Pith. sign in

REVIEW 4 major objections 5 minor 42 references

When Noisy Labels Meet Class Imbalance on Graphs: A Graph Augmentation Method with LLM and Pseudo Label

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read GraphALP proposes that class-imbalanced graphs with noisy labels are best handled by LLM-generated minority nodes, self-supervised pre-training on the balanced graph, and reweighted pseudo-labeling followed by a second LLM oversampling…

desk verdict A coherent LLM+pseudo-label pipeline for a real problem, but the evaluation protocol is ambiguous enough that the headline gains may not be real. read the letter →

arxiv 2507.18153 v2 pith:NRC2SKCP submitted 2025-07-24 cs.LG cs.AI

Riting Xia , Rucong Wang , Yulin Liu , Anchen Li , Xueyan Liu , Yan Zhang This is my paper
classification cs.LGcs.AI
keywords graphnodeclassificationclassimbalancelabelnoiseaugmentationlargelanguagemodelspseudo-labelingoversamplingrobustlearning
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

Class-imbalanced graphs in the real world tend to arrive with noisy labels, yet prior graph methods typically tackle one of those problems and assume the other away. This paper proposes GraphALP, a single pipeline that treats both at once: an LLM writes realistic text for minority nodes with reliable labels, a sentence-embedding model and an MLP projection insert those synthetic nodes into the graph, and a confidence-based pseudo-labeling stage with class reweighting cleans up residual noise. The paper reports consistent gains in ACC, F1, and G-mean over state-of-the-art class-imbalance baselines on Cora, CiteSeer, Pubmed, and Wiki-CS, including G-mean improvements of 8.03 and 6.03 points over GraphDAO. If these results hold, graph classifiers no longer have to choose between fixing skewed classes and tolerating mislabeled nodes.

What carries the argument

The load-bearing object is the two-stage augmentation loop. In stage one, LLM-written minority text is turned into node representations by jina-embedding-v3 and aligned through an MLP projection with the original node attributes; the resulting class-balanced graph is encoded by a GraphSage autoencoder whose reconstruction losses include attribute reconstruction, structural reconstruction, and an edge predictor that wires synthetic nodes to similar real nodes. In stage two, pseudo-labels are selected only when confidence exceeds a threshold and the cross-entropy loss uses class weights $w_{C_i}=\max(1,N/n_i)$, so minority nodes are neither mistaken for noise nor starved of gradient. A second LLM oversampling pass rebalances the pseudo-label-skewed distribution. The mechanism that carries the argument is the combination: accurate synthetic minority labels lower the noise ratio at the source, and then rebalanced pseudo-labels keep the classifier from regressing toward the majority.

What would settle it

Train a discriminator to separate projected LLM-generated minority embeddings from real minority node features; if the two sets are cleanly separable, the alignment assumption fails and the reported oversampling gains cannot be attributed to semantic fidelity.

Watch

Extended reading notes

Core claim

GraphALP's central claim is that the interaction between class imbalance and label noise is the actual obstacle, and a coordinated augmentation loop can remove both. Synthetic minority nodes are generated by prompting an LLM to produce title-and-abstract style text for the minority classes, embedded with jina-embedding-v3, and projected by an MLP to match the original feature space; this class-balanced graph is pre-trained with an autoencoder and GraphSage before fine-tuning with a dynamically weighted cross-entropy loss that selects high-confidence pseudo-labels. A second LLM oversampling step then compensates for the majority-class skew that pseudo-labeling itself introduces. Across four real datasets with controlled imbalance and uniform label noise, the paper reports that GraphALP outperforms all baselines on ACC, F1, and G-mean, with the largest gains concentrated in balanced metrics like G-mean.

Load-bearing premise

The entire pipeline depends on the assumption that after an MLP projection, LLM-written synthetic node texts land in the same feature space as the graph's real node attributes, so that oversampling with them improves minority-class learning instead of injecting out-of-distribution examples.

Editorial extensions

If this is right

  • On the four benchmark graphs, GraphALP reports the highest ACC, F1, and G-mean among all tested baselines, with the largest G-mean advantage over GraphDAO being 8.03 points on Cora and 6.03 points on Wiki-CS.
  • The pseudo-labeling stage cuts the measured label noise ratio on Cora from 30.23% to 10.15%, so most of the supervision the GNN finally trains on is cleaner than the original labels.
  • In robustness sweeps, GraphALP stays ahead as both imbalance and noise get worse, and its edge over GraphDAO on CiteSeer grows from 1.55 points at 10% noise to 9.78 points at 50% noise.
  • The reported runtimes are lower than GraphMixup and GraphDAO on Cora, CiteSeer, and Wiki-CS, so the combined pipeline does not buy robustness at the cost of prohibitive training time.

Reading between the lines

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

  • If the alignment trick generalizes, LLM oversampling could replace interpolation-based synthesis on any text-attributed graph, including knowledge graphs or item-description recommendation graphs, which this paper does not evaluate.
  • A cheap extension would be to report pseudo-label noise after each stage on all datasets, which would separate the contribution of label cleaning from class rebalancing.
  • Because the LLM's knowledge is domain-bound, synthetic nodes for specialized vocabularies such as legal or biomedical text may drift off target; holding the pipeline fixed while varying the prompt template or generator would reveal how much of the gain depends on the specific LLM.
  • The two-stage rebalancing recipe can be bolted onto existing noisy-label GNNs: select confident pseudo-labels, measure the class skew they create, then oversample the shrunken classes.
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 / 5 minor

Summary. The paper studies node classification on graphs that simultaneously exhibit class imbalance and label noise, a setting the authors argue is underexplored. They propose GraphALP, a framework with three components: LLM-based oversampling of minority nodes (with SentenceBERT-style embedding and MLP projection), a self-supervised pre-training module (AE/GAE) on the augmented class-balanced graph, and a fine-tuning module that assigns pseudo-labels to high-confidence unlabeled nodes and uses them as additional training supervision, followed by a second LLM-based rebalancing step. Experiments on Cora, CiteSeer, PubMed, and Wiki-CS report consistent improvements in ACC, F1, and G-mean over several class-imbalance baselines, including gains of up to 8.03 G-mean points over GraphDAO. The paper also includes robustness studies varying imbalance and noise ratios, ablations, and a cost comparison.

Significance. The problem addressed—joint class imbalance and label noise on graphs—is practically important and relatively underexplored, so a well-validated method would be a useful contribution. The design of using LLM-generated synthetic minority nodes is interesting and avoids some limitations of feature-space oversampling. The pseudo-labeling with dynamic weighting is also a reasonable mechanism. However, the empirical evidence as presented is not fully convincing: the main claim rests on averaged results without error bars or significance tests, and the pseudo-labeling protocol risks training on the evaluation set. If those issues are resolved, the contribution could be significant, but at present the central empirical claim is not yet established.

major comments (4)
  1. [Fine-tuning Module with Pseudo-label, Eq. (20)] The paper never states that test nodes are excluded from the set YU used for pseudo-label selection. Under the standard transductive splits for Cora, CiteSeer, and PubMed, all non-training nodes—including the test split—are typically treated as unlabeled. If high-confidence test nodes are assigned pseudo-labels and then integrated into the training set (as stated after Eq. (22)), the reported ACC/F1/G-mean are computed on nodes whose predictions were directly optimized during training. This can inflate the reported gains and makes the comparison to baselines unfair. The authors must specify the exact split used for YU and rerun all experiments with the test split explicitly excluded from pseudo-labeling. If the results change materially, the conclusions must be revised.
  2. [Experiments, Table 1] All reported numbers in Table 1 are averages over 5 runs without standard deviations, confidence intervals, or statistical significance tests. The claimed improvements range from about 1 to 8 percentage points, which may be within run-to-run variation for GNNs on small graph datasets. The authors should report per-seed results with error bars and apply pairwise significance tests (e.g., paired t-test or Wilcoxon signed-rank test) between GraphALP and each baseline, especially for the headline G-mean gains.
  3. [Robust Analysis, Figure 4] The text in "Influence of Label Noise Ratio" states: "As the label noise ratio decreases, the performance of all methods deteriorates." This contradicts the expected and subsequently described behavior: lower noise should improve performance, and the later statement that at noise ratio 0.1 GraphALP beats GraphDAO by 1.55% while at 0.5 the gap is 9.78% implies the opposite trend. This is a load-bearing inconsistency in the robustness narrative and must be corrected and the corresponding numbers verified.
  4. [Experiment Setup and Loss Function, Eqs. (11), (20), (21)] Several free hyperparameters that directly control the method's behavior are not reported: the edge similarity threshold tau in Eq. (11), the pseudo-label confidence threshold tau in Eq. (20), the loss weights alpha, beta, gamma in Eq. (21), and the exact oversampling scale used for each dataset (e.g., Wiki-CS). Without these values, the method cannot be reproduced, and the sensitivity of the reported gains to these thresholds is unknown. The authors should provide the values and include a sensitivity analysis for at least the two thresholds and the loss weights.
minor comments (5)
  1. [Introduction and Related Work] The claims "first explicit study" and "first work to tackle these challenges via LLM" are too strong given existing LLM-based augmentation for imbalanced node classification (e.g., Wang et al. 2024a, cited in the paper) and the existence of noisy-label benchmarks like NoisyGL. Please soften the novelty statements and clearly position against these works.
  2. [Fine-tuning Module with Pseudo-label, Eq. (22)] In Eq. (22), the notation yP_i is used without explicitly defining it as the one-hot encoding of the selected pseudo-label from YP. Please define all symbols and specify how the pseudo-label loss term is masked for nodes that are not selected.
  3. [Ablation Study, Table 3] The ablation names Ours-PL, Ours-RB, Ours-W are not fully consistent with the text: the text says removing (1) pseudo-labeling module, (2) rebalancing module, and (3) weighted cross-entropy loss, but the table caption uses RB and W without defining which is which. Please clarify the definitions and make the table self-contained.
  4. [Throughout] There are multiple typos and formatting artifacts: "effciency" appears twice, Table 1 has "0.3921" instead of "39.21", and equations contain repeated letters such as "XXX", "AAA1", and "ZZZ". These should be cleaned up before submission.
  5. [Case Study and Visualization, Figures 5 and 6] The captions do not specify what is being visualized or how the reported noise ratio is computed. For Figure 6, please state the exact procedure used to compute the noise ratio after pseudo-labeling and whether the visualization is on training or test nodes.

Circularity Check

1 steps flagged · score 6.0 of 10

Reported test-set gains are inflated by using the model's own test-node pseudo-labels as training targets, so for those nodes the final predictions are fit to themselves.

  1. fitted input called prediction [Fine-tuning Module with Pseudo-label, Eq. (20) and Eq. (22); Experiments, Dataset and Evaluation Metrics]
    "YP = { ŷic ∈ ŶU | ŷic > τ } (20), where ŶU is the pseudo-label predictions for the unlabeled nodes. ... we integrate pseudo labels into the training set ... LC = −Σ wCi·yi log ŷi − Σ wCi·yP_i log ŷi (22), where yP_i ∈ YP is the pseudo label."

    The split description never states that test (or validation) nodes are excluded from the unlabeled set U; for the standard transductive splits used on these benchmarks, every non-training node is unlabeled. Eq. (20) therefore admits high-confidence test nodes into YP, and the text says these pseudo-labels are integrated into the training set. Eq. (22) then optimizes the classifier to output yP_i on those same nodes. The reported ACC/F1/G-mean on the pseudo-labeled part of the test set measure how well the model reproduced its own pre-fine-tuning predictions, a fitted input, rather than independent generalization. The comparison to baselines that do not pseudo-label test nodes is accordingly forced for the high-confidence subset.

full rationale

The LLM oversampling chain is not circular: synthetic nodes are generated from class-label prompts by an external LLM and embedded with an external LM; the projection in Eq. (7) is a learned alignment, not a definition of the target. The use of pseudo-labeling as self-training is not by itself a derivation-level circularity. The circular step is in the evaluation protocol: the paper's Eq. (20) collects pseudo-labels from all unlabeled nodes, the text says they are integrated into the training set, and Eq. (22) trains on them, without any statement that the test split is excluded. Under the transductive splits used for these datasets, the unlabeled set includes the test nodes, so the final predictions on those nodes are optimized to match the model's own earlier predictions. Reported gains such as the 8.03 G-mean improvement on Cora are thus, for the high-confidence fraction of the test set, self-consistency with fitted pseudo-labels rather than held-out predictive accuracy. No load-bearing self-citation chain or ansatz-smuggled-from-prior-work step was found; GraphDAO and other baselines serve as external comparators. The central empirical claim is therefore partially circular due to the test-set pseudo-labeling protocol.

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

The central claim rests on assumptions about LLM output reliability and pseudo-label trustworthiness, plus several hand-set hyperparameters whose values are not reported. No new physical or conceptual entities are introduced.

free parameters (4)
  • Oversampling scale = 0.8 (default)
    Controls how many synthetic minority nodes are added; chosen per dataset without sensitivity analysis.
  • Edge similarity threshold tau (Eq. 11) = not reported
    Determines initial edges for synthetic nodes; set by hand and not justified.
  • Pseudo-label confidence threshold tau (Eq. 20) = not reported
    Selects which unlabeled nodes receive pseudo-labels; value directly affects noise reduction and is not reported.
  • Loss weights alpha, beta, gamma (Eq. 21) = not reported
    Balance reconstruction losses; no values given in the paper.
assumptions (3)
  • domain assumption LLM-generated minority node text is label-accurate and semantically representative of the true class distribution.
    Used in Eq. (3) to generate synthetic nodes; if false, oversampling adds misleading data.
  • domain assumption The projection via MLP2 aligns the original feature space and the LLM embedding space well enough for synthetic nodes to be in-distribution.
    Invoked in Eq. (7); not validated in the paper.
  • ad hoc to paper High-confidence pseudo-labels are sufficiently accurate to serve as training targets.
    The pseudo-label selection in Eq. (20) relies on confidence thresholding; this is a common but unproven assumption for this setting.

how reviews work

0 comments
Cite this review

Pith. "Pith review of When Noisy Labels Meet Class Imbalance on Graphs: A Graph Augmentation Method with LLM and Pseudo Label." pith.science (2026). https://pith.science/paper/NRC2SKCP

@misc{pith2026250718153,
  author       = {Pith},
  title        = {Pith review of: When Noisy Labels Meet Class Imbalance on Graphs: A Graph Augmentation Method with LLM and Pseudo Label},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NRC2SKCP}},
  note         = {Machine review of arXiv:2507.18153}
}
read the original abstract

Class-imbalanced graph node classification is a practical yet underexplored research problem. Although recent studies have attempted to address this issue, they typically assume clean and reliable labels when processing class-imbalanced graphs. This assumption often violates the nature of real-world graphs, where labels frequently contain noise. Given this gap, this paper systematically investigates robust node classification for class-imbalanced graphs with noisy labels. We propose GraphALP, a novel Graph Augmentation framework based on Large language models (LLMs) and Pseudo-labeling techniques. Specifically, we design an LLM-based oversampling method to generate synthetic minority nodes, producing label-accurate minority nodes to alleviate class imbalance. Based on the class-balanced graphs, we develop a dynamically weighted pseudo-labeling method to obtain high-confidence pseudo labels to reduce label noise ratio. Additionally, we implement a secondary LLM-guided oversampling mechanism to mitigate potential class distribution skew caused by pseudo labels. Experimental results show that GraphALP achieves superior performance over state-of-the-art methods on class-imbalanced graphs with noisy labels.

Figures

Figures reproduced from arXiv: 2507.18153 by the authors.

Figure 1
Figure 1. The architecture of GraphALP. which generates synthetic minority node text via LLM, en￾codes it into representations via language model (LM), and balances graph class distribution. This module mitigates im￾balance and reduces label noise with high-quality synthetic nodes. (2) Self-Supervised Pre-Training Module, which constructs a class-balanced graph with synthetic nodes and learns representations in a self-supervi… view at source ↗
Figure 2
Figure 2. Prompt and node text generated by LLM on Cora. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 4
Figure 4. Results under different label noise ratio. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figures from the paper (3 more)
Figure 3
Figure 3. Figure 3: Results under different class imbalance ratio. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png]
Figure 5
Figure 5. Figure 5: Visualization of the Cora dataset. iments are performed on PubMed, Wiki-CS, Cora, and Cite￾Seer [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Node representation visualization. Color dots and [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

42 extracted references · 36 canonical work pages

  1. [1]

    Abdi, L.; and Hashemi, S. 2016. To combat multi-class imbalanced problems by means of over-sampling techniques. IEEE Transactions on Knowledge and Data Engineering., 28(1): 238--251

    work page 2016

  2. [2]

    Ando, S.; and Huang, C.-Y. 2017. Deep over-sampling framework for classifying imbalanced data. In PKDD, volume 10534, 770--785. Springer

    work page 2017

  3. [3]

    Buda, M.; Maki, A.; and Mazurowski, M. A. 2018. A systematic study of the class imbalance problem in convolutional neural networks. Neural Networks, 106: 249--259

    work page 2018

  4. [4]

    V.; Bowyer, K

    Chawla, N. V.; Bowyer, K. W.; Hall, L. O.; and Kegelmeyer, W. P. 2002. SMOTE: synthetic minority over-sampling technique. Journal of Artificial Intelligence Research, 16: 321--357

    work page 2002

  5. [5]

    Dai, E.; Aggarwal, C.; and Wang, S. 2021. NRGNN: learning a label noise resistant graph neural network on sparsely and noisily labeled graphs. In KDD, 227--236. ACM

    work page 2021

  6. [6]

    Dai, Q.; Wang, L.; Zhang, J.; Ding, W.; and Chen, L. 2025. GQEO: nearest neighbor graph-based generalized quadrilateral element oversampling for class-imbalance problem. Neural Networks, 184: 107107

    work page 2025

  7. [7]

    DeepSeek - AI; Liu, A.; Feng, B.; Xue, B.; Wang, B.; Wu, B.; Lu, C.; Zhao, C.; Deng, C.; Zhang, C.; Ruan, C.; Dai, D.; Guo, D.; Yang, D.; Chen, D.; Ji, D.; Li, E.; Lin, F.; Dai, F.; Luo, F.; Hao, G.; Chen, G.; Li, G.; Zhang, H.; Bao, H.; Xu, H.; Wang, H.; Zhang, H.; Ding, H.; Xin, H.; Gao, H.; Li, H.; Qu, H.; Cai, J. L.; Liang, J.; Guo, J.; Ni, J.; Li, J....

    arXiv 2024

  8. [8]

    Fang, T.; Zhang, Y.; Yang, Y.; Wang, C.; and Chen, L. 2023. Universal prompt tuning for graph neural networks. In NeurIPS

    work page 2023

Show all 42 references
  1. [9]

    L.; Ying, Z.; and Leskovec, J

    Hamilton, W. L.; Ying, Z.; and Leskovec, J. 2017. Inductive representation learning on large graphs. In NIPS, 1024--1034

    2017

  2. [10]

    Khoshraftar, S.; Abedini, N.; and Hajian, A. 2025. GraphiT: efficient node classification on text-attributed graphs with prompt optimized llms. In WWW, 1824--1829. ACM

    2025

  3. [11]

    Li, A.; and Yang, B. 2025. Dual graph denoising model for social recommendation. In WWW, 347--356. ACM

    2025

  4. [12]

    Li, D.; Li, X.; Gan, Z.; Li, Q.; Qu, B.; and Wang, J. 2025. Rethinking the impact of noisy labels in graph classification: a utility and privacy perspective. Neural Networks, 182: 106919

    2025

  5. [13]

    Li, X.; Li, Q.; Li, D.; Qian, H.; and Wang, J. 2024. Contrastive learning of graphs under label noise. Neural Networks, 172: 106113

    2024

  6. [14]

    Li, Y.; Yin, J.; and Chen, L. 2021. Unified robust training for graph neural networks against label noise. In PAKDD, volume 12712 of Lecture Notes in Computer Science, 528--540. Springer

    2021

  7. [15]

    Liu, Y.; Ao, X.; Qin, Z.; Chi, J.; Feng, J.; Yang, H.; and He, Q. 2021. Pick and choose: a GNN-based imbalanced learning approach for fraud detection. In WWW, 3168--3177. ACM

    2021

  8. [16]

    Liu, Z.; Li, Y.; Chen, N.; Wang, Q.; Hooi, B.; and He, B. 2025. A survey of imbalanced learning on graphs: problems, techniques, and future directions. IEEE Transactions on Knowledge and Data Engineering., 37(6): 3132--3152

    2025

  9. [17]

    Long, L.; Wang, R.; Xiao, R.; Zhao, J.; Ding, X.; Chen, G.; and Wang, H. 2024. On llms-driven synthetic data generation, curation, and evaluation: a survey. In ACL, 11065--11082. Association for Computational Linguistics

    2024

  10. [18]

    Lu, Q.; and Getoor, L. 2003. Link-based classification. In ICML, 496--503

    2003

  11. [19]

    Ma, Y.; Tian, Y.; Moniz, N.; and Chawla, N. V. 2025. Class-imbalanced learning on graphs: a survey. ACM Computing Surveys, 57(8): 207:1--207:16

    2025

  12. [20]

    Mernyei, P.; and Cangea, C. 2020. Wiki-CS: A wikipedia-based benchmark for graph neural networks. CoRR, abs/2007.02901

    2020 arXiv

  13. [21]

    Namata, G.; London, B.; Getoor, L.; and Huang, B. 2012. Query-driven active surveying for collective classification. In MLG

    2012

  14. [22]

    Sen, P.; Namata, G.; Bilgic, M.; Getoor, L.; Gallagher, B.; and Eliassi - Rad, T. 2008. Collective classification in network data. AI MAGAZINE, 29(3): 93--106

    2008

  15. [23]

    Song, H.; Kim, M.; Park, D.; Shin, Y.; and Lee, J.-G. 2023. Learning from noisy labels With deep neural networks: a Survey. IEEE Transactions on Neural Networks and Learning Systems, 34(11): 8135--8153

    2023

  16. [24]

    K.; G \" u nther, M.; Wang, B.; Krimmel, M.; Wang, F.; Mastrapas, G.; Koukounas, A.; Wang, N.; and Xiao, H

    Sturua, S.; Mohr, I.; Akram, M. K.; G \" u nther, M.; Wang, B.; Krimmel, M.; Wang, F.; Mastrapas, G.; Koukounas, A.; Wang, N.; and Xiao, H. 2024. Jina-embeddings-v3: multilingual embeddings With task Lora. CoRR, abs/2409.10173

    2024 arXiv

  17. [25]

    Sun, H.; Li, J.; and Zhu, X. 2025. A novel expandable borderline smote over-sampling method for class imbalance problem. IEEE Transactions on Knowledge and Data Engineering., 37(5): 2183--2199

    2025

  18. [26]

    Tan, Y.; Liu, Y.; Long, G.; Jiang, J.; Lu, Q.; and Zhang, C. 2023. Federated learning on non-iid graphs via structural knowledge sharing. In AAAI, 9953--9961. AAAI

    2023

  19. [27]

    K.; and Zheng, M

    Tao, Y.; Sun, J.; Yang, H.; Chen, L.; Wang, X.; Yang, W.; Du, D. K.; and Zheng, M. 2023. Local and global logit adjustments for long-tailed learning. In ICCV, 11749--11758. IEEE

    2023

  20. [28]

    Wang, L.; Wang, Y.; Ni, B.; Zhao, Y.; and Derr, T. 2024 a . Large language model-based augmentation for imbalanced node classification on text-attributed graphs. arXiv preprint arXiv:2410.16882

    2024

  21. [29]

    Wang, Y.; Chu, Z.; Ouyang, X.; Wang, S.; Hao, H.; Shen, Y.; Gu, J.; Xue, S.; Zhang, J.; Cui, Q.; Li, L.; Zhou, J.; and Li, S. 2024 b . LLMRG: improving recommendations through large language model reasoning graphs. In AAAI, 19189--19196. AAAI

    2024

  22. [30]

    Wang, Z.; Guo, J.; Liang, J.; Liang, J.; Cheng, S.; and Zhang, J. 2025 a . Graph segmentation and contrastive enhanced explainer for graph neural networks. In AAAI, 21393--21401. AAAI

    2025

  23. [31]

    Wang, Z.; Ma, S.; Wang, K.; and Zhuang, Z. 2025 b . Rule-guided graph neural networks for explainable knowledge graph reasoning. In AAAI, 12784--12791. AAAI

    2025

  24. [32]

    Wang, Z.; Sun, D.; Zhou, S.; Wang, H.; Fan, J.; Huang, L.; and Bu, J. 2024 c . NoisyGL: a comprehensive benchmark for graph neural networks under label noise. In NeurIPS

    2024

  25. [33]

    Wu, L.; Xia, J.; Gao, Z.; Lin, H.; Tan, C.; and Li, S. Z. 2022. GraphMixup: improving class-imbalanced node classification by reinforcement mixup and self-supervised context prediction. In PKDD, volume 13716 of Lecture Notes in Computer Science, 519--535. Springer

    2022

  26. [34]

    Xia, J.; Lin, H.; Xu, Y.; Tan, C.; Wu, L.; Li, S.; and Li, S. Z. 2024 a . GNN cleaner: label cleaner for graph structured data. IEEE Transactions on Knowledge and Data Engineering, 36(2): 640--651

    2024

  27. [35]

    Xia, R.; Liu, H.; Li, A.; Liu, X.; Zhang, Y.; Zhang, C.; and Yang, B. 2025. Incomplete graph learning: a comprehensive survey. CoRR, abs/2502.12412

    2025 arXiv

  28. [36]

    Xia, R.; Zhang, C.; Zhang, Y.; Liu, X.; and Yang, B. 2024 b . A novel graph oversampling framework for node classification in class-imbalanced graphs. SCIENCE CHINA Information Sciences, 67(6)

    2024

  29. [37]

    Xu, W.; Wang, P.; Zhao, Z.; Wang, B.; Wang, X.; and Wang, Y. 2024. When imbalance meets imbalance: structure-driven learning for imbalanced graph classification. In WWW, 905--913. ACM

    2024

  30. [38]

    Yuan, B.; and Ma, X. 2012. Sampling + reweighting: boosting the performance of adaBoost on imbalanced datasets. In IJCNN, 1--6. IEEE

    2012

  31. [39]

    Zhao, T.; Zhang, X.; and Wang, S. 2021. GraphSMOTE: imbalanced node classification on graphs with graph neural networks. In WSDM, 833--841. ACM

    2021

  32. [40]

    Zhong, X.; and Liang, H. 2024. On the scale-free property of citation networks: an empirical study. In WWW, 541--544. ACM

    2024

  33. [41]

    , " * 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...

  34. [42]

    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 gl...

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.