Pith. sign in

REVIEW 4 major objections 8 minor 59 references

Edge Contrastive Learning: An Augmentation-Free Graph Contrastive Learning Model

T0 review · 4 major / 8 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 The paper claims that graph contrastive learning can be done at the level of edges, without augmentation, and that this edge-level contrast achieves state-of-the-art link prediction and extremely low-label node classification.

desk verdict Novel edge-level contrastive idea with an uncomputable loss equation as written; the strong empirical results cannot yet be attributed to the stated method. read the letter →

arxiv 2412.11075 v1 pith:5ERZ5W6N submitted 2024-12-15 cs.LG

Yujun Li , Hongyuan Zhang , Yuan Yuan This is my paper · ORCID
classification cs.LG
keywords graphcontrastivelearningself-supervisededge-levelcontrastaugmentation-freelinkpredictionsemi-supervisednodeclassificationneuralnetworkslow-label
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper sets out to show that graph contrastive learning can be done without data augmentation and without moving to node-level or graph-level contrasts: the learning signal can live entirely on edges. The proposed model, AFECL, constructs an edge embedding by concatenating the embeddings of the edge's two endpoint nodes, then trains with a contrastive loss that treats edges sharing a node as positive pairs and all other edges as negative pairs. The claimed payoff is practical: a single-view, augmentation-free objective that produces strong representations for link prediction and for semi-supervised node classification when only one to four labeled nodes per class are available, and that uses less memory than comparable augmentation-free methods. A sympathetic reader would take the contribution to be the demonstration that edge-edge contrast, not node or view contrast, is a sufficient self-supervised objective for graphs.

What carries the argument

The central object is the edge embedding $h_{ij} = \varphi(h_i, h_j)$, obtained by concatenating the endpoint node embeddings $h_i$ and $h_j$ through an identity or learnable map $W$, together with the edge contrastive loss $\ell(h_{ij})$ in Eq. (6). This loss is an InfoNCE-style ratio: the numerator contains the anchor edge and all edges incident to either endpoint (positives), and the denominator contains those positives plus non-incident edges (negatives), so topology enters the objective directly through the choice of pairs. Edge sampling via $A' = A \circ R$, with $R_{ij} \sim \mathrm{Bernoulli}(p_s)$ on existing edges, is the mechanism that keeps the $O(N^2 K F')$ overall complexity manageable on large graphs.

What would settle it

Inspect the released code to check whether the denominator of Eq. (6) actually sums over all node pairs or only over sampled observed edges; if it is the latter, the implemented loss differs from the written one and the reported runtime and accuracy claims would need to be re-derived under the true loss. A separate check is to replace the shared-endpoint positive rule with random edge pairing on a heterophilic dataset and measure whether accuracy drops, which would reveal whether the topological contrast is the source of the reported gain.

Watch

Extended reading notes

Core claim

On its own terms, the paper's discovery is that edges can be the contrastive instances in graph contrastive learning, and that the original unmodified graph is enough to define them: no corruptions, augmentations, or second view are required. The model defines $h_{ij} = \varphi(h_i, h_j)$, the embedding of edge $(v_i, v_j)$ as the (possibly learned) concatenation of its endpoint node embeddings, and optimizes a loss in which an anchor edge $h_{ij}$ is pulled toward itself and toward edges incident to either of its endpoints, while being pushed away from non-incident edges. Across eight datasets, the paper reports that this objective reaches state-of-the-art or competitive accuracy relative to twelve baselines, including two supervised GNNs, in low-label node classification, achieves the best AUC on link prediction for Cora, Citeseer, and Coauthor-CS, and scales to large graphs by sampling edges rather than constructing edge neighborhoods.

Load-bearing premise

The load-bearing premise is that the original graph's topology alone defines a useful notion of contrastive similarity—edges that share an endpoint should be pulled together and every other edge pair pushed apart—so that the objective remains beneficial across both homophilic and heterophilic graphs without any augmentation.

Editorial extensions

If this is right

  • No augmentation tuning: because AFECL uses the original graph as its only view, users do not need to select dataset-specific graph transformations such as edge dropping or attribute masking.
  • Low-label gains: with one to four labeled nodes per class, AFECL reports the best or second-best accuracy on Cora, Citeseer, PubMed, Coauthor-CS, and Amazon-Photo among the compared methods.
  • Direct link prediction strength: since edge embeddings are endpoint concatenations, they can be scored for missing links; AFECL reports higher AUC than the compared methods on Cora, Citeseer, and Coauthor-CS.
  • Comparable cost to node-node GCL: the paper's complexity analysis puts AFECL at $O((NFF' + N^2F')K)$, the same order as representative node-node methods, and its measured memory use is lower than SPGCL and GraphACL on citation graphs.
  • Scalability to large graphs: edge sampling plus the single-view design lets AFECL run on Penn94 and ogbn-arxiv, where it matches or exceeds the reported baselines.

Reading between the lines

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

  • Extension: the paper's positive-pair rule is motivated by topology alone, but the paper does not analyze when shared-node edges are actually semantically similar; on heterophilic graphs the reported gains suggest the objective may be capturing structural or positional information beyond label homophily, which would be worth isolating with synthetic graphs.
  • Extension: a direct testable consequence is that the denominator's treatment of non-neighbor edges determines both memory and behavior; checking the released implementation against Eq. (6) would settle whether the written all-pairs loss is what was evaluated.
  • Extension: the same edge-contrastive loss could be probed as a pretraining objective for heterogeneous or temporal graphs, where edge types carry meaning; the paper only tests the fixed homogeneous-graph setting.
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 / 8 minor

Summary. This paper proposes AFECL, an augmentation-free edge-level graph contrastive learning method. Node embeddings are learned by a multi-head GAT; an edge embedding is defined by concatenating the two endpoint node embeddings and applying a linear map or identity. The pretext task is an edge contrastive loss (Eq. 6) in which the anchor edge is contrasted with edges sharing one of its endpoints, with the paper claiming that edges sharing a node are positives and other edges are negatives. The authors report state-of-the-art results on semi-supervised node classification with 1-20 labels per class across eight datasets, link prediction on three datasets, and scalability experiments on Penn94 and ogbn-arxiv.

Significance. An augmentation-free, edge-level contrastive objective is a potentially valuable alternative to node-level GCL: it avoids handcrafted graph augmentations, exploits the comparatively rich edge set, and the concatenation-based edge embedding is computationally cheap. The paper provides code, extensive comparisons with 12 baselines, ablations, memory measurements, and hyperparameter sensitivity analyses. However, the central loss equation is under-specified: as written it references undefined embeddings and triple-counts the anchor, and the claimed global 'other edges' contrast is not what the equation implements. The empirical claims therefore currently support a method that is not fully described, and the contribution's validity depends on how these specification issues are resolved.

major comments (4)
  1. [Section 3.3, Eq. (6)] The denominator of Eq. (6) contains sums over all k not equal to i or j of exp(theta(hij,hik)/tau) and exp(theta(hij,hkj)/tau). Section 3.2 defines edge embeddings only for actual edges: E ∈ R^{M×D} and hij = phi(hi,hj) for (vi,vj) ∈ E. For k not in Ni, the pair (vi,vk) is not an edge, so hik is undefined. The paper therefore needs to state explicitly how the denominator is computed, for example by restricting the sums to existing edges or by defining node-pair embeddings for all pairs. If the implementation differs from the printed loss, the reported results must be re-evaluated under the implemented loss.
  2. [Section 3.3, Eq. (6) and positive-pair count] Because j ∈ Ni and i ∈ Nj, the self-similarity term exp(theta(hij,hij)/tau) appears three times in the numerator, while the denominator excludes the anchor from both k-sums. The stated positive count |Ni|+|Nj|+1 therefore counts the anchor three times; the distinct number of positive pairs is |Ni|+|Nj|-1. The normalization in Eq. (6) is internally inconsistent and needs to be corrected, or the paper should justify the triple counting.
  3. [Abstract, Section 1, and Section 3.3] The paper repeatedly states that edges that do not share a node are negative pairs. However, the denominator of Eq. (6) only contains terms hik and hkj, meaning the contrast is limited to pairs of edges that share endpoint i or endpoint j; edges disjoint from {vi,vj} never enter the loss. If the intended contrast is global edge-edge, Eq. (6) must include all other edges; if the intended contrast is local, the abstract, introduction, and method description should be revised accordingly. This distinction affects the interpretation of the empirical results.
  4. [Section 3.3, Computational Complexity, and Table 8] The complexity analysis states D'=KF' and derives O(N^2 KF') overall, but Eq. (5) defines W ∈ R^{D'×2KF'} and Table 8 reports values consistent with D'=2KF' (e.g., Cora with K=4 and F'=32 gives D'=256). The complexity claim therefore does not follow from the stated dimensions. In addition, Eq. (7) sums over all N^2 node pairs while the edge sampling procedure changes the number of nonzero terms to M', and the complexity analysis separately uses M' for large graphs; the relationship among Eq. (7), Algorithm 1, and the reported complexity needs to be reconciled.
minor comments (8)
  1. [Section 3.1] The phrase 'optimizing the the edge contrastive loss' contains a duplicated article and should be corrected.
  2. [Section 4.4] The text 'edge–evel contrastive method' should read 'edge-level contrastive method'.
  3. [Appendix A.1] The dataset name 'Chamelon' should be 'Chameleon', and 'Wikipeida' should be 'Wikipedia'.
  4. [Section 2.2] The sentence 'the original graph faded into the encoder' should likely read 'the original graph fed into the encoder'.
  5. [Section 4.4] The sentence 'The results of homophilic and heterophilic graphs node classification accuracy are summarized in Table 2 and From the tables' is grammatically incomplete and should be rephrased.
  6. [Section 4.5] The description of the 'w/o ECL' variant is unclear; the sentence 'define edges corresponding to nodes connected by only one node as negative pairs' should be replaced with a precise statement of which pairs are positive and which are negative.
  7. [Section 4.1 and Appendix A.1] The dataset descriptions state that Amazon-Photo nodes are 'documents' and Coauthor-CS nodes are 'documents'; these should be 'products' and 'authors', respectively.
  8. [Section 1] The claim 'for the first time we study the edge-level pairs for contrast' should be tempered or supported with an explicit discussion of prior edge-level contrastive methods, since the phrase 'first' requires careful literature verification.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's claims are empirical, and the loss design, while possibly underspecified in Eq. (6), is not equivalent to its inputs.

full rationale

The paper's central claims are experimental: AFECL is proposed as a contrastive objective, and its reported value is state-of-the-art performance on link prediction and semi-supervised node classification. There is no derived prediction that reduces to a fitted parameter or to a self-citation. The positive-pair rule (edges sharing a node are similar) is encoded directly into the loss, but this is a modeling choice, not a circular derivation; the paper does not claim to derive the rule from first principles. Self-citations in the references are background material and do not supply the key result. The most serious issue found is that Eq. (6) sums over non-neighbor pairs whose edge embeddings are not defined by Eqs. (4)-(5), and the written denominator is not the global 'other edges' contrast described in the abstract. However, this is an underspecification or implementation defect, not a case of a claim being equivalent to its inputs by construction. No fitted parameter is renamed as a prediction, and no uniqueness theorem or load-bearing self-citation is invoked. Therefore the paper is not circular, even though its stated loss may not be exactly what was evaluated.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The paper does not introduce new physical entities. Its contribution is a new contrastive objective and edge embedding scheme; the central claim depends on five per-dataset hyperparameters (especially the edge sampling probability) and on several unstated inductive biases about edge semantics and GAT encoding.

free parameters (5)
  • Temperature tau = Cora:1, Citeseer:5, PubMed:5, Coauthor-CS:1, Amazon-Photo:1, Actor:1, Chameleon:1, Penn94:0.2
    Tuned per dataset over {0.1,0.2,0.5,1,5,10,15}; directly controls the sharpness of the contrastive distribution.
  • Number of GAT attention heads K = Cora:4, Citeseer:4, PubMed:2, Coauthor-CS:4, Amazon-Photo:2, Actor:32, Chameleon:8, Penn94:32
    Tuned per dataset over {1,2,4,8,16,32}; increases model capacity and edge embedding dimension.
  • Node embedding dimension F' = Cora:32, Citeseer:32, PubMed:32, Coauthor-CS:32, Amazon-Photo:32, Actor:8, Chameleon:32, Penn94:256
    Tuned per dataset over {8,16,32,64,128,256}; determines edge embedding dimension D'=2KF'.
  • Edge sampling probability ps = 1,1,0.5,0.27,0.18,1,1,0.004
    Chosen so the number of sampled edges is around 40,000 on large graphs; recommendation derived from experiments on Coauthor-CS and Amazon-Photo and then applied per dataset.
  • Learning rate and weight decay = lr in {1e-2,5e-2,5e-3,1e-3}, wd in {1e-4,5e-5,5e-4,0}
    Per-dataset optimization hyperparameters, tuned by validation performance.
assumptions (5)
  • domain assumption The graph is simple and undirected with symmetric adjacency matrix, and edge embeddings are only defined for existing edges.
    Stated in Section 3.1; real datasets may be directed or have multi-edges, but the method and experiments preprocess them as simple undirected graphs.
  • domain assumption Edges sharing an endpoint are semantically related and are treated as positive pairs.
    Core positive-pair definition in Section 3.3; no theoretical argument is given for why this inductive bias should help on heterophilic graphs such as Actor and Chameleon.
  • ad hoc to paper Concatenating the two endpoint node embeddings produces an edge representation suitable for contrastive learning.
    Eq. (4)-(5) in Section 3.2; this design choice is specific to AFECL and has no independent empirical or theoretical support in the paper.
  • standard math Cosine similarity in the temperature-scaled InfoNCE loss is an appropriate measure of edge embedding agreement.
    The loss in Eq. (6) follows standard contrastive learning practice from Oord et al. 2018 and Sohn 2016.
  • domain assumption Multi-head GAT is an effective encoder for all eight datasets.
    Section 3.2 fixes multi-head GAT for all benchmarks; no comparison with other GNN encoders is provided to isolate the contribution of the edge contrastive loss.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Edge Contrastive Learning: An Augmentation-Free Graph Contrastive Learning Model." pith.science (2026). https://pith.science/paper/5ERZ5W6N

@misc{pith2026241211075,
  author       = {Pith},
  title        = {Pith review of: Edge Contrastive Learning: An Augmentation-Free Graph Contrastive Learning Model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5ERZ5W6N}},
  note         = {Machine review of arXiv:2412.11075}
}
read the original abstract

Graph contrastive learning (GCL) aims to learn representations from unlabeled graph data in a self-supervised manner and has developed rapidly in recent years. However, edgelevel contrasts are not well explored by most existing GCL methods. Most studies in GCL only regard edges as auxiliary information while updating node features. One of the primary obstacles of edge-based GCL is the heavy computation burden. To tackle this issue, we propose a model that can efficiently learn edge features for GCL, namely AugmentationFree Edge Contrastive Learning (AFECL) to achieve edgeedge contrast. AFECL depends on no augmentation consisting of two parts. Firstly, we design a novel edge feature generation method, where edge features are computed by embedding concatenation of their connected nodes. Secondly, an edge contrastive learning scheme is developed, where edges connecting the same nodes are defined as positive pairs, and other edges are defined as negative pairs. Experimental results show that compared with recent state-of-the-art GCL methods or even some supervised GNNs, AFECL achieves SOTA performance on link prediction and semi-supervised node classification of extremely scarce labels. The source code is available at https://github.com/YujunLi361/AFECL.

Figures

Figures reproduced from arXiv: 2412.11075 by the authors.

Figure 1
Figure 1. The architecture of the proposed AFECL framework. Firstly, the original graph is regarded as a view and fed into the [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 3
Figure 3. t-SNE visualization of representations on bench [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 2
Figure 2. Sensitivity analysis of the hyperparameters [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Experimental results for edge sampling. Orange [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 29 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]

    I.; Baratin, A.; Rajeshwar, S.; Ozair, S.; Bengio, Y.; Courville, A.; and Hjelm, D

    Belghazi, M. I.; Baratin, A.; Rajeshwar, S.; Ozair, S.; Bengio, Y.; Courville, A.; and Hjelm, D. 2018. Mutual information neural estimation. In International conference on machine learning, 531--540. PMLR

    2018

  4. [4]

    Cai, Z.; Jiang, Z.; and Yuan, Y. 2021. Task-related self-supervised learning for remote sensing image change detection. In ICASSP 2021-2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 1535--1539. IEEE

    work page 2021

  5. [5]

    Cao, X.; Shi, Y.; Wang, J.; Yu, H.; Wang, X.; and Yan, Z. 2022. Cross-modal knowledge graph contrastive learning for machine learning method recommendation. In Proceedings of the 30th ACM International Conference on Multimedia, 3694--3702

    work page 2022

  6. [6]

    Chen, T.; Kornblith, S.; Norouzi, M.; and Hinton, G. 2020. A simple framework for contrastive learning of visual representations. In International conference on machine learning, 1597--1607. PMLR

    2020

  7. [7]

    D.; and Varadhan, S

    Donsker, M. D.; and Varadhan, S. S. 1983. Asymptotic evaluation of certain Markov process expectations for large time. IV. Communications on pure and applied mathematics, 36(2): 183--212

    1983

  8. [8]

    Feng, S.; Jing, B.; Zhu, Y.; and Tong, H. 2022. Adversarial graph contrastive learning with information regularization. In Proceedings of the ACM Web Conference 2022, 1362--1371

    work page 2022

Show all 59 references
  1. [9]

    Feng, W.; Zhang, J.; Dong, Y.; Han, Y.; Luan, H.; Xu, Q.; Yang, Q.; Kharlamov, E.; and Tang, J. 2020. Graph random neural networks for semi-supervised learning on graphs. Advances in neural information processing systems, 33: 22092--22103

    2020

  2. [10]

    Gao, T.; Yao, X.; and Chen, D. 2021. Simcse: Simple contrastive learning of sentence embeddings. arXiv preprint arXiv:2104.08821

    2021 arXiv

  3. [11]

    Gong, X.; Yang, C.; and Shi, C. 2023. Ma-gcl: Model augmentation tricks for graph contrastive learning. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, 4284--4292

    2023

  4. [12]

    Gutmann, M.; and Hyv \"a rinen, A. 2010. Noise-contrastive estimation: A new estimation principle for unnormalized statistical models. In Proceedings of the thirteenth international conference on artificial intelligence and statistics, 297--304. JMLR Workshop and Conference Pr...

    2010

  5. [13]

    Hassani, K.; and Khasahmadi, A. H. 2020. Contrastive multi-view representation learning on graphs. In International conference on machine learning, 4116--4126. PMLR

    2020

  6. [14]

    He, D.; Zhao, J.; Huo, C.; Huang, Y.; Huang, Y.; and Feng, Z. 2024. A New Mechanism for Eliminating Implicit Conflict in Graph Contrastive Learning. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, 12340--12348

    2024

  7. [15]

    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

  8. [16]

    Li, Q.; Han, Z.; and Wu, X.-M. 2018. Deeper insights into graph convolutional networks for semi-supervised learning. In Proceedings of the AAAI conference on artificial intelligence, volume 32

    2018

  9. [17]

    Li, Q.; Wu, X.-M.; Liu, H.; Zhang, X.; and Guan, Z. 2019. Label efficient semi-supervised learning via graph filtering. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 9582--9591

    2019

  10. [18]

    Li, X. 2022. Positive-incentive noise. IEEE Transactions on Neural Networks and Learning Systems

    2022

  11. [19]

    Li, Y.; Chen, C.; Zheng, X.; Zhang, Y.; Han, Z.; Meng, D.; and Wang, J. 2023. Making users indistinguishable: Attribute-wise unlearning in recommender systems. In Proceedings of the 31st ACM International Conference on Multimedia, 984--994

    2023

  12. [20]

    Li, Z.; Zhao, B.; and Yuan, Y. 2023. Bio-Inspired Audiovisual Multi-Representation Integration via Self-Supervised Learning. In Proceedings of the 31st ACM International Conference on Multimedia, 3755--3764

    2023

  13. [21]

    L.; Gupta, V.; Bhalerao, O.; and Lim, S

    Lim, D.; Hohne, F.; Li, X.; Huang, S. L.; Gupta, V.; Bhalerao, O.; and Lim, S. N. 2021. Large scale learning on non-homophilous graphs: New benchmarks and strong simple methods. Advances in Neural Information Processing Systems, 34: 20887--20902

    2021

  14. [22]

    Lin, L.; and Chen, J. 2023. Spectral Augmentation for Self-Supervised Learning on Graphs. In The Eleventh International Conference on Learning Representations

    2023

  15. [23]

    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

  16. [24]

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

    2018 arXiv

  17. [25]

    Paszke, A.; Gross, S.; Massa, F.; Lerer, A.; Bradbury, J.; Chanan, G.; Killeen, T.; Lin, Z.; Gimelshein, N.; Antiga, L.; et al. 2019. Pytorch: An imperative style, high-performance deep learning library. Advances in neural information processing systems, 32

    2019

  18. [26]

    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

  19. [27]

    Rao, J.; Zheng, S.; Mai, S.; and Yang, Y. 2022. Communicative Subgraph Representation Learning for Multi-Relational Inductive Drug-Gene Interaction Prediction. arXiv preprint arXiv:2205.05957

    2022 arXiv

  20. [28]

    Sen, P.; Namata, G.; Bilgic, M.; Getoor, L.; Galligher, B.; and Eliassi-Rad, T. 2008. Collective classification in network data. AI magazine, 29(3): 93--93

    2008

  21. [29]

    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

  22. [30]

    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, volume 37, 9782--9791

    2023

  23. [31]

    Shiao, W.; Guo, Z.; Zhao, T.; Papalexakis, E.; Liu, Y.; and Shah, N. 2023. Link Prediction with Non-Contrastive Learning. In Eleventh International Conference on Learning Representations (ICLR)

    2023

  24. [32]

    Sohn, K. 2016. Improved deep metric learning with multi-class n-pair loss objective. Advances in neural information processing systems, 29

    2016

  25. [33]

    Sun, M.; Xing, J.; Wang, H.; Chen, B.; and Zhou, J. 2021. MoCL: data-driven molecular fingerprint via knowledge-aware contrastive learning from molecular graph. In Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining, 3585--3594

    2021

  26. [34]

    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. 2021. Large-scale representation learning on graphs via bootstrapping. arXiv preprint arXiv:2102.06514

    2021 arXiv

  27. [35]

    Van der Maaten, L.; and Hinton, G. 2008. Visualizing data using t-SNE. Journal of machine learning research, 9(11)

    2008

  28. [36]

    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

  29. [37]

    L.; Liò, P.; Bengio, Y.; and Hjelm, R

    Veličković, P.; Fedus, W.; Hamilton, W. L.; Liò, P.; Bengio, Y.; and Hjelm, R. D. 2019. Deep Graph Infomax. In International Conference on Learning Representations

    2019

  30. [38]

    Wan, S.; Zhan, Y.; Liu, L.; Yu, B.; Pan, S.; and Gong, C. 2021. Contrastive graph poisson networks: Semi-supervised learning with extremely limited labels. Advances in Neural Information Processing Systems, 34: 6316--6327

    2021

  31. [39]

    Wang, H.; Zhang, J.; Zhu, Q.; Huang, W.; Kawaguchi, K.; and Xiao, X. 2022. Single-pass contrastive learning can work for both homophilic and heterophilic graph. arXiv preprint arXiv:2211.10890

    2022 arXiv

  32. [40]

    Wang, M.; Zheng, D.; Ye, Z.; Gan, Q.; Li, M.; Song, X.; Zhou, J.; Ma, C.; Yu, L.; Gai, Y.; et al. 2019. Deep graph library: A graph-centric, highly-performant package for graph neural networks. arXiv preprint arXiv:1909.01315

    2019 arXiv

  33. [41]

    Xia, J.; Wu, L.; Chen, J.; Hu, B.; and Li, S. Z. 2022. Simgrace: A simple framework for graph contrastive learning without data augmentation. In Proceedings of the ACM Web Conference 2022, 1070--1079

    2022

  34. [42]

    Xiao, T.; Zhu, H.; Chen, Z.; and Wang, S. 2024. Simple and asymmetric graph contrastive learning without augmentations. Advances in Neural Information Processing Systems, 36

    2024

  35. [43]

    Xie, Y.; Xu, Z.; Zhang, J.; Wang, Z.; and Ji, S. 2022. Self-supervised learning of graph neural networks: A unified review. IEEE transactions on pattern analysis and machine intelligence, 45(2): 2412--2429

    2022

  36. [44]

    Yang, H.; Gao, J.; Yuan, Y.; and Li, X. 2023. Imbalanced Aircraft Data Anomaly Detection. arXiv preprint arXiv:2305.10082

    2023 arXiv

  37. [45]

    Yang, W.; and Mirzasoleiman, B. 2024. Graph Contrastive Learning under Heterophily via Graph Filters. arXiv:2303.06344

    2024 arXiv

  38. [46]

    Yang, Z.; Cohen, W.; and Salakhudinov, R. 2016. Revisiting semi-supervised learning with graph embeddings. In International conference on machine learning, 40--48. PMLR

    2016

  39. [47]

    You, Y.; Chen, T.; Shen, Y.; and Wang, Z. 2021. Graph contrastive learning automated. In International Conference on Machine Learning, 12121--12132. PMLR

    2021

  40. [48]

    You, Y.; Chen, T.; Sui, Y.; Chen, T.; Wang, Z.; and Shen, Y. 2020. Graph contrastive learning with augmentations. Advances in neural information processing systems, 33: 5812--5823

    2020

  41. [49]

    Zhang, G.; Li, Z.; Huang, J.; Wu, J.; Zhou, C.; Yang, J.; and Gao, J. 2022 a . efraudcom: An e-commerce fraud detection system via competitive graph neural networks. ACM Transactions on Information Systems (TOIS), 40(3): 1--29

    2022

  42. [50]

    Zhang, H.; Shi, J.; Zhang, R.; and Li, X. 2023. Non-Graph Data Clustering via-Bipartite Graph Convolution. IEEE Transactions on Pattern Analysis & Machine Intelligence, 45(07): 8729--8742

    2023

  43. [51]

    Zhang, H.; Wu, Q.; Wang, Y.; Zhang, S.; Yan, J.; and Yu, P. S. 2022 b . Localized contrastive learning on graphs. arXiv preprint arXiv:2212.04604

    2022 arXiv

  44. [52]

    Zhang, H.; Wu, Q.; Yan, J.; Wipf, D.; and Yu, P. S. 2021. From canonical correlation analysis to self-supervised graph neural networks. Advances in Neural Information Processing Systems, 34: 76--89

    2021

  45. [53]

    Zhang, H.; Xu, Y.; Huang, S.; and Li, X. 2024 a . Data Augmentation of Contrastive Learning is Estimating Positive-incentive Noise. arXiv preprint arXiv:2408.09929

    2024 arXiv

  46. [54]

    Zhang, H.; Zhu, Y.; and Li, X. 2024. Decouple Graph Neural Networks: Train Multiple Simple GNNs Simultaneously Instead of One. IEEE Transactions on Pattern Analysis and Machine Intelligence

    2024

  47. [55]

    Zhang, S.; Yang, W.; Cao, X.; Zhang, H.; and Huang, Z. 2024 b . StructComp: Substituting Propagation with Structural Compression in Training Graph Contrastive Learning. arXiv:2312.04865

    2024 arXiv

  48. [56]

    Zhu, H.; Sun, K.; and Koniusz, P. 2021. Contrastive laplacian eigenmaps. Advances in Neural Information Processing Systems, 34: 5682--5695

    2021

  49. [57]

    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

  50. [58]

    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

  51. [59]

    Zhu, Z.; Galkin, M.; Zhang, Z.; and Tang, J. 2022. Neural-symbolic models for logical queries on knowledge graphs. In International Conference on Machine Learning, 27454--27478. PMLR

    2022

Pith tools

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