Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

From Graph Diffusion to Graph Classification

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Pith's one-line read A score-based graph diffusion model, trained with a softmax over per-class ELBOs and combined with permutation majority voting, achieves state-of-the-art graph classification accuracy on the evaluated benchmarks.

desk verdict A clear sign error in the central ELBO inequality guts the stated objective's formal justification, but the graph-diffusion-classifier idea is novel and worth a corrected resubmission. read the letter →

arxiv 2411.17236 v1 pith:SXIGU73K submitted 2024-11-26 cs.LG cs.AI

classification cs.LGcs.AI
keywords graphclassificationdiffusionmodelsgenerativeclassifiersscore-basedELBOpermutationinvarianceneuralnetworks
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 asks whether generative graph diffusion models can classify graphs as well as discriminative graph neural networks. It finds that a purely generative training objective does not work for classification, so it introduces a discriminative objective: train the diffusion model so that the per-class ELBOs, placed through a softmax, predict the class. With this objective, random permutation augmentation at training and majority voting over permutations at inference, the model beats the GNN baselines on IMDB-BINARY, IMDB-BINARY with node features, and PROTEINS, and reaches perfect accuracy on a synthetic K-regular graph task. The result matters because it shows a generative model can match or exceed discriminative classifiers in the graph domain, opening a new route for graph classification.

What carries the argument

The load-bearing object is the class-conditional denoising ELBO $L_{DEN}(A,y,\theta)$, a variational lower bound on the log-likelihood of an adjacency matrix under a class-conditioned score-based diffusion model. The paper inserts the vector of per-class ELBOs into a softmax (Eq. 16) to form $L_{CLF}$, so training pushes the relative ELBO values to rank the true class first, and the same vector is used for approximate inference. Around this, random permutation sampling replaces one adjacency matrix by samples from its isomorphism class, with predictions aggregated by majority vote, compensating for the non-permutation-invariant SwinGNN backbone.

What would settle it

For a set of held-out graphs, compute the exact class-conditional log-likelihoods via the ODE solver (Eq. 20) and the per-class ELBOs $L_{DEN}$; if the two produce different class rankings on a non-negligible fraction of graphs, the central approximation is falsified. An even simpler check: the same model should fail to beat a degree-histogram or graphlet-count baseline on a dataset where those features separate classes.

Watch

Extended reading notes

Core claim

The central claim is that a score-based graph diffusion model can act as a generative classifier that outperforms discriminative GNN baselines, provided classification is driven by the lower-bound objective $L_{CLF}$ rather than by the generative ELBO. The paper proposes approximating the true class-conditional log-likelihood $\ln p(A|y)$ with the denoising ELBO $L_{DEN}(A,y,\theta)$ from the diffusion model, then defining the training loss as the softmax cross-entropy over the vector of per-class ELBOs. At inference it estimates $\ln p(A|y) \approx -L_{CLF}(A,y,\theta)$ (approximate inference) or computes the exact likelihood via ODE integration. On the evaluated datasets, the combination of $L_{CLF}$ training with approximate inference and permutation majority voting achieves the best accuracy among the six design choices and exceeds the GNN baselines from Errica et al. (2019), for example 75.4% on PROTEINS and 72.6% on IMDB-BINARY with features. The paper also reports that purely generatively trained diffusion models, matched with exact likelihood inference, perform poorly as zero-shot classifiers.

Load-bearing premise

The per-class ELBO is a tight enough lower bound on the true class-conditional log-likelihood that sorting the ELBOs across classes reproduces the true posterior ranking; if the bound is loose or orders classes differently, the $L_{CLF}$ objective stops tracking classification accuracy.

Editorial extensions

If this is right

  • A generative graph diffusion model trained with $L_{CLF}$ can serve as a standalone classifier that beats the GNN baselines on IMDB-BINARY, IMDB-BINARY with features, and PROTEINS.
  • Approximate inference with the $L_{CLF}$ objective is over 2000 times faster for checkpoint selection than exact ODE-based likelihood, making model selection feasible in practice.
  • Increasing the number of inference-time permutations from 1 to 5 yields 2 to 10 percentage points of accuracy gain, with further gains saturating around 100 permutations.
  • Purely generative training with the $L_{DEN}$ objective and exact likelihood gives weak classification accuracy, so the discriminative objective is necessary for classification.

Reading between the lines

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

  • If the ELBO ranks classes correctly on larger benchmark suites, the same recipe could make diffusion-based generative classifiers competitive on molecular or social graph datasets without task-specific GNN design.
  • A direct test of the ranking property would compare per-class ELBOs against exact ODE log-likelihoods on held-out graphs; disagreement would break the approximation.
  • The permutation majority voting suggests a calibration strategy: variance across permutations could serve as a proxy for classification uncertainty, which the paper does not explore.
  • Fine-tuning a pre-trained generative diffusion model with $L_{CLF}$, which the paper names as future work, would test whether generation quality and classification accuracy can be obtained jointly in one model.
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 / 5 minor

Summary. The paper adapts the SwinGNN score-based graph diffusion model to graph classification. It introduces a class-conditional diffusion model trained with three objectives (the pure denoising loss LDEN, a discriminative loss LCLF, and their sum LSUM) and combines them with two inference schemes (approximate inference using the variational loss and ODE-based likelihood computation), together with random permutation augmentation at both training and test time and majority voting over permutations. Experiments on K-Regular, IMDB-BINARY, IMDB-BINARY with features, and PROTEINS report that LCLF trained with approximate inference and permutation voting achieves the best accuracy on all four dataset configurations and is state of the art compared with GNN baselines from Errica et al. (2019).

Significance. If the results hold, the paper offers a useful demonstration that class-conditional graph diffusion models can serve as generative classifiers for graph-level tasks, complementing analogous results in the image domain. The permutation-based augmentation and the use of the variational loss for efficient checkpoint selection are practical ideas that could benefit future work. However, the central theoretical derivation contains a sign error in the ELBO that breaks the formal connection between the proposed objective and likelihood-based classification, and the empirical protocol deviates from the GNN baselines in several ways. The paper does not release code or provide statistical significance tests, so the state-of-the-art claim is not fully supported.

major comments (3)
  1. [Section 4, Eqs. (15)-(18), and Section 5, Eq. (19)] The ELBO inequality is given in the wrong direction. For a denoising diffusion model, the evidence lower bound has the form log p(x) >= -E[||D_theta(x_tilde, t) - x||^2] + C, so the denoising loss LDEN is an upper bound on the negative log-likelihood, not a lower bound on the log-likelihood as written in Eq. (15). Consequently, placing LDEN directly inside the softmax in Eq. (16) makes minimizing LCLF push the true class's denoising loss upward relative to other classes, which is the opposite of likelihood-based classification. Eq. (18) therefore does not establish that LCLF is a lower bound on the training cross-entropy, and Eq. (19) is also inconsistent because -LCLF is a posterior log-probability, not a class-conditional log-likelihood. The empirical success of LCLF in Table 2 suggests that the implementation actually feeds -LDEN into the softmax; if so, the equations must be corrected and the corrected objective must be stated explicitly. As written, the formal justification of the central claim is internally inconsistent.
  2. [Section 6 and Appendix A.1.2] The evaluation protocol is not directly comparable to the baselines. Hyperparameters are selected on a single fold and then fixed for all folds, only one training run per fold is performed instead of three as in the Errica et al. protocol, and graphs above a size cutoff are removed from training and then counted as incorrect at test time. Each of these deviations can change accuracy by more than the small margins over the best baselines in Table 1 (roughly 0.6-1.7 points). To support the state-of-the-art claim, the authors should either follow the reference protocol exactly or quantify the effect of each deviation, for example by running per-fold hyperparameter selection and multiple seeds on at least one dataset.
  3. [Section 6.1, Table 1] The reported improvements over the best GNN baselines are within one standard deviation for IMDB-BINARY (70.5 +/- 5.7 vs. 69.9 +/- 4.6) and PROTEINS (75.4 +/- 3.4 vs. 73.7 +/- 3.5), and no significance tests or per-fold win/loss counts are reported. Calling these results state-of-the-art overstates the statistical evidence. I recommend reporting paired per-fold accuracies with a significance test and tempering the abstract's claim accordingly.
minor comments (5)
  1. [Section 2] The sentence 'For non-graph i.i.d. data, it is well known that a classification model can be derived from a generative model (?)' contains a missing citation placeholder; please insert the intended reference.
  2. [Section 4, Eq. (14)] The surrounding text says 'for all i' but the formula uses the index j; please unify the notation for class labels.
  3. [Appendix A.1.2] Exact inference is described as using t=4 rather than integrating to the maximum noise level; given this truncation, calling the ODE-based procedure 'exact inference' in Section 5 and Table 2 is misleading. Please either rename it or justify why the truncated likelihood is sufficient.
  4. [Throughout] There are numerous typos and formatting issues, including 'repamatrize', 'ususally', 'trianing', 'permutaiton', the uncited '?' in Section 2, and the garbled 'ORGNN' in Figure 2; these should be corrected.
  5. [Section 6.2, Figure 4] Figure 4 shows accuracy curves without error bars or per-fold variability, although Section 6.2 notes that the effects depend on the data split; adding confidence bands would make the trend more informative.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the training and evaluation loop is a standard supervised classification setup, and the self-citations are architecture choices, not load-bearing derivations.

full rationale

The paper derives a discriminative training objective LCLF from a denoising loss LDEN via Bayes' rule and a claimed ELBO bound. The approximate inference rule (Eq. 19) is explicitly defined to use the same LCLF objective, so the classifier is the training objective applied to held-out graphs; this is not circular because test labels are not used in computing LDEN, and the reported accuracy is measured on held-out folds. The use of SwinGNN as backbone cites prior work with overlapping authorship, but the architecture is used as an implementation choice, not as the justification of the main claim. Checkpoint selection uses validation accuracy, which is standard model selection. The formal issue flagged in the manuscript—the inequality direction in Eq. (15), the sign inconsistency between Eqs. (16) and (18), and Eq. (19) equating a log-likelihood with a log-posterior—is a correctness or sign-error concern, not a circularity: the empirical results could still be valid for the implemented objective, but the stated derivation would not justify them as written. No step reduces a prediction to its own input by construction.

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

The paper introduces no new mathematical or physical entities. It borrows the SwinGNN architecture and the EDM diffusion schedule from prior work. The central methodological assumptions are that the ELBO is a good enough likelihood surrogate for graph classification and that permutation sampling can compensate for the non-equivariance of the backbone. The hand-chosen hyperparameters (permutation count, ODE horizon, graph size cutoff) affect the reported accuracy numbers.

free parameters (3)
  • Inference-time permutation count P = 300
    The best reported results use P=300 permutations; the authors show accuracy saturates around 100, but the exact value is chosen by hand after observing test trends, so it is a hand-selected hyperparameter rather than a prediction.
  • ODE integration horizon for exact inference = t=4 instead of sigma_max=80
    For exact likelihood, the ODE solver is run only up to t=4 because the authors find larger noise levels less accurate and more costly; this truncation changes the likelihood estimate and is an ad hoc choice.
  • Graph size cutoff = 128 nodes (IMDB-BINARY), 192 nodes (PROTEINS)
    Graphs larger than the cutoff are omitted from training and counted as misclassified at test time; the cutoff is chosen by hand as a memory and efficiency trade-off, affecting the evaluated graph population.
assumptions (4)
  • standard math Bayes' theorem and the softmax identity p(y|A) = softmax(log p(A|y) + log p(y))
    Used in Eqs. (13)-(14) to derive class probabilities from class-conditional likelihoods.
  • domain assumption The ELBO of the diffusion model is a valid lower bound on the class-conditional log-likelihood
    Eq. (15) assumes the denoising loss LDEN bounds ln p(A|y); the tightness of this bound for binary adjacency matrices and SwinGNN is not verified.
  • domain assumption The probability flow ODE from Song et al. (2021) gives exact likelihoods for the continuous diffusion process on adjacency matrices
    Eq. (20) adapts the ODE likelihood formula; this assumes continuous Gaussian diffusion over binary adjacency matrices is a well-specified density model for graphs.
  • domain assumption Random permutations during training and majority voting at inference approximate the true isomorphism-invariant class probability
    Section 5 and Algorithm 1 state p(G|y) is approximated by sampling from the isomorphism class of the adjacency matrix, but no convergence guarantee or bias analysis is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of From Graph Diffusion to Graph Classification." pith.science (2026). https://pith.science/paper/SXIGU73K

@misc{pith2026241117236,
  author       = {Pith},
  title        = {Pith review of: From Graph Diffusion to Graph Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SXIGU73K}},
  note         = {Machine review of arXiv:2411.17236}
}
read the original abstract

Generative models such as diffusion models have achieved remarkable success in state-of-the-art image and text tasks. Recently, score-based diffusion models have extended their success beyond image generation, showing competitive performance with discriminative methods in image {\em classification} tasks~\cite{zimmermann2021score}. However, their application to classification in the {\em graph} domain, which presents unique challenges such as complex topologies, remains underexplored. We show how graph diffusion models can be applied for graph classification. We find that to achieve competitive classification accuracy, score-based graph diffusion models should be trained with a novel training objective that is tailored to graph classification. In experiments with a sampling-based inference method, our discriminative training objective achieves state-of-the-art graph classification accuracy.

Figures

Figures reproduced from arXiv: 2411.17236 by the authors.

Figure 1
Figure 1. The overall architecture of our model, adapted from SwinGNN (Yan et al., 2023), processes graph inputs by first transforming them into noisy adjacency matrices. These matrices are then concatenated with label embeddings and passed through multiple layers of the model. After processing through these blocks, the model outputs a denoised, clean adjacency matrix. This clean matrix is subsequently transformed back into g… view at source ↗
Figure 2
Figure 2. Overview of training and testing procedure in our method. Train Time: The input adjacency matrix is first randomly permuted, then input to a GNN along with all possible class labels (separately, but in the same model). The loss LDEN is then computed based on the output of the ground-truth class, the LCLF is computed based on the output of all the labels, and LSUM is the sum of two losses. Test Time: The adjacency ma… view at source ↗
Figure 3
Figure 3. The mean accuracy curve during training using LCLF on IMDB-BINARY dataset when training with or without random permutation of the adjacency matrix. Training with permutations offers significant gains in accuracy. 0 100 200 300 400 500 # Permutations 0.60 0.65 0.70 Accuracy LSUM LDEN LCLF [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: The accuracy curve of varying the number of permu￾tations across three training objectives (LCLF, LDEN, and LSUM) with approximate inference. In general, increasing the number of inference-time permutations improves performance, but the gains saturate around 100 permut…
Figure 5
Figure 5. Figure 5: The mean validation accuracy curve during training using LDEN on IMDB-BINARY dataset when taining with or without random permutation of the adjacency matrix. The blue line is training without permutation, the red line is training with permutation 0 2000 4000 6000 8000 …
Figure 6
Figure 6. Figure 6: Similar as [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]
Figure 7
Figure 7. Figure 7: The test accuracy curve of the three objective using Exact inference for different number of permutation. The yellow, brown and green lines are LCLF, LDEN and LSUM respectively [PITH_FULL_IMAGE:figures/full_fig_p014_7.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. DiffImaginE: Imagine to Verify Entity Types with Diffusio

    cs.AI 2026-08 conditional novelty 6.0 of 10

    A diffusion-based verifier scores candidate entity types by conditional denoising error and beats a matched deterministic control on two Twitter MNER benchmarks.

Reference graph

Works this paper leans on

33 extracted references · 15 canonical work pages · cited by 1 Pith paper

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    M., Ong, C

    Borgwardt, K. M., Ong, C. S., Schönauer, S., Vishwanathan, S. V. N., Smola, A. J., and Kriegel, H.-P. Protein function prediction via graph kernels . Bioinformatics, 21 0 (suppl1): 0 i47--i56, 06 2005

  3. [3]

    Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J. D., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., Agarwal, S., Herbert-Voss, A., Krueger, G., Henighan, T., Child, R., Ramesh, A., Ziegler, D., Wu, J., Winter, C., Hesse, C., Chen, M., Sigler, E., Litwin, M., Gray, S., Chess, B., Clark, J., Berner, C., McCandlish, S., Radford, A.,...

  4. [4]

    I., Bronstein, M., Webb, S., and Rossi, E

    Chamberlain, B., Rowbottom, J., Gorinova, M. I., Bronstein, M., Webb, S., and Rossi, E. Grand: Graph neural diffusion. In Meila, M. and Zhang, T. (eds.), Proceedings of the 38th International Conference on Machine Learning, volume 139 of Proceedings of Machine Learning Research, pp.\ 1407--1418. PMLR, 18--24 Jul 2021. URL https://proceedings.mlr.press/v13...

  5. [5]

    Diffdock: Diffusion steps, twists, and turns for molecular docking, 2023

    Corso, G., Stärk, H., Jing, B., Barzilay, R., and Jaakkola, T. Diffdock: Diffusion steps, twists, and turns for molecular docking, 2023. URL https://arxiv.org/abs/2210.01776

  6. [6]

    A fair comparison of graph neural networks for graph classification

    Errica, F., Podda, M., Bacciu, D., and Micheli, A. A fair comparison of graph neural networks for graph classification. ArXiv, abs/1912.09893, 2019. URL https://api.semanticscholar.org/CorpusID:209439835

  7. [7]

    L., Ying, R., and Leskovec, J

    Hamilton, W. L., Ying, R., and Leskovec, J. Inductive representation learning on large graphs. In Proceedings of the 31st International Conference on Neural Information Processing Systems, NIPS'17, pp.\ 1025–1035, Red Hook, NY, USA, 2017. Curran Associates Inc. ISBN 9781510860964

  8. [8]

    Denoising diffusion probabilistic models, 2020

    Ho, J., Jain, A., and Abbeel, P. Denoising diffusion probabilistic models, 2020. URL https://arxiv.org/abs/2006.11239

Show all 33 references
  1. [9]

    G., Frossard, P., Welling, M., Bronstein, M., and Correia, B

    Igashov, I., Stärk, H., Vignac, C., Satorras, V. G., Frossard, P., Welling, M., Bronstein, M., and Correia, B. Equivariant 3d-conditional diffusion models for molecular linker design, 2022. URL https://arxiv.org/abs/2210.05274

  2. [10]

    Jo, J., Lee, S., and Hwang, S. J. Score-based generative modeling of graphs via the system of stochastic differential equations. In Chaudhuri, K., Jegelka, S., Song, L., Szepesvari, C., Niu, G., and Sabato, S. (eds.), Proceedings of the 39th International Conference on Machine...

  3. [11]

    Elucidating the design space of diffusion-based generative models, 2022

    Karras, T., Aittala, M., Aila, T., and Laine, S. Elucidating the design space of diffusion-based generative models, 2022. URL https://arxiv.org/abs/2206.00364

  4. [12]

    P., Salimans, T., Poole, B., and Ho, J

    Kingma, D. P., Salimans, T., Poole, B., and Ho, J. Variational diffusion models, 2023. URL https://arxiv.org/abs/2107.00630

  5. [13]

    A., and Zhang, C

    Kong, L., Cui, J., Sun, H., Zhuang, Y., Prakash, B. A., and Zhang, C. Autoregressive diffusion model for graph generation. In Krause, A., Brunskill, E., Cho, K., Engelhardt, B., Sabato, S., and Scarlett, J. (eds.), Proceedings of the 40th International Conference on Machine Le...

  6. [14]

    Voicebox: Text-guided multilingual universal speech generation at scale

    Le, M., Vyas, A., Shi, B., Karrer, B., Sari, L., Moritz, R., Williamson, M., Manohar, V., Adi, Y., Mahadeokar, J., and Hsu, W.-N. Voicebox: Text-guided multilingual universal speech generation at scale. In Thirty-seventh Conference on Neural Information Processing Systems, 202...

  7. [15]

    C., Prabhudesai, M., Duggal, S., Brown, E., and Pathak, D

    Li, A. C., Prabhudesai, M., Duggal, S., Brown, E., and Pathak, D. Your diffusion model is secretly a zero-shot classifier. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pp.\ 2206--2217, October 2023

  8. [16]

    Ng, A. Y. and Jordan, M. I. On discriminative vs. generative classifiers: A comparison of logistic regression and naive bayes. In NIPS, volume 14, pp.\ 841--848. MIT Press, 2001

  9. [17]

    OpenAI, Achiam, J., Adler, S., Agarwal, S., Ahmad, L., Akkaya, I., Aleman, F. L., Almeida, D., Altenschmidt, J., Altman, S., Anadkat, S., Avila, R., Babuschkin, I., Balaji, S., Balcom, V., Baltescu, P., Bao, H., Bavarian, M., Belgum, J., Bello, I., Berdine, J., Bernadett-Shapi...

  10. [18]

    and Narasimhan, K

    Radford, A. and Narasimhan, K. Improving language understanding by generative pre-training. 2018. URL https://api.semanticscholar.org/CorpusID:49313245

  11. [19]

    Language models are unsupervised multitask learners

    Radford, A., Wu, J., Child, R., Luan, D., Amodei, D., and Sutskever, I. Language models are unsupervised multitask learners. 2019

  12. [20]

    High-resolution image synthesis with latent diffusion models, 2021

    Rombach, R., Blattmann, A., Lorenz, D., Esser, P., and Ommer, B. High-resolution image synthesis with latent diffusion models, 2021

  13. [21]

    L., Ghasemipour, K., Gontijo Lopes, R., Karagol Ayan, B., Salimans, T., Ho, J., Fleet, D

    Saharia, C., Chan, W., Saxena, S., Li, L., Whang, J., Denton, E. L., Ghasemipour, K., Gontijo Lopes, R., Karagol Ayan, B., Salimans, T., Ho, J., Fleet, D. J., and Norouzi, M. Photorealistic text-to-image diffusion models with deep language understanding. In Koyejo, S., Mohamed...

  14. [22]

    and Komodakis, N

    Simonovsky, M. and Komodakis, N. Dynamic edge-conditioned filters in convolutional neural networks on graphs. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), July 2017

  15. [23]

    P., Kumar, A., Ermon, S., and Poole, B

    Song, Y., Sohl-Dickstein, J., Kingma, D. P., Kumar, A., Ermon, S., and Poole, B. Score-based generative modeling through stochastic differential equations. In International Conference on Learning Representations, 2021. URL https://openreview.net/forum?id=PxTIG12RRHS

  16. [24]

    Digress: Discrete denoising diffusion for graph generation

    Vignac, C., Krawczuk, I., Siraudin, A., Wang, B., Cevher, V., and Frossard, P. Digress: Discrete denoising diffusion for graph generation. In The Eleventh International Conference on Learning Representations, 2023. URL https://openreview.net/forum?id=UaAD-Nu86WX

  17. [25]

    A connection between score matching and denoising autoencoders

    Vincent, P. A connection between score matching and denoising autoencoders. Neural Computation, 23 0 (7): 0 1661--1674, 2011. doi:10.1162/NECO_a_00142

  18. [26]

    E., Bronstein, M

    Wang, Y., Sun, Y., Liu, Z., Sarma, S. E., Bronstein, M. M., and Solomon, J. M. Dynamic graph cnn for learning on point clouds. ACM Transactions on Graphics (TOG), 2019

  19. [27]

    How powerful are graph neural networks? In International Conference on Learning Representations, 2019 a

    Xu, K., Hu, W., Leskovec, J., and Jegelka, S. How powerful are graph neural networks? In International Conference on Learning Representations, 2019 a . URL https://openreview.net/forum?id=ryGs6iA5Km

  20. [28]

    How powerful are graph neural networks? In International Conference on Learning Representations, 2019 b

    Xu, K., Hu, W., Leskovec, J., and Jegelka, S. How powerful are graph neural networks? In International Conference on Learning Representations, 2019 b . URL https://openreview.net/forum?id=ryGs6iA5Km

  21. [29]

    Geodiff: a geometric diffusion model for molecular conformation generation, 2022

    Xu, M., Yu, L., Song, Y., Shi, C., Ermon, S., and Tang, J. Geodiff: a geometric diffusion model for molecular conformation generation, 2022. URL https://arxiv.org/abs/2203.02923

  22. [30]

    Swingnn: Rethinking permutation invariance in diffusion models for graph generation, 2023

    Yan, Q., Liang, Z., Song, Y., Liao, R., and Wang, L. Swingnn: Rethinking permutation invariance in diffusion models for graph generation, 2023

  23. [31]

    and Vishwanathan, S

    Yanardag, P. and Vishwanathan, S. Deep graph kernels. In Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD '15, pp.\ 1365–1374, New York, NY, USA, 2015. Association for Computing Machinery. ISBN 9781450336642. doi:10.1145/2...

  24. [32]

    Hierarchical graph representation learning with differentiable pooling

    Ying, Z., You, J., Morris, C., Ren, X., Hamilton, W., and Leskovec, J. Hierarchical graph representation learning with differentiable pooling. In Bengio, S., Wallach, H., Larochelle, H., Grauman, K., Cesa-Bianchi, N., and Garnett, R. (eds.), Advances in Neural Information Proc...

  25. [33]

    S., Schott, L., Song, Y., Dunn, B

    Zimmermann, R. S., Schott, L., Song, Y., Dunn, B. A., and Klindt, D. A. Score-based generative classifiers. CoRR, abs/2110.00473, 2021

Pith tools

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