, Malin Selleby, Qing Chen, Joakim Odqvist, and Yong Du. Phase equilibria and thermodynamic properties in the fe-cr system. Critical Reviews in Solid State and Materials Sciences, 35(2):125–152, 2010

    2010

  72. [80]

    L. Zhu, B. Bijeljic, and M.J. Blunt. Diffusion model-based generation of three-dimensional multiphase pore-scale images. Transport in Porous Media , 152:22, 2025. doi: 10.1007/ s11242-025-02158-4. URL https://doi.org/10.1007/s11242-025-02158-4 . 15 A Deriving reverse-time tran...

    2025 doi

  73. [85]

    1"s, while very high intensities produce over- saturated, bold structures such as

    is d dtpk,ℓ(t) =rpk−1,ℓ(t) +rpk+1,ℓ(t) +rpk,ℓ−1(t) +rpk,ℓ+1(t)− 4rpk,ℓ(t). (11) Here, we use periodic boundary conditions,pi±N,ℓ =pk,ℓ,pk,ℓ±N =pk,ℓ over the domain (k,ℓ )∈ {1,...N x}×{ 1,...N y}. We can formally treat pk,l as a rank-2 tensor. We can write the one- dimensional ...

    2000

  74. [1976]

    doi: 10.1016/0021-9991(76)90041-3

    ISSN 0021-9991. doi: 10.1016/0021-9991(76)90041-3

    doi

  75. [2004]

    doi: 10.1109/TIP.2003.819861

    2003

  76. [2010]

    URL http://yann.lecun.com/exdb/mnist/

  77. [9606]

    doi: 10.1063/1.1613254

    doi

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.