Pith. sign in

REVIEW 3 major objections 5 minor 34 references

Self-supervised Subgraph Neural Network With Deep Reinforcement Walk Exploration

T0 review · 3 major / 5 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read A walk-based Markov decision process can generate the same connected substructures as subgraph expansion while cutting candidate cost from quadratic to linear.

desk verdict A genuine technical idea—walk-endpoint action space for subgraph sampling—sits under an overclaimed equivalence theorem that only holds for unbounded walks, not the finite L used in practice. read the letter →

arxiv 2502.01809 v1 pith:42STFF7L submitted 2025-02-03 cs.LG

Jianming Huang , Hiroyuki Kasai This is my paper · ORCID
classification cs.LG MSC 68T0768R10
keywords subgraphneuralnetworksgraphreinforcementlearningwalkexplorationself-supervisedclassificationexplainabilitydeepQ-networks
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 proposes a self-supervised framework, RWE-SGNN, that trains a subgraph neural network by having a reinforcement-learning agent walk through the input graph rather than expand subgraphs step by step. The central claim is that a walk-based Markov decision process has equivalent generation capability to the traditional subgraph-generation process, so the agent can extract the same important substructures while evaluating far fewer candidates. If the claim holds, graph classifiers gain a data-driven sampling policy that is both cheaper than algorithm-based subgraph sampling and inherently interpretable: the learned walks mark the substructures that drive predictions. Experiments on seven graph classification benchmarks report top-2 accuracy everywhere and top-1 on four, with extracted walks localizing known chemical groups such as NO2 in MUTAG and circle and house motifs in BA-2motifs.

What carries the argument

The load-bearing object is the walk-exploration-based MDP (Definition 3), whose state space is random walks of length at most $L$ on the input graph, whose actions are neighbors of the current endpoint, and whose transition appends a node to the walk. It is paired with Theorem 1, which proves that every connected subgraph has a walk that visits all of its nodes, giving the walk process the same generation capability as the subgraph-based MDP (Definition 2) while reducing the feasible action space from $\bigcup_{v \in V_s} N(v) \setminus V_s$ to $N(v_l^s)$. A deep Q-network trained with the downstream loss as reward implements the walk policy, and a shallow MPNN provides node-level graphlet-aware embeddings that guide the agent.

What would settle it

Construct a graph whose classification label is determined by a star-shaped substructure with $n$ leaves, set the walk length $L$ to a value smaller than $n+1$, and check whether the extracted walk can visit all leaves. If the walk cannot cover the full star while the subgraph-based MDP can, the claimed equivalence between walk exploration and subgraph generation does not hold under the bounded-length setting actually used.

Watch

Extended reading notes

Core claim

RWE-SGNN replaces the breadth-first subgraph-generation MDP of earlier subgraph neural networks with a depth-first walk-exploration MDP. States are walk sequences of bounded length, actions are neighbors of the current endpoint, and the reward is the reduction in downstream loss obtained by extending the walk. Theorem 1 shows that for any connected subgraph there exists a walk visiting all its nodes, so walk exploration can, in principle, generate the same connected subgraphs as the traditional node-by-node expansion. The candidate count per step drops from the union of border neighborhoods to a single node neighborhood, changing the generation complexity from quadratic to linear in the structure size. The sampling model first uses a shallow MPNN to compute graphlet-aware embeddings, then a deep Q-network chooses walk steps, and a two-stage training loop alternates between optimizing the output model on downstream loss and using that loss as the reward signal for the agent. On the seven datasets, the walk-based MDP outperforms the subgraph-based MDP on six, and the extracted walks match known domain motifs in the visualization experiments.

Load-bearing premise

The method works only if a fixed maximum walk length $L$ (set to 8 or 16 in the experiments) is long enough for a walk to cover every important substructure, which fails for star-shaped substructures where backtracking requires more steps than the number of nodes.

Editorial extensions

If this is right

  • Walk-based sampling can replace subgraph generation in SGNNs, lowering the candidate evaluation cost from $O(N^2 D)$ to $O(L D)$ without losing access to the same connected substructures.
  • The same framework yields explanations: the learned walk sequences identify the substructures that determine predictions, as shown by the NO2 and motif visualizations.
  • The two-stage training loop lets the downstream loss act as the reward, so the sampling policy adapts to the task instead of relying on a fixed algorithm-based policy.
  • Because the method is self-supervised and task-driven, it can be ported to other downstream objectives by swapping the loss function in the reward definition.
  • Performance on seven benchmarks suggests the approach transfers across molecule and social-network domains, reaching top-2 accuracy on all and top-1 on four.

Reading between the lines

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

  • The equivalence theorem says nothing about embedding fidelity: a walk that visits all nodes of a subgraph does not necessarily preserve the subgraph's edge structure in the sequence encoder, so tasks where edge topology matters may need a walk-to-subgraph reconstruction step.
  • The backtracking difficulty for star graphs suggests a natural extension: an adaptive or learned stopping criterion for walk length per graph, instead of a global $L$.
  • The reward difference $l(O(s)) - l(O(s'))$ rewards local improvements; a credit-assignment variant that rewards whole-trajectory outcomes could make the agent more stable on larger graphs.
  • The walk framework could transfer to link prediction or graph generation, where sequential construction of substructures is a core operation and the linear action space would be a direct benefit.
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. This manuscript proposes RWE-SGNN, a self-supervised subgraph neural network in which a reinforcement learning agent generates substructures as bounded-length walk sequences rather than by iterative subgraph expansion. The sampling model combines a shallow MPNN graphlet encoder with a DQN-based walk-exploration MDP, and a two-stage training loop alternates between fitting the output model on the downstream loss and updating the agent using that loss as a reward. The authors claim that walk exploration has generation capability equivalent to subgraph generation, that it reduces candidate-action complexity from quadratic to linear, and that experiments on seven graph benchmarks show top-2 accuracy and explainable substructures.

Significance. The proposed framework is timely, and the same-framework comparison between subgraph-based and walk-based MDPs, together with the ablation over trajectory length and sample number, is a useful empirical contribution. If the equivalence and complexity claims could be made precise, the method would offer a practical trade-off between expressiveness and efficiency. The proof of Theorem 1 is correct, and the empirical study includes a meaningful internal baseline, but the theoretical equivalence is currently overstated and is the main load-bearing weakness; the experimental results alone do not justify the formal claims in the abstract.

major comments (3)
  1. [Section 3.2, Definition 3 and Theorem 1] The equivalence claim is not supported by Theorem 1 because the theorem concerns the complete random walk set of arbitrary length, while Definition 3 restricts states to walks of length at most L. For a star subgraph with center and k leaves, any walk visiting all k+1 nodes has length at least 2k-1; with the reported L=16, the walk MDP cannot represent a 10-node star (k=9) even though the subgraph-generation MDP can produce it in 10 node additions. Therefore the claimed equivalent generation capability is false under the finite-L definition actually used, and this directly affects the L in {8,16} settings of Section 4.1.
  2. [Section 3.2, Theorem 1 and the following paragraph] Even for unbounded walks, the theorem only shows that some walk visits every vertex; it does not show that the walk MDP can generate every connected subgraph as a state, that the edge set of the subgraph is represented, or that optimal policies under the two reward definitions coincide. A walk visiting the three vertices of a triangle can miss one of the triangle's edges unless it is lengthened, so the object encoded by the sequence encoder E_w is not automatically the same object generated by the subgraph MDP. The sentence after Theorem 1 that walk exploration has equivalent subgraph generation capability therefore overstates what has been proved.
  3. [Section 3.5, complexity comparison] The claimed reduction from O(N^2 D) to O(L D) treats L as an independent constant, but to cover a connected substructure of N nodes the walk length L must scale with N (for a star with N=k+1 nodes the minimum visiting walk has length 2k-1). The experiments fix L in {8,16} without evidence that all informative substructures fit within that bound, and no wall-clock or memory measurements are reported, so the practical efficiency advantage is not fully demonstrated.
minor comments (5)
  1. [Algorithm 1, line 29] The output-model parameter update minimizes lQ instead of lO; this appears to be a typographical error in the pseudo-code.
  2. [Algorithm 2] The inference procedure computes the Q-learning loss lQ although it is unused, and line 11 uses Z' without computing it in that algorithm; the pseudo-code should compute the output-model embeddings and remove the Q-loss computation.
  3. [Section 4.2, ablation paragraph] The text says that Figure 6 shows the test accuracy curves of different hyperparameters, but the accuracy curves appear in Figure 4 while Figure 6 is the BA-2motifs visualization; the cross-reference should be corrected.
  4. [Throughout] There are several typographical and formatting errors, including Defintion, trajctory, architechture, enchance, and the inconsistent spacing R WE-SGNN; a careful proofread is needed.
  5. [Table 1] The caption states that top-2 results are marked in bold, but the bolding rule is applied unevenly (e.g., SUGAR NoMI has no reported standard deviation), and the criterion should be stated precisely.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation chain; the central equivalence claim rests on an independent graph-theoretic theorem, and the finite-length gap is a correctness limitation rather than a circular reduction.

full rationale

I find no circularity in this paper's derivation chain. Theorem 1 is an independent graph-theoretic statement: for any connected subgraph, a walk of sufficiently long (unbounded) length can visit all its vertices, and the proof constructs such a walk by concatenating pairwise connecting walks; it does not assume the equivalence it is used to justify. The walk-exploration MDP in Definition 3 is defined independently of the subgraph-generation MDP in Definition 2, and the claimed equivalence is then supported by Theorem 1 rather than by definitional identity. The empirical section does not fit a parameter to a target quantity and then report that quantity as a prediction; the reported accuracies come from implemented experiments with conventional hyperparameters, and the SubgraphMDP versus WalkMDP comparison is a real comparison of two defined variants. The self-supervised reward loop, in which the sampling model is trained using the output model's loss, is a training design and not a logical dependency that makes the output an input to itself. There are no load-bearing self-citations: the authors cite related work for context, but the central theorem and the experimental claims are not reduced to those citations. The finite-trajectory-length issue noted by the skeptic is a genuine correctness limitation of the equivalence claim when L is small, but it is a gap in the proof or a condition on the theorem's applicability, not a circular step in which a quantity is defined in terms of itself or a fitted value is renamed as a prediction.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new entities (particles, forces, etc.). Its central claims rest on standard graph theory, RL assumptions, and tunable hyperparameters (L, K, gamma, beta, epsilon), plus a domain assumption that important substructures are connected and can be represented by walks of bounded length. The theorem itself is a known fact, so the main contribution is the engineering combination of these pieces into a self-supervised SGNN.

free parameters (5)
  • maximum walk length L = 8 or 16 (tuned per dataset)
    Controls the maximum length of the generated walk; central to the claimed equivalence and complexity reduction. Tuned in Section 4.1 without a rigorous justification for sufficiency.
  • number of substructures K = 3, 16, or 32 (tuned per dataset)
    Number of walks sampled per graph; directly affects the bag-of-substructures representation and downstream accuracy. Tuned in Section 4.1.
  • discount factor gamma = 0.9
    Standard RL hyperparameter for DQN, set in Section 4.1.
  • target network smoothing beta = 0.1
    Polyak averaging coefficient for the target Q-network, set in Section 4.1.
  • epsilon-greedy schedule = increases from 0.1 to 0.4 over epochs
    Exploration schedule for the policy, set in Section 4.1.
assumptions (4)
  • domain assumption Substructures of interest are connected subgraphs
    The subgraph-generation-based MDP restricts the state space to connected subgraphs (Definition 2), and this restriction is carried over to the walk-based MDP. The paper argues this improves interpretability and encoding, but it excludes disconnected but potentially informative substructures.
  • standard math Any connected graph has a walk visiting all its vertices
    Used in the proof of Theorem 1 (Section 3.2). This is an elementary graph-theory fact; the proof concatenates paths between consecutive vertices in an arbitrary ordering.
  • standard math Q-learning with a function approximator converges to a useful policy
    The DQN training (Section 3.3) relies on the Bellman equation and off-policy Q-learning, whose convergence is assumed without discussion of approximation error or nonstationarity from the two-stage training.
  • domain assumption A 3-layer GIN produces graphlet-aware node embeddings that guide the walk agent
    The low-level graphlet perceptron uses a 3-layer GIN (Section 4.1) to detect graphlet patterns; the paper does not provide evidence that this architecture actually captures the needed graphlet topology for the agent.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Self-supervised Subgraph Neural Network With Deep Reinforcement Walk Exploration." pith.science (2026). https://pith.science/paper/42STFF7L

@misc{pith2026250201809,
  author       = {Pith},
  title        = {Pith review of: Self-supervised Subgraph Neural Network With Deep Reinforcement Walk Exploration},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/42STFF7L}},
  note         = {Machine review of arXiv:2502.01809}
}
read the original abstract

Graph data, with its structurally variable nature, represents complex real-world phenomena like chemical compounds, protein structures, and social networks. Traditional Graph Neural Networks (GNNs) primarily utilize the message-passing mechanism, but their expressive power is limited and their prediction lacks explainability. To address these limitations, researchers have focused on graph substructures. Subgraph neural networks (SGNNs) and GNN explainers have emerged as potential solutions, but each has its limitations. SGNNs computes graph representations based on the bags of subgraphs to enhance the expressive power. However, they often rely on predefined algorithm-based sampling strategies, which is inefficient. GNN explainers adopt data-driven approaches to generate important subgraphs to provide explanation. Nevertheless, their explanation is difficult to be translated into practical improvements on GNNs. To overcome these issues, we propose a novel self-supervised framework that integrates SGNNs with the generation approach of GNN explainers, named the Reinforcement Walk Exploration SGNN (RWE-SGNN). Our approach features a sampling model trained in an explainer fashion, optimizing subgraphs to enhance model performance. To achieve a data-driven sampling approach, unlike traditional subgraph generation approaches, we propose a novel walk exploration process, which efficiently extracts important substructures, simplifying the embedding process and avoiding isomorphism problems. Moreover, we prove that our proposed walk exploration process has equivalent generation capability to the traditional subgraph generation process. Experimental results on various graph datasets validate the effectiveness of our proposed method, demonstrating significant improvements in performance and precision.

Figures

Figures reproduced from arXiv: 2502.01809 by the authors.

Figure 1
Figure 1. Illustration of the self-supervised framework, where solid arrow represents the direction [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Illustration of the sampling model of RWE-SGNN, which shows the process of extracting [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Illustration of the model architecture, which comprises two main components: the sam [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Test accuracy of walk-based MDP for varying trajectory lengths and varying sample [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]
Figure 5
Figure 5. Figure 5: The extracted subgraphs at different epochs in MUTAG dataset. The nodes in orange [PITH_FULL_IMAGE:figures/full_fig_p014_5.png]
Figure 6
Figure 6. Figure 6: The extracted subgraphs at different epochs in BA-2motifs dataset. The nodes in orange [PITH_FULL_IMAGE:figures/full_fig_p015_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 23 canonical work pages

  1. [1]

    Subgraph neural net- works

    Emily Alsentzer, Samuel Finlayson, Michelle Li, and Marinka Zitnik. Subgraph neural net- works. Advances in Neural Information Processing Systems, 33:8017–8029, 2020

    work page 2020

  2. [2]

    Efficient subgraph gnns by learning effective selection policies

    Beatrice Bevilacqua, Moshe Eliasof, Eli Meirom, Bruno Ribeiro, and Haggai Maron. Efficient subgraph gnns by learning effective selection policies. arXiv preprint arXiv:2310.20082, 2023

    arXiv 2023

  3. [3]

    Equivariant subgraph ag- gregation networks

    Beatrice Bevilacqua, Fabrizio Frasca, Derek Lim, Balasubramaniam Srinivasan, Chen Cai, Gopinath Balamurugan, Michael M Bronstein, and Haggai Maron. Equivariant subgraph ag- gregation networks. arXiv preprint arXiv:2110.02910, 2021

    arXiv

  4. [4]

    Protein function prediction via graph kernels

    Karsten M Borgwardt, Cheng Soon Ong, Stefan Sch¨ onauer, SVN Vishwanathan, Alex J Smola, and Hans-Peter Kriegel. Protein function prediction via graph kernels. Bioinformatics, 21(suppl 1):i47–i56, 2005

    work page 2005

  5. [5]

    Structure-activity relationship of mutagenic aromatic and heteroaromatic nitro compounds

    Asim Kumar Debnath, Rosa L Lopez de Compadre, Gargi Debnath, Alan J Shusterman, and Corwin Hansch. Structure-activity relationship of mutagenic aromatic and heteroaromatic nitro compounds. correlation with molecular orbital energies and hydrophobicity. Journal of medicinal chemistry, 34(2):786–797, 1991

    work page 1991

  6. [6]

    Fast graph representation learning with pytorch geometric

    Matthias Fey and Jan Eric Lenssen. Fast graph representation learning with pytorch geometric. arXiv preprint arXiv:1903.02428, 2019

    arXiv 1903

  7. [7]

    Inductive representation learning on large graphs

    Will Hamilton, Zhitao Ying, and Jure Leskovec. Inductive representation learning on large graphs. Advances in neural information processing systems, 30, 2017

    2017

  8. [8]

    King, Stefan Kramer, and Ashwin Srinivasan

    Christoph Helma, Ross D. King, Stefan Kramer, and Ashwin Srinivasan. The predictive toxi- cology challenge 2000–2001. Bioinformatics, 17(1):107–108, 2001

    work page 2000

Show all 34 references
  1. [9]

    Finding frequent subgraphs in longitudinal social network data using a weighted graph mining approach

    Chuntao Jiang, Frans Coenen, and Michele Zito. Finding frequent subgraphs in longitudinal social network data using a weighted graph mining approach. In Advanced Data Mining and Applications: 6th International Conference, ADMA 2010, Chongqing, China, November 19-21, 2010, Proc...

    2010

  2. [10]

    Variational graph auto-encoders

    Thomas N Kipf and Max Welling. Variational graph auto-encoders. arXiv preprint arXiv:1611.07308, 2016

    2016 arXiv

  3. [11]

    Kipf and Max Welling

    Thomas N. Kipf and Max Welling. Semi-supervised classification with graph convolutional networks. arXiv preprint: arXiv:1609.02907, 2017

    2017 arXiv

  4. [12]

    Orphicx: A causality-inspired latent variable model for interpreting graph neural networks

    Wanyu Lin, Hao Lan, Hao Wang, and Baochun Li. Orphicx: A causality-inspired latent variable model for interpreting graph neural networks. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 13729–13738, 2022

    2022

  5. [13]

    Parameterized explainer for graph neural network

    Dongsheng Luo, Wei Cheng, Dongkuan Xu, Wenchao Yu, Bo Zong, Haifeng Chen, and Xiang Zhang. Parameterized explainer for graph neural network. Advances in neural information processing systems, 33:19620–19631, 2020

    2020

  6. [14]

    Provably powerful graph networks

    Haggai Maron, Heli Ben-Hamu, Hadar Serviansky, and Yaron Lipman. Provably powerful graph networks. Advances in neural information processing systems, 32, 2019. 18

    2019

  7. [15]

    Playing atari with deep reinforcement learning

    Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Alex Graves, Ioannis Antonoglou, Daan Wierstra, and Martin Riedmiller. Playing atari with deep reinforcement learning. arXiv preprint arXiv:1312.5602, 2013

    2013 arXiv

  8. [16]

    Kriege, Franka Bause, Kristian Kersting, Petra Mutzel, and Mar- ion Neumann

    Christopher Morris, Nils M. Kriege, Franka Bause, Kristian Kersting, Petra Mutzel, and Mar- ion Neumann. Tudataset: A collection of benchmark datasets for learning with graphs. In International Conference on Machine Learning (ICML), 2020

    2020

  9. [17]

    Weisfeiler and leman go neural: Higher-order graph neural networks

    Christopher Morris, Martin Ritzert, Matthias Fey, William L Hamilton, Jan Eric Lenssen, Gaurav Rattan, and Martin Grohe. Weisfeiler and leman go neural: Higher-order graph neural networks. In Proceedings of the AAAI conference on artificial intelligence, volume 33, pages 4602–...

    2019

  10. [18]

    Biological network comparison using graphlet degree distribution

    Nataˇ sa Prˇ zulj. Biological network comparison using graphlet degree distribution. Bioinformatics, 23(2):e177–e183, 2007

    2007

  11. [19]

    Reinforcement learning enhanced explainer for graph neural networks

    Caihua Shan, Yifei Shen, Yao Zhang, Xiang Li, and Dongsheng Li. Reinforcement learning enhanced explainer for graph neural networks. Advances in Neural Information Processing Systems, 34:22523–22533, 2021

    2021

  12. [20]

    Shervashidze, P

    N. Shervashidze, P. Schweitzer, E. J. v Leeuwen, K. Mehlhorn, and K. M. Borgwardt. Weisfeiler- lehman graph kernels. The Journal of Machine Learning Research, 12:2539–2561, 2011

    2011

  13. [21]

    Sugar: Subgraph neural network with reinforcement pooling and self-supervised mutual information mechanism

    Qingyun Sun, Jianxin Li, Hao Peng, Jia Wu, Yuanxing Ning, Philip S Yu, and Lifang He. Sugar: Subgraph neural network with reinforcement pooling and self-supervised mutual information mechanism. In Proceedings of the Web Conference 2021, pages 2081–2091, 2021

    2021

  14. [22]

    Spline-fitting with a genetic algorithm: A method for developing classification structure-activity relationships

    Jeffrey J Sutherland, Lee A O’brien, and Donald F Weaver. Spline-fitting with a genetic algorithm: A method for developing classification structure-activity relationships. Journal of chemical information and computer sciences, 43(6):1906–1915, 2003

    1906

  15. [23]

    Optimal transport for structured data with application on graphs

    Vayer Titouan, Nicolas Courty, Romain Tavenard, and R´ emi Flamary. Optimal transport for structured data with application on graphs. In International Conference on Machine Learning (ICML), 2019

    2019

  16. [24]

    Wasserstein weisfeiler-lehman graph kernels

    Matteo Togninalli, Elisabetta Ghisu, Felipe Llinares-L´ opez, Bastian Rieck, and Karsten Borg- wardt. Wasserstein weisfeiler-lehman graph kernels. In Advances in Neural Information Processing Systems (NIPS), 2019

    2019

  17. [25]

    Graphopt: Learning optimization models of graph formation

    Rakshit Trivedi, Jiachen Yang, and Hongyuan Zha. Graphopt: Learning optimization models of graph formation. In International Conference on Machine Learning, pages 9603–9613. PMLR, 2020

    2020

  18. [26]

    Comparison of descriptor spaces for chemical compound retrieval and classification

    Nikil Wale, Ian A Watson, and George Karypis. Comparison of descriptor spaces for chemical compound retrieval and classification. Knowledge and Information Systems, 14:347–375, 2008

    2008

  19. [27]

    Dynamic graph cnn for learning on point clouds

    Yue Wang, Yongbin Sun, Ziwei Liu, Sanjay E Sarma, Michael M Bronstein, and Justin M Solomon. Dynamic graph cnn for learning on point clouds. ACM Transactions on Graphics (tog), 38(5):1–12, 2019

    2019

  20. [28]

    Q-learning

    Christopher JCH Watkins and Peter Dayan. Q-learning. Machine learning, 8:279–292, 1992. 19

    1992

  21. [29]

    A new perspective on” how graph neural networks go beyond weisfeiler-lehman?”

    Asiri Wijesinghe and Qing Wang. A new perspective on” how graph neural networks go beyond weisfeiler-lehman?”. In International Conference on Learning Representations, 2021

    2021

  22. [30]

    How powerful are graph neural networks? In International Conference on Learning Representations (ICLR), 2019

    Keyulu Xu, Weihua Hu, Jure Leskovec, and Stefanie Jegelka. How powerful are graph neural networks? In International Conference on Learning Representations (ICLR), 2019

    2019

  23. [31]

    Deep graph kernels

    Pinar Yanardag and SVN Vishwanathan. Deep graph kernels. In ACM Special Interest Group on Knowledge Discovery and Data Mining (SIGKDD), 2015

    2015

  24. [32]

    Gnnexplainer: Generating explanations for graph neural networks

    Zhitao Ying, Dylan Bourgeois, Jiaxuan You, Marinka Zitnik, and Jure Leskovec. Gnnexplainer: Generating explanations for graph neural networks. Advances in neural information processing systems, 32, 2019

    2019

  25. [33]

    On explainability of graph neural networks via subgraph explorations

    Hao Yuan, Haiyang Yu, Jie Wang, Kang Li, and Shuiwang Ji. On explainability of graph neural networks via subgraph explorations. In International conference on machine learning, pages 12241–12252. PMLR, 2021

    2021

  26. [34]

    A survey on deep graph generation: Methods and applications

    Yanqiao Zhu, Yuanqi Du, Yinkai Wang, Yichen Xu, Jieyu Zhang, Qiang Liu, and Shu Wu. A survey on deep graph generation: Methods and applications. In Learning on Graphs Conference, pages 47–1. PMLR, 2022. 20

    2022

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.