Pith. sign in

REVIEW 4 major objections 4 minor 49 references

GNN-Transformer Cooperative Architecture for Trustworthy Graph Contrastive Learning

T0 review · 4 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Triple neighborhood consensus yields trustworthy graph positive pairs and top or near-top node-classification accuracy.

desk verdict The core recipe is real, but the trustworthiness proof is circular; send to review with a demand for a rewritten theory section. read the letter →

arxiv 2412.16218 v4 pith:67V3FI5V submitted 2024-12-18 cs.LG

Jianqing Liang , Xinkai Wei , Min Chen , Zhiqiang Wang , Jiye Liang This is my paper · ORCID
classification cs.LG
keywords graphcontrastivelearningaugmentation-freeneuralnetworkstransformerspositivepairsamplingk-nearestneighborsnodeclassificationtrustworthyrepresentation
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 tries to establish that the best positive pairs for graph contrastive learning come from consensus rather than random augmentation. GTCA defines a positive set for each node as the intersection of three k-nearest-neighbour lists: one from a GNN embedding, one from a Transformer embedding, and one from graph hop distance, so nodes must agree across a local, a global, and a topological view. The loss gives each anchor multiple positives instead of the usual single one. The paper reports a multi-view learning bound, a proposition that the intersection produces the most trustworthy positive pairs, and node-classification results on five benchmarks where GTCA ranks first three times and second twice.

What carries the argument

The central object is the positive set $\mathcal{P}_i = \mathcal{B}^\theta_i \cap \mathcal{B}^\varphi_i \cap \mathcal{T}_i$, where $\mathcal{B}^\theta_i$ and $\mathcal{B}^\varphi_i$ are the k nearest neighbours of node $v_i$ in the GCN and NodeFormer embedding spaces and $\mathcal{T}_i$ is its k nearest neighbours by hop count. This triple intersection is the mechanism: it converts two representation views and one topology view into multiple positives per anchor without random augmentation. Proposition 1 argues that the intersection has minimum cardinality, hence the smallest numerator in the InfoNCE-style loss of Equation (7), making the loss an upper bound whose minimization yields trustworthy contrastive learning.

What would settle it

Run GTCA on a heterophilous graph where adjacent nodes often have different classes and record the fraction of nodes in $\mathcal{P}_i$ that share the anchor's ground-truth label. If that fraction is not clearly above random choice, or if the method's classification accuracy does not beat a plain GCN on such graphs, then the claim that Equation (5) generates the most trustworthy positive pairs is empirically false.

Watch

Extended reading notes

Core claim

The paper's central claim is that for graph contrastive learning, the most trustworthy positive pairs of a node are the consensus of three independent neighbourhoods: its k nearest neighbours in a GCN embedding, its k nearest neighbours in a NodeFormer embedding, and its k nearest neighbours by graph hops. Proposition 1 states that Equation (5), the intersection of these three sets, generates the most trustworthy positive pairs because it contains the fewest candidates, which minimizes the numerator of the contrastive loss and turns the loss into an upper bound; minimizing that upper bound is what the paper calls trustworthy GCL. Empirically, the intersection's label-correctness exceeds 80 percent on Cora and Amazon-Photo, while either encoder alone stays below 40 percent, and GTCA reports top or near-top accuracy on all five datasets.

Load-bearing premise

The load-bearing premise is that a node close to another in both learned embedding spaces and in graph hops probably shares the same class; if that consensus is wrong or empty, for example in heterophilous graphs or early in training, the positive pairs are unreliable and the proposition's bound does not guarantee label-correctness.

Editorial extensions

If this is right

  • Removing random augmentation avoids the main failure mode of GCL: altering a graph can change the meaning of a node or molecule, as the paper's opening example illustrates.
  • Each anchor gets $2|\mathcal{P}_i|+1$ positives instead of a single one, so the loss pulls multiple similar nodes together rather than pushing all non-positives away.
  • On Cora, Coauthor-CS, and Amazon-Photo, GTCA reports the highest accuracy among the 12 baselines, and it ranks second on Wiki-CS and Amazon-Computers.
  • The intersection rule raises positive-pair correctness above 80 percent on Cora and Amazon-Photo, versus below 40 percent for a single encoder and 30–60 percent for two encoders without the topology view.
  • Ablations show that dropping the topology view or replacing one encoder type with a duplicate of the other lowers accuracy, so the local GNN view, the global Transformer view, and hop distance all contribute.

Reading between the lines

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

  • Beyond the paper: the same consensus rule could be carried to graph-level contrastive learning by replacing hop distance with a structure-aware distance such as diffusion or shortest-path similarity.
  • Beyond the paper: because Proposition 1 bounds the paper's own loss rather than classification error, the trustworthiness claim is currently an empirical property; a label-dependent bound would be a stronger theoretical test.
  • Beyond the paper: on heterophilous graphs, hop-based neighbourhoods can mix classes, so a signed or class-aware topology distance might preserve the consensus idea where vanilla k-hop fails.
  • Beyond the paper: the sensitivity analysis shows k controls a precision–recall trade-off, so an automatic rule for choosing k (for example, where intersection precision stops improving) could replace per-dataset grid search.
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

4 major / 4 minor

Summary. The paper proposes GTCA, an augmentation-free graph contrastive learning method that combines a GCN encoder and a NodeFormer (linear Transformer) encoder. Positive pairs for each anchor node are defined as the intersection of the GCN k-NN set, the NodeFormer k-NN set, and a hop-based topological k-NN set, and the training loss is a multi-positive InfoNCE-style objective. The authors claim that theoretical analysis verifies the trustworthiness of this positive-pair selection and report node-classification accuracies on five benchmark datasets, with ablations and sensitivity analyses.

Significance. The combination of a GNN view and a Transformer view for augmentation-free GCL is a reasonable and timely idea, and the paper is transparent in providing code and ablations. The empirical setup is standard and the method is simple to reproduce. However, the central theoretical claim in the title and abstract is not supported: the proof of Proposition 1 is circular in that it defines trustworthiness in terms of the paper's own loss and then concludes trustworthiness, and Theorem 1 is an imported multi-view bound that is not connected to the proposed intersection sampling. The label-correctness evidence in Figure 4 covers only two homophilous datasets, and the accuracy improvements in Table 3 are mostly within one standard deviation. If the trustworthiness language were removed or replaced with a direct analysis of label-correctness, the paper would be a useful empirical contribution, but in its current form the claims outrun the evidence.

major comments (4)
  1. [Contrastive Loss Function, Proposition 1] The proof of Proposition 1 is circular. The term "trustworthy" is never defined, and the proof only observes that the intersection P_i = B_i^theta ∩ B_i^phi ∩ T_i minimizes the numerator of Eq. (7) and therefore maximizes the loss ℓ. An upper bound on a contrastive loss is not a statement about label correctness: a set of positives can be perfectly aligned with this loss and still contain mostly wrong-class nodes. Since the abstract claims that "theoretical analysis verifies the trustworthiness of the proposed method," this missing external correctness criterion is load-bearing and must be addressed.
  2. [Method, Theorem 1] Theorem 1 is stated without proof and is not connected to the proposed method. It is a generic multi-view risk bound that does not mention k-NN sets, the intersection operation, class labels, or Eq. (5). The notation is also incomplete: r, η, R_N, C, and H∘G_M are not defined. Consequently, Theorem 1 cannot verify the trustworthiness of the positive-pair sampling used in GTCA. Please either remove the trustworthiness claim or supply a theorem that directly bounds label disagreement among sampled positives under stated assumptions.
  3. [Experiments, Table 3 and Figure 4] The empirical support for the central claim is modest. In Table 3, GTCA ranks first on 3 of 5 datasets and second on the other two, but the margins over the strongest baselines are within one standard deviation (e.g., 82.5±1.3 vs 82.2±1.6 on Cora; 92.5±0.6 vs 91.2±0.9 on Coauthor-CS). Figure 4 measures label-correctness of positive pairs only on Cora and Amazon-Photo, which are both homophilous; no heterophilous graph is tested. Because the topology-based k-NN is a central component and is negatively correlated with label agreement under heterophily, the paper should either extend the empirical analysis to heterophilous datasets or explicitly scope the claims.
  4. [Node Sampling and Eq. (7)] The training objective treats every node outside P_i as a negative pair. If P_i contains false positives, the loss actively reinforces them, and true same-class nodes outside P_i are pushed away. No theorem or experiment bounds the fraction of false positives under early uninformative embeddings or under heterophilous topology. Since the method is presented as a general graph contrastive learning approach, this is a substantive gap that should be discussed or closed.
minor comments (4)
  1. [Algorithm 1, line 9] "teturn H for downstream tasks" should be "return H for downstream tasks".
  2. [Ablation Study, Table 4 paragraph] The phrase "ranks recond" should be "ranks second".
  3. [Method, Eq. (2)] The notation in Eq. (2) is difficult to parse (e.g., "egj/τ" appears to be a superscript that is not formatted). Please rewrite the equation so that the exponential terms are unambiguous.
  4. [Conclusion] The conclusion states that the method has "quadratic complexity," but NodeFormer is described earlier as a linear-attention Transformer. Please clarify the overall complexity of GTCA and specify which component is quadratic.

Circularity Check

1 steps flagged · score 8.0 of 10

Proposition 1 proves only that Pi minimizes the loss's numerator; 'trustworthy' is defined by the loss, so the trustworthiness claim is forced by construction.

  1. self definitional [Proposition 1 and its proof, Section 'Contrastive Loss Function' (Eqs. 5, 7, 11)]
    "Proof. It is obvious that the intersection set contains the minimum number of positive pairs with specificBθ i , Bφ i , and Ti. The numerator of Equation (7) sums the exponential calculation with respect to Pi, whereas the denominator sums the exponential calculation with respect to both Pi and Ni. Therefore, when calculatingPi with Equation (5), we get the lower bound of the numerator, and also the upper bound of the contrastive loss in Equation (7). When we minimize the upper bound of the loss in Equation (11), trustworthy GCL can be achieved."

    The proposition's conclusion is that Equation (5) generates 'the most trustworthy positive pairs for GCL,' but the proof never defines trustworthiness by any external criterion such as label agreement or downstream classification correctness. It shows only that Pi, being an intersection of three k-NN sets, makes the numerator of the GTCA loss in Equation (7) as small as possible, thereby upper-bounding that loss. Minimizing the loss is then declared to be 'trustworthy GCL.' Thus the claimed theoretical result reduces to the tautology that the set chosen to minimize the loss upper bound is the set that minimizes the loss upper bound.

full rationale

The central theoretical claim of the paper is circular. Proposition 1 asserts that Equation (5) generates 'the most trustworthy positive pairs,' but the proof equates trustworthiness with minimization of the paper's own contrastive loss: the intersection Pi is shown to lower-bound the numerator and upper-bound Equation (7), and that is taken as proof of trustworthiness. Since 'trustworthy' is never defined with respect to class labels or any external correctness measure, the theorem is true by construction and says nothing about whether Pi contains semantically similar nodes. The separate empirical correct-ratio analysis in Figure 4 is a genuine external check, but it is not part of the theoretical proof and only covers two homophilous datasets. Theorem 1 is an imported generic multi-view risk bound from Huang et al. (2021) and does not mention Pi, k-NN sets, or the intersection operation, so it does not repair the circularity. The empirical comparisons and ablations are self-contained and independently meaningful, so the paper is not wholly circular; however, its flagship 'theoretical analysis verifies trustworthiness' claim is forced by definition, warranting a score of 8.

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

The ledger captures the tuned hyperparameters and the assumptions behind the positive-pair sampling. The main cost is that the trustworthiness claim rests on an assumption about homophily between graph topology, embedding neighborhoods, and labels, which is not guaranteed by the proofs.

free parameters (3)
  • k (neighborhood size) = 240 to 520 per dataset
    Tuned per dataset in Table 2; sensitivity analysis in Figure 5 shows accuracy depends on k.
  • lambda (output interpolation weight) = 0.4 to 0.8 per dataset
    Tuned per dataset in Table 2; Figure 5 shows an optimum near 0.7.
  • embedding dimension E = 400 to 512 per dataset
    Tuned per dataset in Table 2; Figure 5 shows accuracy increases with E.
assumptions (4)
  • domain assumption Graph nodes are drawn i.i.d. from an unknown distribution D.
    Assumed in Theorem 1 and used for the generalization bound; graph nodes are typically interdependent, so this is an idealization.
  • domain assumption Topological k-NN sets T_i reflect semantic similarity.
    Equation (5) intersects T_i with learned k-NN sets; on heterophilous graphs hop proximity is not label proximity, so the positive set can be wrong.
  • domain assumption Learned embeddings H_theta and H_phi are informative enough that their k-NN sets are non-empty and meaningful at every training epoch.
    Positive pairs are sampled from k-NN sets of the encoders being trained, a chicken-and-egg dependence that can collapse if embeddings are poor early in training.
  • standard math Rademacher complexity bounds of the form R_N(F) <= sqrt(C(F)/N) hold for the encoder function classes.
    Invoked in the remark after Theorem 1 to simplify the bound.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GNN-Transformer Cooperative Architecture for Trustworthy Graph Contrastive Learning." pith.science (2026). https://pith.science/paper/67V3FI5V

@misc{pith2026241216218,
  author       = {Pith},
  title        = {Pith review of: GNN-Transformer Cooperative Architecture for Trustworthy Graph Contrastive Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/67V3FI5V}},
  note         = {Machine review of arXiv:2412.16218}
}
read the original abstract

Graph contrastive learning (GCL) has become a hot topic in the field of graph representation learning. In contrast to traditional supervised learning relying on a large number of labels, GCL exploits augmentation strategies to generate multiple views and positive/negative pairs, both of which greatly influence the performance. Unfortunately, commonly used random augmentations may disturb the underlying semantics of graphs. Moreover, traditional GNNs, a type of widely employed encoders in GCL, are inevitably confronted with over-smoothing and over-squashing problems. To address these issues, we propose GNN-Transformer Cooperative Architecture for Trustworthy Graph Contrastive Learning (GTCA), which inherits the advantages of both GNN and Transformer, incorporating graph topology to obtain comprehensive graph representations. Theoretical analysis verifies the trustworthiness of the proposed method. Extensive experiments on benchmark datasets demonstrate state-of-the-art empirical performance.

Figures

Figures reproduced from arXiv: 2412.16218 by the authors.

Figure 1
Figure 1. Augmentations on image (a) keep the semantic [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The architecture of GTCA. Given a graph, the GNN encoder [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The red nodes denote the positive pair and the blue [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Correct ratio of positive pairs with various [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Accuracy vs. hyperparameters k, E and λ on Cora and Amazon-Photo datasets [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Visualization of DGI, NCLA and GTCA embeddings on Cora dataset with PCA. [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 40 canonical work pages

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    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 word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Abdi, H.; and Williams, L. J. 2010. Principal Component Analysis. Wiley Interdisciplinary Reviews: Computational Statistics, 433--459

    work page 2010

  4. [4]

    Garg, V.; Jegelka, S.; and Jaakkola, T. 2020. Generalization and representational limits of graph neural networks. In Proceedings of International Conference on Machine Learning, 3419--3430

    work page 2020

  5. [5]

    Guo, X.; Wang, Y.; Wei, Z.; and Wang, Y. 2023. Architecture matters: Uncovering implicit mechanisms in graph contrastive learning. In Proceedings of Advances in Neural Information Processing Systems, 28585--28610

    work page 2023

  6. [6]

    Hamilton, W.; Ying, Z.; and Leskovec, J. 2017. Inductive representation learning on large graphs. In Proceedings of Advances in Neural Information Processing Systems

    work page 2017

  7. [7]

    Hassani, K.; and Khasahmadi, A. H. 2020. Contrastive multi-view representation learning on graphs. In Proceedings of International Conference on Machine Learning, 4116--4126

    work page 2020

  8. [8]

    D.; Fedorov, A.; Lavoie-Marchildon, S.; Grewal, K.; Bachman, P.; Trischler, A.; and Bengio, Y

    Hjelm, R. D.; Fedorov, A.; Lavoie-Marchildon, S.; Grewal, K.; Bachman, P.; Trischler, A.; and Bengio, Y. 2019. Learning Deep Representations by Mutual Information Estimation and Maximization. In Proceedings of the International Conference on Learning Representations

    work page 2019

Show all 49 references
  1. [9]

    Huang, Y.; Du, C.; Xue, Z.; Chen, X.; Zhao, H.; and Huang, L. 2021. What makes multi-modal learning better than single (provably). 10944--10956

    2021

  2. [10]

    N.; and Welling, M

    Kipf, T. N.; and Welling, M. 2016. Semi-supervised classification with graph convolutional networks. arXiv preprint arXiv:1609.02907

    2016 arXiv

  3. [11]

    N.; and Welling, M

    Kipf, T. N.; and Welling, M. 2017. Semi-supervised classification with graph convolutional networks. In Proceedings of International Conference on Learning Representations

    2017

  4. [12]

    Kreuzer, D.; Beaini, D.; Hamilton, W.; L \'e tourneau, V.; and Tossou, P. 2021. Rethinking graph transformers with spectral attention. In Proceedings of Advances in Neural Information Processing Systems, 21618--21629

    2021

  5. [13]

    Lee, N.; Lee, J.; and Park, C. 2022. Augmentation-free self-supervised learning on graphs. In Proceedings of the AAAI Conference on Artificial Intelligence, 7372--7380

    2022

  6. [14]

    Li, S.; Wang, X.; Zhang, A.; Wu, Y.; He, X.; and Chua, T.-S. 2022. Let invariant rationale discovery inspire graph contrastive learning. In Proceedings of International Conference on Machine Learning, 13052--13065

    2022

  7. [15]

    Liang, J.; Chen, M.; and Liang, J. 2024. Graph External Attention Enhanced Transformer. In Proceedings of International Conference on Machine Learning

    2024

  8. [16]

    Liu, M.; Gao, H.; and Ji, S. 2020. Towards deeper graph neural networks. In Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 338--348

    2020

  9. [17]

    Liu, Y.; Zhao, Y.; Wang, X.; Geng, L.; and Xiao, Z. 2024. Multi-scale subgraph contrastive learning. arXiv preprint arXiv:2403.02719

    2024 arXiv

  10. [18]

    Mernyei, P.; and Cangea, C. 2020. Wiki-cs: A wikipedia-based benchmark for graph neural networks. arXiv preprint arXiv:2007.02901

    2020 arXiv

  11. [19]

    Mo, Y.; Peng, L.; Xu, J.; Shi, X.; and Zhu, X. 2022. Simple unsupervised graph representation learning. In Proceedings of the AAAI Conference on Artificial Intelligence, 7797--7805

    2022

  12. [20]

    Oord, A. v. d.; Li, Y.; and Vinyals, O. 2018. Representation learning with contrastive predictive coding. arXiv preprint arXiv:1807.03748

    2018 arXiv

  13. [21]

    Peng, Z.; Huang, W.; Luo, M.; Zheng, Q.; Rong, Y.; Xu, T.; and Huang, J. 2020. Graph representation learning via graphical mutual information maximization. In Proceedings of The Web Conference 2020, 259--270

    2020

  14. [22]

    Qiu, J.; Chen, Q.; Dong, Y.; Zhang, J.; Yang, H.; Ding, M.; Wang, K.; and Tang, J. 2020. Gcc: Graph contrastive coding for graph neural network pre-training. In Proceedings of the 26th ACM SIGKDD International Conference on knowledge Discovery & Data Mining, 1150--1160

    2020

  15. [23]

    Sen, P.; Namata, G.; Bilgic, M.; Getoor, L.; Galligher, B.; and Eliassi-Rad, T. 2008. Collective classification in network data. AI magazine, 93--93

    2008

  16. [24]

    Shchur, O.; Mumme, M.; Bojchevski, A.; and G \"u nnemann, S. 2018. Pitfalls of graph neural network evaluation. arXiv preprint arXiv:1811.05868

    2018 arXiv

  17. [25]

    Shen, X.; Sun, D.; Pan, S.; Zhou, X.; and Yang, L. T. 2023. Neighbor contrastive learning on learnable graph augmentation. In Proceedings of the AAAI Conference on Artificial Intelligence, 9782--9791

    2023

  18. [26]

    Shorten, C.; and Khoshgoftaar, T. M. 2019. A survey on image data augmentation for deep learning. Journal of Big Data, 1--48

    2019

  19. [27]

    Suresh, S.; Li, P.; Hao, C.; and Neville, J. 2021. Adversarial graph augmentation to improve graph contrastive learning. In Proceedings of Advances in Neural Information Processing Systems, 15920--15933

    2021

  20. [28]

    Tan, S.; Li, D.; Jiang, R.; Zhang, Y.; and Okumura, M. 2024. Community-Invariant Graph Contrastive Learning. In Proceedings of International Conference on Machine Learning

    2024

  21. [29]

    G.; Azabou, M.; Dyer, E

    Thakoor, S.; Tallec, C.; Azar, M. G.; Azabou, M.; Dyer, E. L.; Munos, R.; Veli c kovi \'c , P.; and Valko, M. 2022. Large-scale representation learning on graphs via bootstrapping. In Proceedings of International Conference on Learning Representations

    2022

  22. [30]

    Veli c kovi \'c , P.; Cucurull, G.; Casanova, A.; Romero, A.; Lio, P.; and Bengio, Y. 2017. Graph attention networks. arXiv preprint arXiv:1710.10903

    2017 arXiv

  23. [31]

    L.; Li \`o , P.; Bengio, Y.; and Hjelm, R

    Velickovic, P.; Fedus, W.; Hamilton, W. L.; Li \`o , P.; Bengio, Y.; and Hjelm, R. D. 2019. Deep graph infomax. In Proceedings of International Conference on Learning Representations

    2019

  24. [32]

    Wan, S.; Pan, S.; Yang, J.; and Gong, C. 2021 a . Contrastive and generative graph convolutional networks for graph-based semi-supervised learning. In Proceedings of AAAI Conference on Artificial Intelligence, 10049--10057

    2021

  25. [33]

    Wan, S.; Zhan, Y.; Liu, L.; Yu, B.; Pan, S.; and Gong, C. 2021 b . Contrastive graph poisson networks: Semi-supervised learning with extremely limited labels. In Proceedings of Advances in Neural Information Processing Systems, 6316--6327

    2021

  26. [34]

    Wei, C.; Wang, Y.; Bai, B.; Ni, K.; Brady, D.; and Fang, L. 2023. Boosting graph contrastive learning via graph contrastive saliency. In Proceedings of International Conference on Machine Learning, 36839--36855

    2023

  27. [35]

    P.; and Yan, J

    Wu, Q.; Zhao, W.; Li, Z.; Wipf, D. P.; and Yan, J. 2022. Nodeformer: A scalable graph structure learning transformer for node classification. In Proceedings of Advances in Neural Information Processing Systems, 27387--27401

    2022

  28. [36]

    Wu, Q.; Zhao, W.; Yang, C.; Zhang, H.; Nie, F.; Jiang, H.; Bian, Y.; and Yan, J. 2024. Simplifying and empowering transformers for large-graph representations. In Proceedings of Advances in Neural Information Processing Systems

    2024

  29. [37]

    Xu, K.; Hu, W.; Leskovec, J.; and Jegelka, S. 2019. How powerful are graph neural networks? In Proceedings of International Conference on Learning Representations

    2019

  30. [38]

    Yang, Z.; Cohen, W.; and Salakhudinov, R. 2016. Revisiting semi-supervised learning with graph embeddings. In Proceedings of International Conference on Machine Learning, 40--48

    2016

  31. [39]

    Yin, Y.; Wang, Q.; Huang, S.; Xiong, H.; and Zhang, X. 2022. Autogcl: Automated graph contrastive learning via learnable view generators. In Proceedings of the AAAI conference on artificial intelligence, 8892--8900

    2022

  32. [40]

    You, Y.; Chen, T.; Sui, Y.; Chen, T.; Wang, Z.; and Shen, Y. 2020. Graph contrastive learning with augmentations. In Proceedings of Advances in Neural Information Processing Systems, 5812--5823

    2020

  33. [41]

    Yu, Y.; Wang, X.; Zhang, M.; Liu, N.; and Shi, C. 2024. Provable training for graph contrastive learning. In Proceedings of Advances in Neural Information Processing Systems

    2024

  34. [42]

    Yun, S.; Jeong, M.; Kim, R.; Kang, J.; and Kim, H. J. 2019. Graph transformer networks. In Proceedings of Advances in Neural Information Processing Systems

    2019

  35. [43]

    Zhang, Y.; Zhu, H.; Meng, Z.; Koniusz, P.; and King, I. 2022 a . Graph-adaptive rectified linear unit for graph neural networks. In Proceedings of the ACM Web Conference 2022, 1331--1339

    2022

  36. [44]

    Zhang, Y.; Zhu, H.; Song, Z.; Koniusz, P.; and King, I. 2022 b . COSTA: covariance-preserving feature augmentation for graph contrastive learning. In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2524--2534

    2022

  37. [45]

    Zhang, Y.; Zhu, H.; Song, Z.; Koniusz, P.; and King, I. 2023 a . Spectral feature augmentation for graph contrastive learning and beyond. In Proceedings of the AAAI Conference on Artificial Intelligence, 11289--11297

    2023

  38. [46]

    Zhang, Y.; Zhu, H.; Song, Z.; Koniusz, P.; and King, I. 2023 b . Spectral feature augmentation for graph contrastive learning and beyond. In Proceedings of the AAAI Conference on Artificial Intelligence, 11289--11297

    2023

  39. [47]

    Zhu, H.; Sun, K.; and Koniusz, P. 2021. Contrastive laplacian eigenmaps. In Proceedings of Advances in neural information processing systems, 5682--5695

    2021

  40. [48]

    Zhu, Y.; Xu, Y.; Yu, F.; Liu, Q.; Wu, S.; and Wang, L. 2020. Deep graph contrastive representation learning. arXiv preprint arXiv:2006.04131

    2020 arXiv

  41. [49]

    Zhu, Y.; Xu, Y.; Yu, F.; Liu, Q.; Wu, S.; and Wang, L. 2021. Graph contrastive learning with adaptive augmentation. In Proceedings of the Web Conference 2021, 2069--2080

    2021

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.