Pith. sign in

REVIEW 3 major objections 5 minor 84 references

A Generative Graph Contrastive Learning Model with Global Signal

T0 review · 3 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 A graph contrastive learning framework called CSG2L, built on an SVD low-rank augmented view and pseudo-label adaptive reweighting, claims consistent node-classification gains over base GNNs on six benchmark graphs.

desk verdict The empirical story is consistent, but the paper misdescribes its own loss: Eq. (7) treats every non-i pair as negative, so the claimed hard-positive reweighting never happens. read the letter →

arxiv 2504.18148 v1 pith:CH6O3ZWQ submitted 2025-04-25 cs.LG

Xiaofan Wei , Binyan Zhang This is my paper · ORCID
classification cs.LG
keywords graphcontrastivelearningnodeclassificationsingularvaluedecompositionlow-rankapproximationadaptivereweightinghardsamplepairsheterophilicgraphsInfoNCEloss
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 tries to establish that graph contrastive learning can be made more accurate by replacing random graph perturbations with a singular value decomposition (SVD) low-rank view and by weighting hard sample pairs more than easy ones. The proposed framework, CSG2L, wraps a base GNN with a contrastive loss on two views: the original normalized adjacency matrix and its low-rank approximation. A pseudo-label-based reweighting term adjusts InfoNCE so that high-confidence same-class pairs with low similarity and high-confidence different-class pairs with high similarity contribute more. The reported experiments claim consistent improvements for GCN, GIN, and GPRGNN across six real graphs, including four heterophilic datasets, with the GPRGNN variant best on all six. A sympathetic reader would care because, if true, this is a general performance booster for GNN node classification that does not change the underlying GNN architecture.

What carries the argument

The load-bearing object is the low-rank SVD reconstruction of the graph, $\hat{A}_{\text{SVD}} = \hat{U}_q\hat{S}_q\hat{V}_q^T$, with $q = 5$ in the experiments, computed by the randomized ApproxSVD algorithm. It serves as a denoised, globally informed second view of the graph, avoiding the noise introduced by random edge perturbation or attribute masking. The second mechanism is the pseudo-label adaptive reweighting inside InfoNCE: an MLP classifier maps the original-view embeddings to predictions, high-confidence nodes define the set $\mathcal{O}$ and binary pair matrix $Q$, and the weight $R(h_i,z_k)=|Q_{ik}-\mathrm{Norm}(s(h_i,z_k))|$ replaces the uniform weight in Eq. (7). Together these produce the augmented contrastive signal and decide which sample pairs should be pulled together or pushed apart.

What would settle it

Re-run CSG2L on the same six datasets with the pseudo-label pair matrix replaced by random labels while keeping the SVD view and the same loss; if accuracy does not drop meaningfully, the adaptive reweighting is not what carries the reported gains. A second check is to sweep the unspecified confidence threshold that defines the high-confidence set: the paper's explanation predicts accuracy peaks where the threshold tracks pseudo-label quality, so a flat accuracy curve across all thresholds would falsify the claimed mechanism.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that the two failure modes it identifies in existing graph contrastive learning—noisy augmented views and uniform treatment of sample pairs—can be addressed jointly. The SVD-aug module computes a rank-$q$ approximation $\hat{A}_{\text{SVD}}=\hat{U}_q\hat{S}_q\hat{V}_q^T$ of the normalized adjacency matrix and uses it as the augmented view, so the contrastive signal carries global structure instead of random noise. The LGDL module then takes both views through a shared GNN and defines a reweighting function $R(h_i,z_k)=|Q_{ik}-\mathrm{Norm}(s(h_i,z_k))|$ for high-confidence nodes, where $Q$ encodes pseudo-label agreement from the classifier being trained. This up-weights hard positive pairs (same pseudo-label, low similarity) and hard negative pairs (different pseudo-labels, high similarity). The experimental section reports that CSG2L improves the average accuracy of GCN by 3.04%, GIN by 3.89%, and GPRGNN by 2.25% over ten splits of six datasets, and that CSG2L_GPRGNN achieves the best result on every benchmark.

Load-bearing premise

The load-bearing premise is that the binary pseudo-label matrix $Q$, computed from the same classifier being trained, is accurate enough to decide which contrastive pairs are positive and which are negative; the paper never states the confidence threshold that defines high-confidence nodes, and if high-confidence predictions are systematically wrong, the adaptive reweighting amplifies those mistakes instead of correcting them.

Editorial extensions

If this is right

  • Any GNN that can consume a normalized adjacency matrix can be wrapped with CSG2L, making the framework a plug-in booster rather than a new architecture.
  • The reported gains are largest on heterophilic graphs (Texas +4.26, Cornell +3.05, Wisconsin +2.83 over GPRGNN), suggesting that a global SVD signal is most valuable where local neighborhoods are unreliable.
  • Because the SVD view replaces random perturbation, the same contrastive objective can be applied without the noise term that random augmentations introduce.
  • The reweighting function is differentiable in the embeddings through the similarity term, so the whole objective can be trained end-to-end with the classification loss.
  • The ablation study indicates each component contributes in order: random augmentation over the base model, then SVD-aug, then adaptive reweighting, with the full model best.

Reading between the lines

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

  • A natural extension the paper does not test: if pseudo-labels are the main driver, the benefit should shrink when the number of labeled nodes is very small or labels are noisy, since the classifier's own errors then define the contrastive pairs.
  • Comparing CSG2L with equal-weight InfoNCE on the same SVD view would isolate how much of the reported gain comes from the reweighting term rather than from the SVD augmentation alone.
  • The low-rank SVD view is essentially a denoised graph filter, which suggests the same recipe could transfer to link prediction or graph-level tasks, though the paper only evaluates node classification.
  • The confidence threshold that defines the high-confidence set $\mathcal{O}$ is never stated, so a practitioner would need to tune it per dataset; if the optimal threshold varies, the plug-in claim depends on an unreported hyperparameter.
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

3 major / 5 minor

Summary. The paper proposes CSG2L, a graph contrastive learning framework for node classification that combines an SVD-directed augmented view (SVD-aug) with a local-global dependency learning module (LGDL) that adaptively reweights contrastive pairs. The authors report consistent accuracy improvements over three base GNNs (GCN, GIN, GPRGNN) on six benchmark datasets, with the full framework outperforming state-of-the-art baselines in Table II. The central claim is that the SVD-aug module captures global interactions while LGDL provides better contrastive signals by emphasizing hard sample pairs.

Significance. If the empirical results are reproducible, CSG2L is a simple, general plug-in that improves multiple GNN backbones on both homophilic and heterophilic graphs. The paper's strengths are the breadth of the evaluation (18 model-dataset combinations, all with positive gains), the clear ablation evidence that SVD-aug alone helps, and the falsifiable accuracy claim. However, the paper does not provide code, significance tests, or a mathematically consistent description of the LGDL mechanism, and the pseudo-label construction has a circularity risk. The significance of the contribution is therefore contingent on correcting these issues.

major comments (3)
  1. [Section III.C, Eq. (7)] The adaptive reweighting of hard positive pairs described in the text is not realized by the loss in Eq. (7). The positive term is exactly one: e^{R(hi,zi)θ(hi,zi)}. For every k≠i, R(hi,zk) and R(hi,hk) appear exclusively in the denominator, so a high-confidence pair (i,k) with Qik=1 is treated as a negative pair; minimizing the loss decreases its similarity, which is the opposite of pulling same-pseudo-label pairs together. Moreover, R∈[0,1], so hard pairs are never assigned a weight greater than the default 1, contradicting the claim that they receive 'greater importance.' At best the loss implements a hard-negative reweighting scheme. The ablation gain attributed to LGDL therefore cannot be attributed to the claimed positive-pair weighting unless the loss is corrected or the interpretation is revised. Note also that R is defined only for arguments (hi,zk), but Eq. (7) applies it to (hi,hk) as well.
  2. [Section III.C, Eqs. (5)-(6)] The pseudo-label matrix Q is derived from the predictions of the MLP classifier c(H) that is being trained jointly with the contrastive objective. For high-confidence nodes, the reweighting in Eq. (6) therefore reinforces the model's own predictions rather than providing an independent contrastive signal, and any systematic errors in c(H) are amplified. The paper never states the confidence threshold that defines the set O, nor does it validate the pseudo-labels against any external signal. The assumption that equal pseudo-labels indicate true positives is load-bearing: the claimed 'accurate contrastive signals' are only as good as the classifier's self-consistency. Please report the threshold, analyze its sensitivity, and discuss the risk of confirmation bias.
  3. [Section IV.D, Tables II and III] The paper claims 'significant improvement' but reports no significance tests, and several improvements are small relative to the reported standard deviations (e.g., Table III: Cora +0.49±0.62, Citeseer +0.52±1.26 for GCN). Since results are averaged over ten splits, paired statistical tests (e.g., paired t-test or Wilcoxon signed-rank test) across the same splits should be reported to substantiate the central claim of consistent improvement.
minor comments (5)
  1. [Section III.A] The heading 'Problem Defination' should be 'Problem Definition', and the condition aij=1 when (ni,nj)∈N should refer to the edge set E rather than the node set N.
  2. [Section III.C, Eq. (7)] Equation (7) uses R(hi,hk) although the definition of R in Eq. (6) only covers arguments of the form (hi,zk); the definition of R should be extended or the loss should be corrected to make the notation consistent.
  3. [Section IV.E] The hyper-parameter analysis sweeps only λ; the SVD rank q, the temperature τ, and the high-confidence threshold for O are fixed without sensitivity analysis, so the robustness of the reported gains to these choices is unclear.
  4. [Figures 2 and 3] The hyper-parameter and ablation results are shown only as figures without numeric tables; providing exact values in the text or in supplementary material would improve reproducibility.
  5. [References] The reference list contains many entries (e.g., [32]-[42], [44]-[54], [57]-[82]) that are not cited in the body or are only tangentially related; the bibliography should be pruned to the relevant literature.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported gains are benchmarked against held-out true labels and are not forced by the model's own equations.

full rationale

The derivation chain is empirical rather than definitional. SVD-aug computes a low-rank view of the normalized adjacency matrix (Eq. 1), and LGDL combines this view with the original graph through an InfoNCE-style contrastive loss (Eqs. 4-8) plus supervised cross-entropy (Eq. 9). The final node-classification accuracy in Tables II and III is measured against ground-truth test labels on standard splits, so the improvement claim does not reduce by construction to the training objective. The pseudo-label matrix Q (Eq. 5) is indeed computed from the same MLP classifier being trained, which is a self-training feedback loop rather than an independent signal; however, this does not make the benchmark accuracy a tautology, because the reported performance is still externally evaluated and the loss includes supervised cross-entropy. The principal correctness concern is that Eq. 7 places reweighted cross-sample terms in the denominator only, so the text's claim that same-pseudo-label hard positives are pulled closer is not implemented literally; that is an internal consistency or implementation issue, not a circularity in the sense of a prediction being equivalent to its inputs by construction.

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

The framework's gains rest on two nontrivial domain assumptions: the top-q singular vectors of the normalized adjacency matrix encode useful global structure (Section III.B), and high-confidence predictions of the classifier are valid pseudo-labels for contrastive pair definition (Section III.C). It also carries five hand-set or per-dataset free parameters, one of which (the confidence threshold) is never specified. No invented entities are introduced; the 'global signal' is a mathematical construction, the SVD reconstruction. The pseudo-label assumption is the most consequential because it embeds a circular component into the contrastive objective.

free parameters (5)
  • SVD rank q = 5
    Controls the truncation in SVD-aug (Eq. 1) and therefore the information content of the augmented view. Set to 5 for graphs of 183 to 3327 nodes, with no reported sensitivity analysis.
  • Contrastive loss weight lambda = 0.1
    Selected by hyperparameter analysis (Fig. 2) on Cora, Texas, and Chameleon, all of which also appear in the final accuracy tables; the tuning is not isolated on a held-out split.
  • Temperature tau = 0.5
    Hand-set in Section IV.C; controls the sharpness of the softmax in the InfoNCE losses (Eq. 4 and 7).
  • High-confidence threshold for set O = not specified
    Defines which nodes receive pseudo-labels in Section III.C ('exceed a given threshold') but the value is never stated, making it an uncontrolled free parameter.
  • Weight decay = chosen from {1e-3, 5e-3, 5e-4, 5e-5, 5e-6}
    Reported in Section IV.C as selected per dataset, but per-dataset values are not given.
assumptions (5)
  • domain assumption The top-q singular values and vectors of the normalized adjacency matrix capture the graph's essential global structure, while the discarded tail is noise.
    Underpins SVD-aug in Section III.B. For heterophilic graphs, high-frequency (small singular value) structure can be informative, so this premise is nontrivial and is not tested directly.
  • domain assumption High-confidence predictions of the MLP classifier c(H) are accurate enough to serve as pseudo-labels for defining contrastive pairs.
    Section III.C builds Q from pseudo-labels Y (Eq. 5). If high-confidence predictions are systematically wrong, reweighting amplifies errors instead of correcting them.
  • ad hoc to paper Pairs with equal pseudo-labels are true positives and pairs with different pseudo-labels are true negatives.
    Eq. (5) binarizes class agreement into Q, a hard modeling choice specific to this paper that ignores graded similarity.
  • standard math The randomized SVD algorithm (ApproxSVD, Halko et al. [21]) yields a sufficiently accurate low-rank approximation for the augmented view.
    Invoked in Eq. (1) without quality or error analysis in the context of adjacency matrix approximation.
  • domain assumption The ten 60/20/20 random splits with early stopping on validation provide an unbiased estimate of test accuracy.
    Section IV.A; standard benchmark practice, but no confidence intervals or significance tests are derived from the ten runs.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Generative Graph Contrastive Learning Model with Global Signal." pith.science (2026). https://pith.science/paper/CH6O3ZWQ

@misc{pith2026250418148,
  author       = {Pith},
  title        = {Pith review of: A Generative Graph Contrastive Learning Model with Global Signal},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CH6O3ZWQ}},
  note         = {Machine review of arXiv:2504.18148}
}
read the original abstract

Graph contrastive learning (GCL) has garnered significant attention recently since it learns complex structural information from graphs through self-supervised learning manner. However, prevalent GCL models may suffer from performance degradation due to inappropriate contrastive signals. Concretely, they commonly generate augmented views based on random perturbation, which leads to biased essential structures due to the introduction of noise. In addition, they assign equal weight to both hard and easy sample pairs, thereby ignoring the difference in importance of the sample pairs. To address these issues, this study proposes a novel Contrastive Signal Generative Framework for Accurate Graph Learning (CSG2L) with the following two-fold ideas: a) building a singular value decomposition (SVD)-directed augmented module (SVD-aug) to obtain the global interactions as well as avoiding the random noise perturbation; b) designing a local-global dependency learning module (LGDL) with an adaptive reweighting strategy which can differentiate the effects of hard and easy sample pairs. Extensive experiments on benchmark datasets demonstrate that the proposed CSG2L outperforms the state-of-art baselines. Moreover, CSG2L is compatible with a variety of GNNs.

Figures

Figures reproduced from arXiv: 2504.18148 by the authors.

Figure 1
Figure 1. a) The overall structure of CSG2L. It includes two components: b) SVD-aug module performs SVD-directed reconstruction of the original graph to obtain augmented graph. c) LGDL module inputs both graphs into a shared GNN, and further input the embedding of original graph into an MLP classifier to compute the classification loss. In addition, an adaptive reweighting strategy is introduced for contrastive learning. Fina… view at source ↗
Figure 2
Figure 2. Hyper-parameter analysis results of CSG2Ls on three benchmark datasets [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. Ablation study results on three benchmark datasets. [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

84 extracted references · 69 canonical work pages

  1. [1]

    Y . Wang, X. Hu, Q. Gan, X. Huang, X. Qiu, and D. Wipf, ”Efficient link prediction via gnn layers induced by negative sampling,” IEEE Transactions on Knowledge and Data Engineering, 2024

    2024

  2. [2]

    L. Wei, H. Zhao, Z. He, and Q. Yao, ”Neural architecture search for GNN-based graph classification,” ACM Transactions on Information Systems, vol. 42, no. 1, pp. 1–29, 2023

    2023

  3. [3]

    Ju et al., ”Towards Graph Contrastive Learning: A Survey and Beyond,” arXiv preprint arXiv:2405.11868, 2024

    W. Ju et al., ”Towards Graph Contrastive Learning: A Survey and Beyond,” arXiv preprint arXiv:2405.11868, 2024

    arXiv 2024

  4. [4]

    Y . Liu, K. Ding, J. Wang, V . Lee, H. Liu, and S. Pan, ”Learning strong graph neural networks with weak information,” in Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2023, pp. 1559–1571

    2023

  5. [5]

    Welling and T

    M. Welling and T. N. Kipf, ”Semi-supervised classification with graph convolutional networks,” in Proc. Int. Conf. on Learning Represenations, 2016

    2016

  6. [6]

    Hamilton, Z

    W. Hamilton, Z. Ying, and J. Leskovec, ”Inductive representation learning on large graphs,” Advances in neural information processing systems, vol. 30, 2017

    work page 2017

  7. [7]

    Veli ˇckovi´c, G

    P. Veli ˇckovi´c, G. Cucurull, A. Casanova, A. Romero, P. Lio, and Y . Bengio, ”Graph attention networks,” in Proceedings of the International Conference on Learning Representations, 2018

    work page 2018

  8. [8]

    Frasca, E

    F. Frasca, E. Rossi, D. Eynard, B. Chamberlain, M. Bronstein, and F. Monti, ”Sign: Scalable inception graph neural networks,” arXiv preprint arXiv:2004.11198, 2020

    arXiv 2004

Show all 84 references
  1. [9]

    Li et al., ”Finding global homophily in graph neural networks when meeting heterophily,” in International Conference on Machine Learning, PMLR, 2022, pp

    X. Li et al., ”Finding global homophily in graph neural networks when meeting heterophily,” in International Conference on Machine Learning, PMLR, 2022, pp. 13242–13256

    2022

  2. [10]

    Defferrard, X

    M. Defferrard, X. Bresson, and P. Vandergheynst, ”Convolutional neural networks on graphs with fast localized spectral filtering,” Advances in neural information processing systems, vol. 29, 2016

    2016

  3. [11]

    F. Wu, A. Souza, T. Zhang, C. Fifty, T. Yu, and K. Weinberger, ”Simplifying graph convolutional networks,” in International conference on machine learning, PMLR, 2019, pp. 6861–6871

    2019

  4. [12]

    Gasteiger, S

    J. Gasteiger, S. Weißenberger, and S. G ¨unnemann, ”Diffusion improves graph learning,” Advances in neural information processing systems, vol. 32, 2019

    2019

  5. [13]

    Predict then Propagate: Graph Neural Networks meet Personalized PageRank,

    J. Gasteiger, A. Bojchevski, and S. G ¨unnemann, “Predict then Propagate: Graph Neural Networks meet Personalized PageRank,” in ICLR, Feb. 2019. 14

    2019

  6. [14]

    Y . You, T. Chen, Y . Sui, T. Chen, Z. Wang, and Y . Shen, ”Graph contrastive learning with augmentations,” Advances in neural information processing systems, vol. 33, pp. 5812–5823, 2020

    2020

  7. [15]

    Y . Zhu, Y . Xu, F. Yu, Q. Liu, S. Wu, and L. Wang, ”Graph Contrastive Learning with Adaptive Augmentation,” in Proceedings of the Web Conference 2021, Apr. 2021, pp. 2069–2080. doi: 10.1145/3442381. 3449802

    2021 doi

  8. [16]

    Sun et al., ”Sugar: Subgraph neural network with reinforcement pooling and self-supervised mutual information mechanism,” in Proceedings of the web conference 2021, 2021, pp

    Q. Sun et al., ”Sugar: Subgraph neural network with reinforcement pooling and self-supervised mutual information mechanism,” in Proceedings of the web conference 2021, 2021, pp. 2081–2091

    2021

  9. [17]

    Thakoor, C

    S. Thakoor, C. Tallec, M. G. Azar, R. Munos, P. Veli ˇckovi´c, and M. Valko, ”Bootstrapped representation learning on graphs,” in ICLR 2021 Workshop on Geometrical and Topological Representation Learning, 2021

    2021

  10. [18]

    J. Xia, L. Wu, G. Wang, J. Chen, and S. Z. Li, ”ProGCL: Rethinking Hard Negative Mining in Graph Contrastive Learning,” in International Conference on Machine Learning, PMLR, 2022, pp. 24332–24346

    2022

  11. [19]

    G. Chu, X. Wang, C. Shi, and X. Jiang, ”CuCo: Graph Representation with Currculum Contrastive Learning,” in IJCAI, 2021, pp. 2300–2306

    2021

  12. [20]

    Rajwade, A

    A. Rajwade, A. Rangarajan, and A. Banerjee, ”Image denoising using the highe order singular value decomposition,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 35, no. 4, pp. 849–862, 2012

    2012

  13. [21]

    Halko, P.-G

    N. Halko, P.-G. Martinsson, and J. A. Tropp, ”Finding structure withrandomness: Probabilistic algorithms for constructing approximate matrix decompositions,” SIAM review, vol. 53, no. 2, pp. 217–288, 2011

    2011

  14. [22]

    H. Pei, B. Wei, K. C.-C. Chang, Y . Lei, and B. Yang, ”Geom-gcn: Geometric graph convolutional networks,” arXiv preprint arxiv: 2002.05287, 2020

    2002 arXiv

  15. [23]

    Rozemberczki, C

    B. Rozemberczki, C. Allen, and R. Sarkar, ”Multi-scale attributed node embedding,” Journal of Complex Networks, vol. 9, no. 2, p. cnab014, 2021

    2021

  16. [24]

    P. Sen, G. Namata, M. Bilgic, L. Getoor, B. Galligher, and T. Eliassi-Rad, ”Collective classification in network data,” AI magazine, vol. 29, no. 3, pp. 93–93, 2008

    2008

  17. [25]

    K. Xu, W. Hu, J. Leskovec, and S. Jegelka, ”How powerful are graph neural networks?,” arXiv preprint arXiv:1810.00826, 2018

    2018 arXiv

  18. [26]

    Chien, J

    E. Chien, J. Peng, P. Li, and O. Milenkovic, ”Adaptive universal generalized pagerank graph neural network,” arXiv preprint arxiv: 2006.07988, 2020

    2006 arXiv

  19. [27]

    Abu-El-Haija et al., ”Mixhop: Higher-order graph convolutional architectures via sparsified neighborhood mixing,” in international conference on machine learning, PMLR, 2019, pp

    S. Abu-El-Haija et al., ”Mixhop: Higher-order graph convolutional architectures via sparsified neighborhood mixing,” in international conference on machine learning, PMLR, 2019, pp. 21–29

    2019

  20. [28]

    J. Zhu, Y . Yan, L. Zhao, M. Heimann, L. Akoglu, and D. Koutra, ”Beyond homophily in graph neural networks: Current limitations and effective designs,” Advances in neural information processing systems, vol. 33, pp. 7793–7804, 2020

    2020

  21. [29]

    Zheng, H

    Y . Zheng, H. Zhang, V . Lee, Y . Zheng, X. Wang, and S. Pan, ”Finding the missing-half: Graph complementary learning for homophily-prone and heterophily-prone graphs,” in International Conference on Machine Learning, PMLR, 2023, pp. 42492–42505

    2023

  22. [30]

    Y . Song, C. Zhou, X. Wang, and Z. Lin, ”Ordered gnn: Ordering message passing to deal with heterophily and over-smoothing,” arXiv preprint arXiv:2302.01524, 2023

    2023 arXiv

  23. [31]

    J. Guo, K. Huang, R. Zhang, and X. Yi, ”ES-GNN: Generalizing graph neural networks beyond homophily with edge splitting,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024

    2024

  24. [33]

    L. Wang, K. Liu, and Y . Yuan, ”GT-A2T: Graph tensor alliance attention network”, IEEE/CAA Journal of Automatica Sinica, doi: 10.1109/JAS.2024.124863, 2024

    2024

  25. [34]

    J. Chen, Y . Yuan, and X. Luo, ”Sdgnn: Symmetry-preserving dual-stream graph neural networks”, IEEE/CAA journal of automatica sinica, vol. 11, no. 7, pp. 1717–1719, 2024

    2024

  26. [35]

    F. Bi, T. He, Y .-S. Ong, and X. Luo, ”Graph Linear Convolution Pooling for Learning in Incomplete High-Dimensional Data”, IEEE Transactions on Knowledge and Data Engineering, 2024

    2024

  27. [36]

    Y . Yuan, Y . Wang, and X. Luo, ”A Node-Collaboration-Informed Graph Convolutional Network for Highly Accurate Representation to Undirected Weighted Graph”, IEEE Transactions on Neural Networks and Learning Systems, 2024, DOI: 10.1109/TNNLS.2024.3514652, 2024

    2024

  28. [37]

    Y . Yuan, S. Lu, and X. Luo, ”A Proportional Integral Controller-Enhanced Non-negative Latent Factor Analysis Model”, IEEE/CAA Journal of Automatica Sinica, doi: 10.1109/JAS.2024.125055, 2024. 15

    2024

  29. [38]

    Y . Yuan, J. Li, and X. Luo, ”A fuzzy PID-incorporated stochastic gradient descent algorithm for fast and accurate latent factor analysis”, IEEE Transactions on Fuzzy Systems,vol. 32, no. 7, pp. 4049-4061, 2024

    2024

  30. [39]

    Y . Yuan, X. Luo, and M. Zhou, ”Adaptive divergence-based non-negative latent factor analysis of high-dimensional and incomplete matrices from industrial applications”, IEEE Transactions on Emerging Topics in Computational Intelligence, vol. 8, no. 2, pp. 1209–1222, 2024

    2024

  31. [40]

    Y . Yuan, X. Luo, M. Shang, and Z. Wang, ”A Kalman-filter-incorporated latent factor analysis model for temporally dynamic sparse data”, IEEE Transactions on Cybernetics, vol. 53, no. 9, pp. 5788–5801, 2022

    2022

  32. [41]

    Y . Yuan, R. Wang, G. Yuan, and L. Xin, ”An adaptive divergence-based non-negative latent factor model”, IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 53, no. 10, pp. 6475–6487, 2023

    2023

  33. [42]

    Y . Yuan, Q. He, X. Luo, and M. Shang, ”A multilayered-and-randomized latent factor model for high-dimensional and sparse matrices”, IEEE transactions on big data, vol. 8, no. 3, pp. 784–794, 2020

    2020

  34. [43]

    X. Luo, L. Wang, P. Hu, and L. Hu, ”Predicting protein-protein interactions using sequence and network information via variational graph autoencoder”, IEEE/ACM Transactions on Computational Biology and Bioinformatics, vol. 20, no. 5, pp. 3182–3194, 2023

    2023

  35. [44]

    J. Li, Y . Yuan, and X. Luo, ”Learning Error Refinement in Stochastic Gradient Descent-Based Latent Factor Analysis via Diversified PID Controllers”, IEEE Transactions on Emerging Topics in Computational Intelligence, DOI: 10.1109/TETCI.2025.3547854, 2025

  36. [45]

    X. Luo, J. Chen, Y . Yuan, and Z. Wang, ”Pseudo gradient-adjusted particle swarm optimization for accurate adaptive latent factor analysis”, IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 54, no. 4, pp. 2213–2226, 2024

    2024

  37. [46]

    J. Li, X. Luo, Y . Yuan, and S. Gao, ”A nonlinear PID-incorporated adaptive stochastic gradient descent algorithm for latent factor analysis”, IEEE Transactions on Automation Science and Engineering, 2023

    2023

  38. [47]

    Chen et al., ”A State-Migration Particle Swarm Optimizer for Adaptive Latent Factor Analysis of High-Dimensional and Incomplete Data”, IEEE/CAA Journal of Automatica Sinica, vol

    J. Chen et al., ”A State-Migration Particle Swarm Optimizer for Adaptive Latent Factor Analysis of High-Dimensional and Incomplete Data”, IEEE/CAA Journal of Automatica Sinica, vol. 11, no. 11, pp. 2220–2235, 2024

    2024

  39. [48]

    X. Luo, Y . Yuan, S. Chen, N. Zeng, and Z. Wang, ”Position-transitional particle swarm optimization-incorporated latent factor analysis”, IEEE Transactions on Knowledge and Data Engineering, vol. 34, no. 8, pp. 3958–3970, 2020

    2020

  40. [49]

    Shang, Y

    M. Shang, Y . Yuan, X. Luo, and M. Zhou, ”An α–β-divergence-generalized recommender for highly accurate predictions of missing user preferences”, IEEE transactions on cybernetics, vol. 52, no. 8, pp. 8006–8018, 2021

    2021

  41. [50]

    L. Xin, Y . Yuan, M. Zhou, Z. Liu, and M. Shang, ”Non-negative latent factor model based on β-divergence for recommender systems”, IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 51, no. 8, pp. 4612–4623, 2019

    2019

  42. [51]

    Y . Yuan, X. Luo, M. Shang, and D. Wu, ”A generalized and fast-converging non-negative latent factor model for predicting user preferences in recommender systems”, in Proceedings of The Web Conference 2020, 2020, pp. 498–507

    2020

  43. [52]

    Y . Yuan, M. Shang, and X. Luo, ”Temporal Web Service QoS Prediction via Kalman Filter-Incorporated Latent Factor Analysis”, in ECAI 2020, IOS Press, 2020, pp. 561–568

    2020

  44. [53]

    J. Chen, Y . Yuan, T. Ruan, J. Chen, and X. Luo, ”Hyper-parameter-evolutionary latent factor analysis for high-dimensional and sparse data from recommender systems”, Neurocomputing, vol. 421, pp. 316–328, 2021

    2021

  45. [54]

    J. Li, Y . Yuan, T. Ruan, J. Chen, and X. Luo, ”A proportional-integral-derivative-incorporated stochastic gradient descent-based latent factor analysis model”, Neurocomputing, vol. 427, pp. 29–39, 2021

    2021

  46. [55]

    Y . Yuan, X. Luo, and M.-S. Shang, ”Effects of preprocessing and training biases in latent factor models for recommender systems”, Neurocomputing, vol. 275, pp. 2019–2030, 2018

    2019

  47. [56]

    D. Wu, X. Luo, G. Wang, M. Shang, Y . Yuan, and H. Yan, ”A highly accurate framework for self-labeled semisupervised classification in industrial applications”, IEEE Transactions on Industrial Informatics, vol. 14, no. 3, pp. 909–920, 2017

    2017

  48. [57]

    L. Hu, S. Yang, X. Luo, H. Yuan, K. Sedraoui, and M. Zhou, ”A distributed framework for large-scale protein-protein interaction data analysis and prediction using mapreduce”, IEEE/CAA Journal of Automatica Sinica, vol. 9, no. 1, pp. 160–172, 2021

    2021

  49. [58]

    L. Hu, J. Zhang, X. Pan, X. Luo, and H. Yuan, ”An effective link-based clustering algorithm for detecting overlapping protein complexes in protein-protein interaction networks”, IEEE Transactions on Network Science and Engineering, vol. 8, no. 4, pp. 3275–3289, 2021

    2021

  50. [59]

    X. Luo, H. Wu, Z. Wang, J. Wang, and D. Meng, ”A novel approach to large-scale dynamically weighted directed network representation”, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 44, no. 12, pp. 9756–9773, 2021

    2021

  51. [60]

    F. Bi, T. He, and X. Luo, ”A fast nonnegative autoencoder-based approach to latent feature analysis on high-dimensional and incomplete data”, IEEE Transactions on Services Computing, vol. 17, no. 3, pp. 733–746, 2023

    2023

  52. [61]

    H. Wu, X. Luo, M. Zhou, M. J. Rawa, K. Sedraoui, and A. Albeshri, ”A PID-incorporated latent factorization of tensors approach to dynamically weighted directed network analysis”, IEEE/CAA Journal of Automatica Sinica, vol. 9, no. 3, pp. 533–546, 2021. 16

    2021

  53. [62]

    F. Bi, X. Luo, B. Shen, H. Dong, and Z. Wang, ”Proximal alternating-direction-method-of-multipliers-incorporated nonnegative latent factor analysis”, IEEE/CAA Journal of Automatica Sinica, vol. 10, no. 6, pp. 1388–1406, 2023

    2023

  54. [63]

    W. Li, X. Luo, H. Yuan, and M. Zhou, ”A momentum-accelerated Hessian-vector-based latent factor analysis model”, IEEE Transactions on Services Computing, vol. 16, no. 2, pp. 830–844, 2022

    2022

  55. [64]

    F. Bi, T. He, Y . Xie, and X. Luo, ”Two-stream graph convolutional network-incorporated latent feature analysis”, IEEE Transactions on Services Computing, vol. 16, no. 4, pp. 3027–3042, 2023

    2023

  56. [65]

    D. Wu, M. Shang, X. Luo, and Z. Wang, ”An L 1-and-L 2-norm-oriented latent factor model for recommender systems”, IEEE Transactions on Neural Networks and Learning Systems, vol. 33, no. 10, pp. 5775–5788, 2021

    2021

  57. [66]

    Z. Liu, G. Yuan, and X. Luo, ”Symmetry and nonnegativity-constrained matrix factorization for community detection”, IEEE/CAA Journal of Automatica Sinica, vol. 9, no. 9, pp. 1691–1693, 2022

    2022

  58. [67]

    Qin and X

    W. Qin and X. Luo, ”Asynchronous parallel fuzzy stochastic gradient descent for high-dimensional incomplete data representation”, IEEE Transactions on Fuzzy Systems, vol. 32, no. 2, pp. 445–459, 2023

    2023

  59. [68]

    X. Shi, Q. He, X. Luo, Y . Bai, and M. Shang, ”Large-scale and scalable latent factor analysis via distributed alternative stochastic gradient descent for recommender systems”, IEEE Transactions on Big Data, vol. 8, no. 2, pp. 420–431, 2020

    2020

  60. [69]

    X. Luo, Y . Zhou, Z. Liu, and M. Zhou, ”Fast and accurate non-negative latent factor analysis of high-dimensional and sparse matrices in recommender systems”, IEEE Transactions on Knowledge and Data Engineering, vol. 35, no. 4, pp. 3897–3911, 2021

    2021

  61. [70]

    D. Wu, X. Luo, M. Shang, Y . He, G. Wang, and X. Wu, ”A data-characteristic-aware latent factor model for web services QoS prediction”, IEEE Transactions on Knowledge and Data Engineering, vol. 34, no. 6, pp. 2525–2538, 2020

    2020

  62. [71]

    X. Luo, H. Wu, and Z. Li, ”NeuLFT: A novel approach to nonlinear canonical polyadic decomposition on high-dimensional incomplete tensors”, IEEE Transactions on Knowledge and Data Engineering, vol. 35, no. 6, pp. 6148–6166, 2022

    2022

  63. [72]

    D. Wu, P. Zhang, Y . He, and X. Luo, ”A double-space and double-norm ensembled latent factor model for highly accurate web service QoS prediction”, IEEE Transactions on Services Computing, vol. 16, no. 2, pp. 802–814, 2022

    2022

  64. [73]

    D. Wu, X. Luo, Y . He, and M. Zhou, ”A prediction-sampling-based multilayer-structured latent factor model for accurate representation to high-dimensional and sparse data”, IEEE transactions on neural networks and learning systems, vol. 35, no. 3, pp. 3845–3858, 2022

    2022

  65. [74]

    L. Hu, S. Yang, X. Luo, and M. Zhou, ”An algorithm of inductively identifying clusters from attributed graphs”, IEEE Transactions on Big Data, vol. 8, no. 2, pp. 523–534, 2020

    2020

  66. [75]

    X. Luo, M. Chen, H. Wu, Z. Liu, H. Yuan, and M. Zhou, ”Adjusting learning depth in nonnegative latent factorization of tensors for accurately modeling temporal patterns in dynamic QoS data”, IEEE Transactions on Automation Science and Engineering, vol. 18, no. 4, pp. 2142–2155, 2021

    2021

  67. [76]

    X. Luo, Y . Zhong, Z. Wang, and M. Li, ”An alternating-direction-method of multipliers-incorporated approach to symmetric non-negative latent factor analysis”, IEEE Transactions on Neural Networks and Learning Systems, vol. 34, no. 8, pp. 4826–4840, 2021

    2021

  68. [77]

    Z. He, M. Lin, X. Luo, and Z. Xu, ”Structure-Preserved Self-Attention for Fusion Image Information in Multiple Color Spaces”, IEEE Transactions on Neural Networks and Learning Systems, 2024

    2024

  69. [78]

    Wu et al., ”An outlier-resilient autoencoder for representing high-dimensional and incomplete data”, IEEE Transactions on Emerging Topics in Computational Intelligence, 2024

    D. Wu et al., ”An outlier-resilient autoencoder for representing high-dimensional and incomplete data”, IEEE Transactions on Emerging Topics in Computational Intelligence, 2024

    2024

  70. [79]

    Z. Liu, X. Luo, and M. Zhou, ”Symmetry and graph bi-regularized non-negative matrix factorization for precise community detection”, IEEE Transactions on Automation Science and Engineering, vol. 21, no. 2, pp. 1406–1420, 2023

    2023

  71. [80]

    M. Chen, L. Tao, J. Lou, and X. Luo, ”Latent-factorization-of-tensors-incorporated battery cycle life prediction”, IEEE/CAA Journal of Automatica Sinica, 2024

    2024

  72. [81]

    D. Wu, Z. Li, Z. Yu, Y . He, and X. Luo, ”Robust low-rank latent feature analysis for spatiotemporal signal recovery”, IEEE Transactions on Neural Networks and Learning Systems, 2023

    2023

  73. [82]

    M. Lin, J. Liu, H. Chen, X. Xu, X. Luo, and Z. Xu, ”A 3D Convolution-Incorporated Dimension Preserved Decomposition Model for Traffic Data Prediction”, IEEE Transactions on Intelligent Transportation Systems, 2024

    2024

  74. [83]

    H. Yang, M. Lin, H. Chen, X. Luo, and Z. Xu, ”Latent Factor Analysis Model With Temporal Regularized Constraint for Road Traffic Data Imputation”, IEEE Transactions on Intelligent Transportation Systems, 2024

    2024

  75. [84]

    D. Wu, P. Zhang, Y . He, and X. Luo, ”Mmlf: Multi-metric latent feature analysis for high-dimensional and incomplete data”, IEEE transactions on services computing, vol. 17, no. 2, pp. 575–588, 2023

    2023

  76. [85]

    W. Qin, X. Luo, and M. Zhou, ”Adaptively-accelerated parallel stochastic gradient descent for high-dimensional and incomplete data representation learning”, IEEE Transactions on Big Data, vol. 10, no. 1, pp. 92–107, 2023

    2023

Pith tools

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