Pith. sign in

REVIEW 3 major objections 4 minor 61 references

Training-free Heterogeneous Graph Condensation via Data Selection

T0 review · 3 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read FreeHGC claims that heterogeneous graph condensation can be done without any model training, and that this training-free approach beats the trained baseline HGCond in accuracy, speed, and generalization.

desk verdict First training-free heterogeneous graph condensation with a strong empirical showing; the theory section and the topology decomposition need cleanup before I'd trust the framework's generality. read the letter →

arxiv 2412.16250 v1 pith:PVIQO77S submitted 2024-12-20 cs.LG

Yuxuan Liang , Wentao Zhang , Xinyi Gao , Ling Yang , Chong Chen , Hongzhi Yin , Yunhai Tong , Bin Cui This is my paper · ORCID
classification cs.LG MSC 68T0705C85
keywords heterogeneousgraphcondensationtraining-freedataselectionneuralnetworksmeta-pathsubmodularoptimizationPersonalizedPageRanknodeclassification
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 argues that heterogeneous graph condensation does not need to train a relay model: it can be reframed as a data selection problem. FreeHGC scores target-type nodes by combining receptive-field coverage with meta-path diversity, then condenses the remaining node types with Personalized PageRank and mean-aggregation synthesis. The claim is that this training-free procedure beats the existing trained method HGCond in accuracy, speed, and generalization across seven heterogeneous graph datasets, while allowing the condensation ratio to grow without the overfitting degradation seen in trained condensation. If true, graph condensation becomes a fast pre-processing step that any heterogeneous graph neural network can consume.

What carries the argument

The load-bearing object is the unified data selection score for target nodes, $F(S) = R(S)/|\hat{R}| + (1 - J(S))$, where $R(S)$ is the size of the union of receptive fields activated by the selected nodes under a given meta-path and $J(S)$ is the normalized Jaccard similarity of that meta-path with all other meta-paths. Because both summands are argued to be submodular, the score enables a greedy selection with an approximation guarantee. The second piece is the root-father-leaf taxonomy of node types: father types are condensed by ranking nodes with Personalized PageRank on symmetrized meta-path adjacency matrices, and leaf types are synthesized by mean-aggregating the features of each father's neighbors, with reverse edges added to preserve father-to-father two-hop information.

What would settle it

Run FreeHGC on a heterogeneous graph where the father and leaf roles can be permuted among the non-target types; if the best role assignment varies unpredictably with the dataset and FreeHGC's advantage over random sampling disappears under another assignment, the root-father-leaf premise is falsified.

Watch

Extended reading notes

Core claim

FreeHGC's central claim is that a training-free, data-selection approach can outperform the trained gradient-matching method HGCond for heterogeneous graph condensation. The target-type selection score $F(S) = R(S)/|\hat{R}| + (1 - J(S))$ combines submodular receptive-field maximization with a Jaccard-based meta-path diversity term, and the greedy algorithm inherits the $(1 - 1/e)$ approximation guarantee for submodular maximization. Other-type nodes are condensed by treating them as 'father' nodes selected by Personalized PageRank importance and 'leaf' nodes synthesized by mean aggregation, based on the observation that neighbor attention in HGNNs is dispensable. Across ACM, DBLP, IMDB, Freebase, MUTAG, AM, and AMiner, FreeHGC reports higher test accuracy than HGCond and coreset baselines, at a fraction of the condensation time and with accuracy that keeps rising as the condensation ratio grows.

Load-bearing premise

The load-bearing assumption is that every heterogeneous graph can be cleanly split into root, father, and leaf node types, and that condensing fathers by Personalized PageRank plus synthesizing leaves by mean aggregation preserves the structure that matters.

Editorial extensions

If this is right

  • Heterogeneous graph condensation becomes a pre-processing routine: no relay model, no bi-level optimization, no nested training loops, and no clustering or OPS initialization.
  • Accuracy improves as the condensation ratio grows, so practitioners can pick any budget without the overfitting-induced performance drops reported for HGCond.
  • The condensed graph transfers across meta-path-based and meta-path-free HGNNs (HAN, HGT, HGB, SeHGNN), so one condensed graph can serve many downstream models.
  • On the tested graphs the procedure cuts storage by roughly 95 to 99 percent and reduces downstream training time to a small fraction of whole-graph training.

Reading between the lines

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

  • This suggests that heterogeneous graph condensation with a fixed budget is an influence-maximization problem on a multiplex of meta-path adjacency matrices, which may admit coreset-style guarantees beyond node classification.
  • The root-father-leaf taxonomy is a hidden prior; for knowledge graphs with eight or more node types and many relations, defining father and leaf roles may require per-dataset design, and automatic role discovery would be a natural extension.
  • The mean-aggregation synthesis step ties the method to the finding that neighbor attention is unnecessary; if a future HGNN relies on fine-grained neighbor weighting, FreeHGC's condensed leaf nodes would likely need to store richer statistics than a mean.
  • PageRank-based father selection could be replaced or augmented by other importance measures, so the framework is a template rather than a single fixed algorithm.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper proposes FreeHGC, a training-free method for heterogeneous graph condensation. Target-type nodes are selected with a unified score combining receptive-field coverage and a meta-path Jaccard diversity term; other-type nodes are condensed with Personalized PageRank for 'father' types and mean-aggregation synthesis for 'leaf' types. Experiments on ACM, DBLP, IMDB, Freebase, AMiner, MUTAG, and AM report higher accuracy than HGCond and coreset baselines at most condensation ratios, better generalization across HGB/HGT/HAN/SeHGNN, and large speedups. The paper also claims a submodularity-based approximation guarantee for the selection criterion.

Significance. If the results hold, FreeHGC is a useful contribution: it decouples condensation from trained relay models and makes condensation much faster and model-agnostic. The empirical evaluation is broad (seven datasets, including one large-scale), reports mean plus variance, includes generalization to four HGNN models, and the code is public, which are genuine strengths. The core empirical claim is credible. However, two load-bearing gaps—an unspecified father/leaf decomposition for complex datasets and a mismatch between the stated submodularity theory and the implemented algorithm—must be addressed before the paper's full claims are acceptable.

major comments (3)
  1. [§IV-B, Eq. (8) and Algorithm 1] The theoretical claim is not matched by the implementation. The paper states that F(S) is submodular and that greedy maximization gives a (1−1/e) approximation, but Algorithm 1 (line 10) does not perform greedy set maximization; it computes per-node scores and selects the top-k. The greedy guarantee therefore does not apply to the actual algorithm. In addition, submodularity of 1−J(S) is asserted for the specific aggregate Jaccard measure of Eqs. (5)–(7) rather than proved. This does not invalidate the empirical results, but the theoretical section must either be aligned with the implementation (e.g., by presenting a genuinely greedy variant) or the (1−1/e) approximation claim should be removed or explicitly made conditional on a different selection procedure.
  2. [§IV-C, Figure 5, and Algorithm 2] The root/father/leaf decomposition is not specified for datasets with many node types, which is load-bearing because the two other-type condensation strategies are defined only after this decomposition. The paper never states which node types of Freebase (8 types), MUTAG (7 types), or AM (7 types) are treated as father types versus leaf types, nor does it give a general rule for assigning these roles. The ablation study in Table VIII covers only ACM, DBLP, and AMiner, so the necessity and correctness of the father/leaf split are empirically unchecked exactly on the datasets where the topology is least obvious. As written, the Freebase, MUTAG, and AM results are not reproducible from the paper description; the authors should provide the per-dataset type mappings or a precise algorithm that derives them automatically.
  3. [§IV-C, Time Complexity] The stated complexity for condensing target-type nodes, O(α N_tgt^2 + N_tgt log N_tgt), is not derived and is not credible for the reported large-scale experiments. For AMiner, N_tgt ≈ 4.89×10^6 and α ranges from 0.0005 to 0.008, which gives roughly 10^10 to 10^11 operations under this formula; this is inconsistent with the modest condensation times shown in Figure 8. The authors should either identify which step causes the quadratic term, provide a sparse/approximate implementation that avoids it, or correct the complexity expression.
minor comments (4)
  1. [Table IV, Freebase row] The reported average for Herding-HG on Freebase is 40.29, but the four preceding accuracies (49.81, 47.22, 43.44, 51.17) average to approximately 47.91; please correct this arithmetic or transcription error.
  2. [Figure 7 caption] The caption says 'Left: ACM dataset. Left: IMDB dataset.'; the second label should be 'Right'.
  3. [§V-A] The text repeatedly uses 'corset methods'; the standard term is 'coreset methods'.
  4. [§V-G] The text says 'Figure VI shows that FreeHGC performs best', but the referenced results are in Table VI; please correct the cross-reference.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: FreeHGC's training-free selection criterion is a structural heuristic validated on held-out test nodes; self-citations are background only.

full rationale

FreeHGC's derivation chain is self-contained and does not reduce any prediction to its own inputs by construction. The target-type selection criterion F(S) = R(S)/|R| + (1 - J(S)) is built from graph-structural quantities: receptive fields computed from normalized meta-path adjacency products (Eq. 1) and a Jaccard-based diversity term (Eqs. 4-8). No fitted parameter is embedded in this criterion, and class proportions are used only to split the condensation budget across classes, not to set the score. The father-type condensation uses Personalized PageRank (Eq. 11) on symmetric meta-path matrices, and the leaf-type synthesis uses mean aggregation motivated by the external SeHGNN finding [18], not by FreeHGC's own outputs; generalization is then checked on HGB, HGT, HAN, and SeHGNN (Table IV) using a held-out 70% test split. The claimed root/father/leaf topology decomposition (Section IV-C, Figure 5) is an asserted modeling assumption and a reproducibility/correctness risk on Freebase and AM, but it is not circular: the method does not define the target accuracy in terms of that decomposition. Self-citations [25], [43], [44], and [45] appear only as related-work background and are not load-bearing in the derivation; the submodularity argument cites external results [48]-[50], [53]-[55]. No equation in the paper is equivalent to another by construction, and no fitted value is renamed as a prediction. Therefore, the appropriate finding is no significant circularity.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The central claim rests on a small set of structural assumptions: submodularity of the score, a fixed root/father/leaf topology, mean aggregation as a universal proxy for HGNN neighbor attention, and Personalized PageRank as a measure of node importance. None of these is machine-checked, and the submodularity step is asserted without a proof matching the implemented algorithm.

free parameters (3)
  • K, meta-path hop limit = 3,4,5,2,1,1,2 for ACM, DBLP, IMDB, Freebase, MUTAG, AM, AMiner
    Chosen per dataset; no sensitivity analysis or selection criterion reported.
  • alpha (PPR restart probability) = not reported
    Appears in Eq. (11) and controls father-node influence scores; no default or tuning protocol is given.
  • objective weight between coverage and diversity = 1 (equal weights)
    Eq. (8) sums R/|R| and 1-J without a learned or tuned coefficient; no ablation of this weight is reported.
assumptions (6)
  • standard math R(S), the union of receptive fields, is a submodular coverage function.
    Standard coverage submodularity, used in Section IV-B Eq. (3).
  • ad hoc to paper 1-J(S) is submodular for the meta-path Jaccard similarity defined in Eq. (5)-(7).
    Asserted with citations [53,54] but not proved for this exact normalized sum; the implemented Algorithm 1 does not run greedy maximization.
  • domain assumption Replacing node-level attention with mean aggregation preserves HGNN accuracy.
    Borrowed from SeHGNN's empirical finding [18] to justify Eq. (14) leaf synthesis.
  • domain assumption Every dataset has a root/father/leaf hierarchy with father types bridging root and leaf.
    Used in Algorithm 2 and Figure 5; Freebase and AM have many node types and no check for general topology.
  • domain assumption Personalized PageRank on symmetrized meta-path adjacency measures father-node importance.
    Eq. (10)-(11) assumes PPR scores correlate with condensation value.
  • domain assumption Enumerating all meta-paths up to length K captures useful semantics without expert meta-path selection.
    Section IV-A; K itself is tuned per dataset.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Training-free Heterogeneous Graph Condensation via Data Selection." pith.science (2026). https://pith.science/paper/PVIQO77S

@misc{pith2026241216250,
  author       = {Pith},
  title        = {Pith review of: Training-free Heterogeneous Graph Condensation via Data Selection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PVIQO77S}},
  note         = {Machine review of arXiv:2412.16250}
}
read the original abstract

Efficient training of large-scale heterogeneous graphs is of paramount importance in real-world applications. However, existing approaches typically explore simplified models to mitigate resource and time overhead, neglecting the crucial aspect of simplifying large-scale heterogeneous graphs from the data-centric perspective. Addressing this gap, HGCond introduces graph condensation (GC) in heterogeneous graphs and generates a small condensed graph for efficient model training. Despite its efficacy in graph generation, HGCond encounters two significant limitations. The first is low effectiveness, HGCond excessively relies on the simplest relay model for the condensation procedure, which restricts the ability to exert powerful Heterogeneous Graph Neural Networks (HGNNs) with flexible condensation ratio and limits the generalization ability. The second is low efficiency, HGCond follows the existing GC methods designed for homogeneous graphs and leverages the sophisticated optimization paradigm, resulting in a time-consuming condensing procedure. In light of these challenges, we present the first Training \underline{Free} Heterogeneous Graph Condensation method, termed FreeHGC, facilitating both efficient and high-quality generation of heterogeneous condensed graphs. Specifically, we reformulate the heterogeneous graph condensation problem as a data selection issue, offering a new perspective for assessing and condensing representative nodes and edges in the heterogeneous graphs. By leveraging rich meta-paths, we introduce a new, high-quality heterogeneous data selection criterion to select target-type nodes. Furthermore, two training-free condensation strategies for heterogeneous graphs are designed to condense and synthesize other-types nodes effectively.

Figures

Figures reproduced from arXiv: 2412.16250 by the authors.

Figure 1
Figure 1. Comparisons of HGCond and our proposed training [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Empirical analysis of low accuracy and efficiency. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. The workflow of FreeHGC, where L represents the number of meta-paths. plex training, thereby releasing the advantages of HGNN to further improve effectiveness. Once condensation is isolated, the goal of graph condensation is transformed into selecting and condensing high-quality data to protect graph structure information, which has two benefits. First, this allows us to intuitively use a flexible condensation ratio… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: An illustration of meta-paths similaritiy problem. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Topological structures of different datasets. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: An illustration of information loss minimization. [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Performance at different condensation ratio. Left: ACM [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

61 extracted references · 52 canonical work pages

  1. [1]

    Stochastic weight completion for road networks using graph convolutional networks,

    J. Hu, C. Guo, B. Yang, and C. S. Jensen, “Stochastic weight completion for road networks using graph convolutional networks,” in 35th IEEE International Conference on Data Engineering, ICDE 2019, Macao, China, April 8-11, 2019 , 2019, pp. 1274–1285

    work page 2019

  2. [2]

    Predicting path failure in time-evolving graphs,

    J. Li, Z. Han, H. Cheng, J. Su, P. Wang, J. Zhang, and L. Pan, “Predicting path failure in time-evolving graphs,” in Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 2019, pp. 1279–1289

    work page 2019

  3. [3]

    Medical entity disambiguation using graph neural networks,

    A. Vretinaris, C. Lei, V . Efthymiou, X. Qin, and F. ¨Ozcan, “Medical entity disambiguation using graph neural networks,” in SIGMOD ’21: International Conference on Management of Data, Virtual Event, China, June 20-25, 2021 , 2021, pp. 2310–2318

    work page 2021

  4. [4]

    Graph transformation policy network for chemical reaction prediction,

    K. Do, T. Tran, and S. Venkatesh, “Graph transformation policy network for chemical reaction prediction,” in Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 2019, pp. 750–760

    work page 2019

  5. [5]

    Creating embed- dings of heterogeneous relational datasets for data integration tasks,

    R. Cappuzzo, P. Papotti, and S. Thirumuruganathan, “Creating embed- dings of heterogeneous relational datasets for data integration tasks,” in Proceedings of the 2020 International Conference on Management of Data, SIGMOD, 2020, pp. 1335–1349

    work page 2020

  6. [6]

    Estimating node importance values in heterogeneous information net- works,

    C. Huang, Y . Fang, X. Lin, X. Cao, W. Zhang, and M. E. Orlowska, “Estimating node importance values in heterogeneous information net- works,” in 38th IEEE International Conference on Data Engineering, ICDE, 2022, pp. 846–858

    work page 2022

  7. [7]

    Discovering maximal motif cliques in large heterogeneous information networks,

    J. Hu, R. Cheng, K. C. Chang, A. Sankar, Y . Fang, and B. Y . H. Lam, “Discovering maximal motif cliques in large heterogeneous information networks,” in 35th IEEE International Conference on Data Engineering, ICDE. IEEE, 2019, pp. 746–757

    work page 2019

  8. [8]

    Are we really making much progress?: Revisiting, benchmarking and refining heterogeneous graph neural networks,

    Q. Lv, M. Ding, Q. Liu, Y . Chen, W. Feng, S. He, C. Zhou, J. Jiang, Y . Dong, and J. Tang, “Are we really making much progress?: Revisiting, benchmarking and refining heterogeneous graph neural networks,” in KDD ’21: The 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , 2021, pp. 1150–1160

    work page 2021

Show all 61 references
  1. [9]

    Heterogeneous graph attention network,

    X. Wang, H. Ji, C. Shi, B. Wang, Y . Ye, P. Cui, and P. S. Yu, “Heterogeneous graph attention network,” in The World Wide Web Conference, WWW 2019 , 2019, pp. 2022–2032

    2019

  2. [10]

    Poskhg: A position-aware knowledge hypergraph model for link prediction,

    Z. Chen, X. Wang, C. Wang, and Z. Li, “Poskhg: A position-aware knowledge hypergraph model for link prediction,” Data Science and Engineering, vol. 8, no. 2, pp. 135–145, 2023

    2023

  3. [11]

    Few-shot relation prediction of knowledge graph via convolutional neural network with self-attention,

    S. Zhong, J. Wang, K. Yue, L. Duan, Z. Sun, and Y . Fang, “Few-shot relation prediction of knowledge graph via convolutional neural network with self-attention,” Data Science and Engineering , vol. 8, no. 4, pp. 385–395, 2023

    2023

  4. [12]

    Multiple types of disease-associated rnas identification for disease prognosis and therapy using heterogeneous graph learning,

    W. Zhang, H. Wei, W. Zhang, H. Wu, and B. Liu, “Multiple types of disease-associated rnas identification for disease prognosis and therapy using heterogeneous graph learning,” Science China Information Sci- ences, vol. 67, no. 8, p. 189103, 2024

    2024

  5. [13]

    Label-aware chinese event detection with heterogeneous graph attention network,

    S.-Y . Cui, B.-W. Yu, X. Cong, T.-W. Liu, Q.-F. Tan, and J.-Q. Shi, “Label-aware chinese event detection with heterogeneous graph attention network,” Journal of Computer Science and Technology , vol. 39, no. 1, pp. 227–242, 2024

    2024

  6. [14]

    Meta-learning based few-shot link prediction for emerging knowledge graph,

    Y .-F. Zhang, W. Chen, P.-P. Zhao, J.-J. Xu, J.-H. Fang, and L. Zhao, “Meta-learning based few-shot link prediction for emerging knowledge graph,” Journal of Computer Science and Technology , vol. 39, no. 5, pp. 1058–1077, 2024

    2024

  7. [15]

    Heterogeneous graph trans- former,

    Z. Hu, Y . Dong, K. Wang, and Y . Sun, “Heterogeneous graph trans- former,” in WWW ’20: The Web Conference 2020, 2020, pp. 2704–2710

    2020

  8. [16]

    Modeling relational data with graph convolutional networks,

    M. S. Schlichtkrull, T. N. Kipf, P. Bloem, R. van den Berg, I. Titov, and M. Welling, “Modeling relational data with graph convolutional networks,” in The Semantic Web - 15th International Conference, ESWC 2018, vol. 10843, 2018, pp. 593–607

    2018

  9. [17]

    Scalable graph neural networks for heterogeneous graphs,

    L. Yu, J. Shen, J. Li, and A. Lerer, “Scalable graph neural networks for heterogeneous graphs,” arXiv preprint arXiv:2011.09679 , 2020

    2011 arXiv

  10. [18]

    Simple and efficient het- erogeneous graph neural network,

    X. Yang, M. Yan, S. Pan, X. Ye, and D. Fan, “Simple and efficient het- erogeneous graph neural network,” in Thirty-Seventh AAAI Conference on Artificial Intelligence, 2023. , 2023, pp. 10 816–10 824

    2023

  11. [19]

    Multi-stage self-supervised learning for graph convolutional networks on graphs with few labeled nodes,

    K. Sun, Z. Lin, and Z. Zhu, “Multi-stage self-supervised learning for graph convolutional networks on graphs with few labeled nodes,” in The Thirty-Fourth AAAI Conference on Artificial Intelligence, AAAI 2020 , 2020, pp. 5892–5899

    2020

  12. [20]

    Graph attention mlp with reliable label utilization,

    W. Zhang, Z. Yin, Z. Sheng, W. Ouyang, X. Li, Y . Tao, Z. Yang, and B. Cui, “Graph attention mlp with reliable label utilization,” arXiv preprint arXiv:2108.10097, 2021

    2021 arXiv

  13. [21]

    Random search and reproducibility for neural architecture search,

    L. Li and A. Talwalkar, “Random search and reproducibility for neural architecture search,” in Proceedings of the Thirty-Fifth Conference on Uncertainty in Artificial Intelligence, UAI 2019, Tel Aviv, Israel, July 22-25, 2019, vol. 115, 2019, pp. 367–377

    2019

  14. [22]

    Heterogeneous graph sparsification for efficient representation learning,

    C. Chunduru, C. J. Zhu, B. Gains, and J. Bi, “Heterogeneous graph sparsification for efficient representation learning,” inIEEE International Conference on Bioinformatics and Biomedicine, BIBM 2022, Las Vegas, NV , USA, December 6-8, 2022, pp. 1891–1896

    2022

  15. [23]

    Relation structure- aware heterogeneous graph neural network,

    S. Zhu, C. Zhou, S. Pan, X. Zhu, and B. Wang, “Relation structure- aware heterogeneous graph neural network,” in 2019 IEEE International Conference on Data Mining, ICDM 2019 , 2019, pp. 1534–1539

    2019

  16. [24]

    Graph condensation for graph neural networks,

    W. Jin, L. Zhao, S. Zhang, Y . Liu, J. Tang, and N. Shah, “Graph condensation for graph neural networks,” in The Tenth International Conference on Learning Representations, ICLR 2022, Virtual Event, April 25-29, 2022 , 2022

    2022

  17. [25]

    Graph condensation: A survey,

    X. Gao, J. Yu, W. Jiang, T. Chen, W. Zhang, and H. Yin, “Graph condensation: A survey,” arXiv preprint arXiv:2401.11720 , 2024

    2024 arXiv

  18. [26]

    Heterogeneous graph condensation,

    J. Gao, J. Wu, and J. Ding, “Heterogeneous graph condensation,” IEEE Trans. Knowl. Data Eng. , vol. 36, no. 7, pp. 3126–3138, 2024

    2024

  19. [27]

    Arnetminer: extraction and mining of academic social networks,

    J. Tang, J. Zhang, L. Yao, J. Li, L. Zhang, and Z. Su, “Arnetminer: extraction and mining of academic social networks,” in Proceedings of the 14th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Las Vegas, Nevada, USA, August 24-27, 2008, 2008, pp. 990–998

    2008

  20. [28]

    Freebase: a collaboratively created graph database for structuring human knowledge,

    K. D. Bollacker, C. Evans, P. K. Paritosh, T. Sturge, and J. Taylor, “Freebase: a collaboratively created graph database for structuring human knowledge,” in Proceedings of the ACM SIGMOD International Conference on Management of Data , 2008, pp. 1247–1250

    2008

  21. [29]

    metapath2vec: Scalable rep- resentation learning for heterogeneous networks,

    Y . Dong, N. V . Chawla, and A. Swami, “metapath2vec: Scalable rep- resentation learning for heterogeneous networks,” in Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , 2017, pp. 135–144

    2017

  22. [30]

    MAGNN: metapath aggregated graph neural network for heterogeneous graph embedding,

    X. Fu, J. Zhang, Z. Meng, and I. King, “MAGNN: metapath aggregated graph neural network for heterogeneous graph embedding,” in WWW ’20: The Web Conference 2020 , 2020, pp. 2331–2341

    2020

  23. [31]

    A comprehensive survey on graph reduction: Sparsification, coarsening, and condensation,

    M. Hashemi, S. Gong, J. Ni, W. Fan, B. A. Prakash, and W. Jin, “A comprehensive survey on graph reduction: Sparsification, coarsening, and condensation,” CoRR, vol. abs/2402.03358, 2024

    2024 arXiv

  24. [32]

    Spectral sparsification of graphs,

    D. A. Spielman and S. Teng, “Spectral sparsification of graphs,” SIAM J. Comput., vol. 40, no. 4, pp. 981–1025, 2011

    2011

  25. [33]

    A unified lottery ticket hypothesis for graph neural networks,

    T. Chen, Y . Sui, X. Chen, A. Zhang, and Z. Wang, “A unified lottery ticket hypothesis for graph neural networks,” in Proceedings of the 38th International Conference on Machine Learning, ICML 2021, 18-24 July 2021, Virtual Event , ser. Proceedings of Machine Learning Research...

    2021

  26. [34]

    A novel graph oversampling framework for node classification in class-imbalanced graphs,

    R. Xia, C. Zhang, Y . Zhang, X. Liu, and B. Yang, “A novel graph oversampling framework for node classification in class-imbalanced graphs,” Science China Information Sciences , vol. 67, no. 6, pp. 1–16, 2024

    2024

  27. [35]

    Herding dynamical weights to learn,

    M. Welling, “Herding dynamical weights to learn,” in Proceedings of the 26th Annual International Conference on Machine Learning, ICML 2009, Montreal, Quebec, Canada, June 14-18, 2009 , ser. ACM International Conference Proceeding Series, vol. 382, 2009, pp. 1121– 1128

    2009

  28. [36]

    Facility location: concepts, models, algorithms and case studies. series: Contributions to management science,

    G. W. Wolf, “Facility location: concepts, models, algorithms and case studies. series: Contributions to management science,” Int. J. Geogr. Inf. Sci., vol. 25, no. 2, pp. 331–333, 2011

    2011

  29. [37]

    Active learning for convolutional neural networks: A core-set approach,

    O. Sener and S. Savarese, “Active learning for convolutional neural networks: A core-set approach,” in 6th International Conference on Learning Representations, ICLR 2018, Vancouver, BC, Canada, April 30 - May 3, 2018, Conference Track Proceedings , 2018

    2018

  30. [38]

    Scaling up graph neural networks via graph coarsening,

    Z. Huang, S. Zhang, C. Xi, T. Liu, and M. Zhou, “Scaling up graph neural networks via graph coarsening,” in KDD ’21: The 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, Virtual Event, Singapore, August 14-18, 2021 , 2021, pp. 675–684

    2021

  31. [39]

    Condensing graphs via one-step gradient matching,

    W. Jin, X. Tang, H. Jiang, Z. Li, D. Zhang, J. Tang, and B. Yin, “Condensing graphs via one-step gradient matching,” in KDD ’22: The 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2022, pp. 720–730

    2022

  32. [40]

    Structure-free graph condensation: From large-scale graphs to con- densed graph-free data,

    X. Zheng, M. Zhang, C. Chen, Q. V . H. Nguyen, X. Zhu, and S. Pan, “Structure-free graph condensation: From large-scale graphs to con- densed graph-free data,” in Advances in Neural Information Processing Systems 36: Annual Conference on Neural Information Processing Systems, 2023

    2023

  33. [41]

    Graph condensation via eigenbasis match- ing,

    Y . Liu, D. Bo, and C. Shi, “Graph condensation via eigenbasis match- ing,” CoRR, vol. abs/2310.09202, 2023

    2023 arXiv

  34. [42]

    Graph-skeleton: ˜1% nodes are sufficient to represent billion-scale graph,

    L. Cao, H. Deng, Y . Yang, C. Wang, and L. Chen, “Graph-skeleton: ˜1% nodes are sufficient to represent billion-scale graph,” in Proceedings of the ACM on Web Conference 2024, WWW , 2024, pp. 570–581

    2024

  35. [43]

    Graph condensation for open-world graph learning,

    X. Gao, T. Chen, W. Zhang, Y . Li, X. Sun, and H. Yin, “Graph condensation for open-world graph learning,” in Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , 2024, pp. 851–862

    2024

  36. [44]

    Rethinking and accelerating graph condensation: A training- free approach with class partition,

    X. Gao, T. Chen, W. Zhang, J. Yu, G. Ye, Q. V . H. Nguyen, and H. Yin, “Rethinking and accelerating graph condensation: A training- free approach with class partition,” arXiv preprint arXiv:2405.13707 , 2024

    2024 arXiv

  37. [45]

    Robgc: Towards robust graph condensation,

    X. Gao, H. Yin, T. Chen, G. Ye, W. Zhang, and B. Cui, “Robgc: Towards robust graph condensation,” arXiv preprint arXiv:2406.13200 , 2024

    2024 arXiv

  38. [46]

    Graph condensation via receptive field distribution matching,

    M. Liu, S. Li, X. Chen, and L. Song, “Graph condensation via receptive field distribution matching,” CoRR, vol. abs/2206.13697, 2022

    2022 arXiv

  39. [47]

    Graph distillation with eigenbasis matching,

    L. Yang, D. Bo, and C. Shi, “Graph distillation with eigenbasis matching,” in Proceedings of the International Conference on Machine Learning., 2024

    2024

  40. [48]

    Maximizing the spread of influence through a social network,

    D. Kempe, J. M. Kleinberg, and ´E. Tardos, “Maximizing the spread of influence through a social network,” in Proceedings of the Ninth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. ACM, 2003, pp. 137–146

    2003

  41. [49]

    Maximizing the spread of influence through a social network,

    ——, “Maximizing the spread of influence through a social network,” in Proceedings of the Ninth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Washington, DC, USA, August 24 - 27, 2003 , 2003, pp. 137–146

    2003

  42. [50]

    An analysis of approximations for maximizing submodular set functions - I,

    G. L. Nemhauser, L. A. Wolsey, and M. L. Fisher, “An analysis of approximations for maximizing submodular set functions - I,” Math. Program., vol. 14, no. 1, pp. 265–294, 1978

    1978

  43. [51]

    Scalable and parallelizable influence maximization with random walk ranking and rank merge pruning,

    S. Kim, D. Kim, J. Oh, J. Hwang, W. Han, W. Chen, and H. Yu, “Scalable and parallelizable influence maximization with random walk ranking and rank merge pruning,” Inf. Sci., vol. 415, pp. 171–189, 2017

    2017

  44. [52]

    Comparing sets of patterns with the jaccard index,

    S. Fletcher and M. Z. Islam, “Comparing sets of patterns with the jaccard index,” Australas. J. Inf. Syst. , vol. 22, 2018

    2018

  45. [53]

    The lov ´asz hinge: A novel convex surrogate for submodular losses,

    J. Yu and M. B. Blaschko, “The lov ´asz hinge: A novel convex surrogate for submodular losses,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 42, no. 3, pp. 735–748, 2020

    2020

  46. [54]

    The lov ´asz-softmax loss: A tractable surrogate for the optimization of the intersection-over-union measure in neural networks,

    M. Berman, A. R. Triki, and M. B. Blaschko, “The lov ´asz-softmax loss: A tractable surrogate for the optimization of the intersection-over-union measure in neural networks,” in 2018 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2018 , 2018, pp. 4413–4421

    2018

  47. [55]

    A class of submodular functions for document summarization,

    H. Lin and J. A. Bilmes, “A class of submodular functions for document summarization,” in The 49th Annual Meeting of the Association for Com- putational Linguistics: Human Language Technologies, Proceedings of the Conference, 2011, pp. 510–520

    2011

  48. [56]

    Scaling graph neural networks with approximate pagerank,

    A. Bojchevski, J. Klicpera, B. Perozzi, A. Kapoor, M. Blais, B. R ´ozemberczki, M. Lukasik, and S. G ¨unnemann, “Scaling graph neural networks with approximate pagerank,” in KDD ’20: The 26th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, Virtual Event, CA, USA,...

    2020

  49. [57]

    Degree centrality, betweenness centrality, and closeness centrality in social network,

    J. Zhang and Y . Luo, “Degree centrality, betweenness centrality, and closeness centrality in social network,” in 2017 2nd international confer- ence on modelling, simulation and applied mathematics (MSAM2017) , 2017, pp. 300–303

    2017

  50. [58]

    Hubs, authorities, and communities,

    J. M. Kleinberg, “Hubs, authorities, and communities,” ACM Comput. Surv., vol. 31, no. 4es, p. 5, 1999

    1999

  51. [59]

    Deep graph library: A graph-centric, highly-performant package for graph neural networks,

    M. Wang, D. Zheng, Z. Ye, Q. Gan, M. Li, X. Song, J. Zhou, C. Ma, L. Yu, Y . Gai, T. Xiao, T. He, G. Karypis, J. Li, and Z. Zhang, “Deep graph library: A graph-centric, highly-performant package for graph neural networks,” arXiv preprint arXiv:1909.01315 , 2019

    1909 arXiv

  52. [60]

    A collection of benchmark datasets for systematic evaluations of machine learning on the semantic web,

    P. Ristoski, G. K. D. de Vries, and H. Paulheim, “A collection of benchmark datasets for systematic evaluations of machine learning on the semantic web,” in The Semantic Web - ISWC 2016 - 15th International Semantic Web Conference, ser. Lecture Notes in Computer Science, vol. ...

    2016

  53. [61]

    Visualizing data using t-sne

    L. Van der Maaten and G. Hinton, “Visualizing data using t-sne.” Journal of machine learning research , vol. 9, no. 11, 2008

    2008

Pith tools

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