Pith. sign in

REVIEW 2 major objections 6 minor 49 references

Enhancing Homophily-Heterophily Separation: Relation-Aware Learning in Heterogeneous Graphs

T0 review · 2 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Heterogeneous graphs can separate homophily from heterophily by learning relation importance, and this paper claims the resulting representations beat existing methods on node classification, clustering, and similarity search.

desk verdict RASH is a genuinely new mechanism with strong benchmark results, but its central homophily/heterophily separation is an unvalidated heuristic that needs a direct label-agreement check before the claim is credible. read the letter →

arxiv 2506.20980 v1 pith:RHVUU5KJ submitted 2025-06-26 cs.SI cs.AI

Ziyu Zheng , Yaming Yang , Ziyu Guan , Wei Zhao , Weigang Lu This is my paper · ORCID
classification cs.SIcs.AI
keywords HeterogeneousgraphneuralnetworksHeterophilySelf-supervisedlearningContrastiveDualhypergraphRelationimportanceNodeclassificationrepresentation
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

Heterogeneous graphs—networks with multiple node and edge types—often connect nodes that carry different labels, a property known as heterophily. Existing graph-learning methods usually flatten such graphs into homogeneous ones before trying to model heterophily, and in doing so lose the meaning carried by the different relation types. This paper tries to establish that both heterogeneity and heterophily can be handled together: a contrastive framework called RASH learns how important each heterogeneous relation is, uses that importance to split same-type node pairs into homophilic (same-label) and heterophilic (different-label) groups, and aligns the resulting views with the original graph representations. If right, it would mean that predefined meta-paths and homogeneous simplification are not needed to capture heterophily, and that one learned set of relation weights can serve multiple node types and downstream tasks. On DBLP, ACM, IMDB, and YELP, RASH reports the best node-classification, clustering, and similarity-search numbers among the compared methods.

What carries the argument

The load-bearing object is the dual heterogeneous hypergraph transform (DHHT) together with relation-aware importance scoring. In DHHT, each relation's bipartite subgraph is written as an incidence matrix whose columns are original edges, so original edges become hyperedges; hypergraph convolution then learns edge-level representations via two-stage message passing between nodes and hyperedges, capturing higher-order interaction among heterogeneous edges. A scoring function converts those representations into edge importance weights through Gumbel-Softmax sampling. Homophilic and heterophilic graphs are constructed dynamically from two-hop products of the importance weights and their complements, and a multi-relation contrastive loss aligns the low-pass filtered homophilic view, the high-pass filtered heterophilic view, and the original heterogeneous view in a shared space.

What would settle it

Train RASH on one benchmark, extract the node pairs ranked most homophilic by equation (9) and most heterophilic by equation (10), and compare the fraction of each set that actually shares ground-truth node labels; if the 'homophilic' set is not clearly more label-agreeing than the 'heterophilic' set—or no better than random edge weights—the separation mechanism is not learning homophily.

Watch

Extended reading notes

Core claim

The central claim is that heterophily in a heterogeneous graph can be learned from the graph's own relational structure rather than from a homogeneous projection of it. RASH converts each relation-specific bipartite subgraph into a dual heterogeneous hypergraph, runs hypergraph convolution to obtain edge representations, and assigns each edge a Gumbel-softmax importance weight. For a pair of same-type nodes, the two-hop product of these weights (equation 9) is treated as a homophilic edge weight, and the product of the complement weights (equation 10) as a heterophilic edge weight. Low-pass filtering on the homophilic graph and high-pass filtering on the heterophilic graph produce views that are pulled toward the heterogeneous encoder by a multi-relation InfoNCE-style contrastive loss. The paper reports that this yields state-of-the-art node classification, clustering, and similarity search on DBLP, ACM, IMDB, and YELP, and that the gains persist when node features are replaced by random vectors.

Load-bearing premise

The load-bearing premise is that along a two-hop path, the product of learned edge-importance weights marks node pairs that share a label, while the product of the complement weights marks node pairs with different labels; this label-interpretation is assumed, not verified against ground-truth edge labels.

Editorial extensions

If this is right

  • Because relation importance is learned end-to-end, no predefined meta-paths are needed to separate homophilic from heterophilic structure.
  • One set of learned importance weights yields homophilic and heterophilic views for any target node type without retraining the model for each type.
  • The dual hypergraph supplies structural signal even when node features are replaced with random vectors, so the method is not dependent on informative raw features.
  • Under random edge deletion, the reported classification accuracy degrades more slowly than the compared methods, indicating that the separated views provide redundancy against missing edges.

Reading between the lines

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

  • An untested corollary of the two-hop construction is that equation (9) could serve as a general-purpose heterophily detector for heterogeneous graphs, usable even when no labels or downstream task is specified.
  • A direct test of the mechanism would correlate learned edge importance with ground-truth label agreement; if the correlation is weak, the contrastive loss may be aligning with graph structure rather than with homophily.
  • The dual-hypergraph edge encoder is detachable and could be inserted into supervised heterogeneous graph models, not only contrastive pipelines.
  • If low importance weights mostly mark noisy or irrelevant relations rather than heterophily, equation (10) would need an explicit irrelevance term; this is the part of the design a follow-up should examine.
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

2 major / 6 minor

Summary. The paper proposes RASH, a self-supervised contrastive framework for heterogeneous graphs. It encodes relation-specific bipartite subgraphs through a dual heterogeneous hypergraph, derives edge importance weights via a Gumbel-style soft sampling (Eq. 8), and constructs homophilic and heterophilic graphs by multiplying learned importance scores along two-hop paths (Eqs. 9-10). Low- and high-pass filters extract view representations, and a multi-relation InfoNCE loss (Eqs. 13-16) aligns these views with the heterogeneous encoder output. Experiments on DBLP, ACM, IMDB, YELP, and a large-scale Aminer graph report state-of-the-art or competitive results in node classification, clustering, and similarity search, along with ablations, robustness to edge deletion, randomized-feature tests, and hyperparameter sensitivity.

Significance. If the central assumption of Eqs. 9-10 is valid, RASH is a meaningful step: it addresses heterophily in heterogeneous graphs without collapsing the graph to predefined meta-paths, and it does so in an end-to-end, self-supervised manner. The paper's strengths include public code, consistent gains across four benchmark datasets plus a 439k-node large-scale experiment, a sensible ablation suite, and robustness analyses. The significance is currently conditional, however, because the paper never tests whether the constructed homophilic and heterophilic graphs correspond to ground-truth label agreement. That validation is the difference between a principled separation mechanism and a self-generated artifact; the authors should be required to provide it.

major comments (2)
  1. [Section 4.3, Eqs. (9)-(10)] The central assumption that a^{r,ho}_{i,j} and a^{r,he}_{i,j} track ground-truth label agreement among target nodes is never validated. A direct test would rank same-type node pairs by these scores and measure whether high-ho pairs connect same-label nodes and high-he pairs connect different-label nodes more often than chance, for example by reporting label-agreement precision or AUC. The ablations in Table 4 remove whole modules (w/o Homo_CL, w/o Hete_CL, w/o RAE) but do not isolate whether the constructed graphs themselves are meaningful. If low learned importance encodes noise or irrelevance rather than heterophily, the contrastive losses in Eqs. (13)-(16) may align representations to self-generated artifacts, and the reported gains could stem from the heterogeneous encoder plus contrastive regularization rather than from genuine homophily/heterophily separation. This is load-bearing because the title and abstract attribute the performance gains to exactly this separation.
  2. [Section 4.2.2, Eq. (8) and Section 5.3] There is an internal contradiction in the treatment of the noise parameter δ. Eq. (8) defines δ ~ Uniform(0,1) and uses log δ - log(1-δ), which is a logistic (Binary Concrete) noise term, not a Gumbel variate as claimed; Section 5.3 then states that 'δ was set to 1e-4', which makes the noise a constant and removes the stochasticity that the reparameterization claim requires. Please correct the formula or the hyperparameter description, and state explicitly whether w is stochastic or deterministic. This matters for reproducibility and for the claim that the model learns discrete edge sampling via Gumbel-Max reparameterization.
minor comments (6)
  1. [Section 4.2.1, Eq. (3)] Equation (3) contains typos in the incidence-matrix definition: the relation index should appear on the entries (M^r_{i,e}, M^r_{j,e}), the condition 'iff A^r_{i,j}=1, <i,j>=e' is malformed, and the following sentence 'M^r denote the transpose matrix' should read 'M^{r\top}'.
  2. [Section 4.3, Eq. (9)] Equation (9) sums over k ∈ N^r_j, but the described two-hop path i → k under relation r and k → j under r^{-1} requires k to be in the intersection of the relevant neighbor sets; please clarify the intended index sets for both Eq. (9) and Eq. (10).
  3. [Section 5.3] The sentence 'The range of the number of positive samples was set from 0 to 5' appears inconsistent with the top-k positive sampling described in Section 4.4; with k=0 no positive sample would exist, so the searched range is presumably 1 to 5.
  4. [Section 5.7 and Reference [37]] The text says features are drawn from the 'Xavier uniform distribution [37]', but reference [37] is the heterogeneous graph convolution paper, not the original source of Xavier initialization; please fix the citation.
  5. [Figures 4 and 5] The axis label 'Value of c' in Figure 4(b) should read τ_c, and the labels 'llow' and 'lhigh' in Figure 5 are undefined; please provide explicit axis labels and a legend or caption definition.
  6. [Section 5.5, Table 3] The statement that RASH's NMI and ARI on YELP are 'improved by 13.74% and 14.21%' over the second-best method should be phrased as percentage-point differences (i.e., 74.93 vs. 61.19 and 77.08 vs. 62.87), to avoid confusion with relative improvements.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: RASH's homophily/heterophily separation is a self-supervised bootstrap evaluated on external benchmarks, not a derivation that reduces to its inputs.

full rationale

The central mechanism (Eqs. 9-10) defines homophilic and heterophilic graphs as products and complements of learned relation-importance weights, and the contrastive loss (Eqs. 13-16) trains the same encoder with those self-generated views. This is a self-supervised bootstrap rather than a circular derivation: the paper never claims to predict an externally defined homophily label, and its reported gains are measured against held-out labels on DBLP, ACM, IMDB, and YELP in classification, clustering, and similarity search. The concern that high two-hop importance may not track label agreement is a genuine empirical assumption about the inductive bias, but it is not an equation-level equivalence between input and output. The only self-citation that appears in the method is reference [37], used in Sec. 4.1 to justify the standard node-aggregation and type-aggregation encoder; that choice is not load-bearing for the proposed relational-importance separation or the contrastive losses. No fitted parameter is renamed as a prediction, no uniqueness theorem is imported from the authors' prior work, and no known result is repackaged under new coordinates. The paper is therefore not circular, though the validity of the homophily/heterophily heuristic remains an open empirical question.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central claims rest on the learned edge importance being a meaningful indicator of homophily, and on hypergraph convolution providing useful edge contexts. Neither is externally validated in the paper. The Gumbel noise inconsistency and the self-referential positive sampling add further uncertainty.

free parameters (6)
  • Number of positive samples K = Tuned per dataset in range 0-5; Figure 4a shows decreasing performance with larger K
    Top-K positive samples for the contrastive loss. K=0 would make the loss undefined, so the practical setting is dataset-specific and not given by a principle.
  • Contrastive temperature tau_c = 0.4 for ACM, DBLP, IMDB; 0.6 for YELP
    Tuned in {0.4, 0.8} with step 0.1; affects sharpness of the InfoNCE distribution.
  • Gumbel temperature tau = 1.0
    Fixed constant controlling how close the edge weight distribution is to binary.
  • Delta (Gumbel noise) = 1e-4
    Eq 8 defines delta as sampled from Uniform(0,1), but Section 5.3 sets delta to 1e-4, making the sampling deterministic and contradicting the Gumbel-Max reparameterization claim.
  • Number of low-pass and high-pass filtering layers = 1 or 2 per dataset
    Tuned per dataset; Figure 5 shows optimal within 2 layers for the four datasets.
  • Feature dimensions = 64 to 512
    Searched via grid; the final dimension per dataset is not reported.
assumptions (4)
  • domain assumption Hypergraph convolution (Eq 6) models higher-order relationships of heterogeneous edges
    The dual hypergraph with size-2 hyperedges in M^r and node-level hyperedges in M^{rT} is assumed to capture contextual edge information; this is inherited from [14] but not proved for heterogeneous graphs.
  • ad hoc to paper Product of learned edge importances along a two-hop path is a valid proxy for label homophily/heterophily (Eq 9-10)
    No theoretical or empirical validation that high-importance products correspond to same-label nodes. The method never tests the constructed graphs against ground truth.
  • ad hoc to paper Low edge importance indicates heterophily rather than irrelevance (Eq 10)
    The model treats low importance as evidence of different labels, which is not justified; a low score could mean a noisy or uninformative relation.
  • domain assumption Contrastive InfoNCE with top-K positive samples selected from the model's own representations and weights yields task-relevant representations
    Standard contrastive learning assumption, but the positive sample selection is self-referential because the weights and representations come from the same network.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Enhancing Homophily-Heterophily Separation: Relation-Aware Learning in Heterogeneous Graphs." pith.science (2026). https://pith.science/paper/RHVUU5KJ

@misc{pith2026250620980,
  author       = {Pith},
  title        = {Pith review of: Enhancing Homophily-Heterophily Separation: Relation-Aware Learning in Heterogeneous Graphs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RHVUU5KJ}},
  note         = {Machine review of arXiv:2506.20980}
}
read the original abstract

Real-world networks usually have a property of node heterophily, that is, the connected nodes usually have different features or different labels. This heterophily issue has been extensively studied in homogeneous graphs but remains under-explored in heterogeneous graphs, where there are multiple types of nodes and edges. Capturing node heterophily in heterogeneous graphs is very challenging since both node/edge heterogeneity and node heterophily should be carefully taken into consideration. Existing methods typically convert heterogeneous graphs into homogeneous ones to learn node heterophily, which will inevitably lose the potential heterophily conveyed by heterogeneous relations. To bridge this gap, we propose Relation-Aware Separation of Homophily and Heterophily (RASH), a novel contrastive learning framework that explicitly models high-order semantics of heterogeneous interactions and adaptively separates homophilic and heterophilic patterns. Particularly, RASH introduces dual heterogeneous hypergraphs to encode multi-relational bipartite subgraphs and dynamically constructs homophilic graphs and heterophilic graphs based on relation importance. A multi-relation contrastive loss is designed to align heterogeneous and homophilic/heterophilic views by maximizing mutual information. In this way, RASH simultaneously resolves the challenges of heterogeneity and heterophily in heterogeneous graphs. Extensive experiments on benchmark datasets demonstrate the effectiveness of RASH across various downstream tasks. The code is available at: https://github.com/zhengziyu77/RASH.

Figures

Figures reproduced from arXiv: 2506.20980 by the authors.

Figure 1
Figure 1. This image is a real movie-actor heterogeneous [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The overall framework of the proposed model (RASH). It utilizes the importance of the heterogeneous edges obtained [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 4
Figure 4. The sensitivity of hyper-parameter. 1 2 3 4 5 lhigh 1 2 3 4 5 l l o w 91.14 92.20 92.74 92.48 91.17 91.73 93.22 92.30 92.16 91.07 91.90 93.75 93.49 92.44 91.50 91.81 92.20 92.40 91.96 89.88 91.70 91.94 91.79 90.69 89.37 ACM 90 91 92 93 1 2 3 4 5 lhigh 1 2 3 4 5 l l o w 94.00 93.41 92.97 92.49 92.61 92.81 93.28 92.95 92.10 91.70 92.87 93.55 92.14 92.28 91.93 93.53 92.94 92.62 92.49 91.57 93.25 93.13 92.39 91.83 91.56… view at source ↗
Figures from the paper (3 more)
Figure 5
Figure 5. Figure 5: The sensitivity of low-pass filtering and high-pass [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Robustness analysis on ACM and IMDB 6 Conclusion In this paper, we investigate the challenge of mining heterophily from a homogeneous perspective with existing heterogeneous graphs without preserving both heterophily and heterogeneity through a real movie actor example…
Figure 7
Figure 7. Figure 7: Stability analysis on YELP and ACM [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 29 canonical work pages

  1. [1]

    Deyu Bo, Xiao Wang, Chuan Shi, and Huawei Shen. 2021. Beyond low-frequency information in graph convolutional networks. InProceedings of the AAAI confer- ence on artificial intelligence, Vol. 35. 3950–3957

    2021

  2. [2]

    P Kingma Diederik. 2014. Adam: A method for stochastic optimization.(No Title) (2014)

    work page 2014

  3. [3]

    Yuxiao Dong, Nitesh V Chawla, and Ananthram Swami. 2017. metapath2vec: Scalable representation learning for heterogeneous networks. InSIGKDD. ACM, 135–144

    work page 2017

  4. [4]

    Haoran Duan, Cheng Xie, and Linyu Li. 2024. Reserving-masking-reconstruction model for self-supervised heterogeneous graph representation. InProceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. 689–700

    work page 2024

  5. [5]

    Yifan Feng, Haoxuan You, Zizhao Zhang, Rongrong Ji, and Yue Gao. 2019. Hy- pergraph neural networks. InProceedings of the AAAI conference on artificial intelligence, Vol. 33. 3558–3565

    2019

  6. [6]

    Xinyu Fu, Jiani Zhang, Ziqiao Meng, and Irwin King. 2020. Magnn: Metap- ath aggregated graph neural network for heterogeneous graph embedding. In Proceedings of the web conference 2020. 2331–2341

    work page 2020

  7. [7]

    Gongde Guo, Hui Wang, David Bell, Yaxin Bi, and Kieran Greer. 2003. KNN model-based approach in classification. InOn The Move to Meaningful Internet Systems 2003: CoopIS, DOA, and ODBASE: OTM Confederated International Confer- ences, CoopIS, DOA, and ODBASE 2003, Catania, Sicily, Italy, November 3-7, 2003. Proceedings. Springer, 986–996

    2003

  8. [8]

    Jiayan Guo, Lun Du, Wendong Bi, Qiang Fu, Xiaojun Ma, Xu Chen, Shi Han, Dongmei Zhang, and Yan Zhang. 2023. Homophily-oriented heterogeneous graph rewiring. InProceedings of the ACM Web Conference 2023. 511–522

    work page 2023

Show all 49 references
  1. [9]

    Ziniu Hu, Yuxiao Dong, Kuansan Wang, and Yizhou Sun. 2020. Heterogeneous graph transformer. InProceedings of the web conference 2020. 2704–2710

    2020

  2. [10]

    Jing Huang and Jie Yang. 2021. Unignn: a unified framework for graph and hypergraph neural networks.arXiv preprint arXiv:2105.00956(2021)

    2021 arXiv

  3. [11]

    Eric Jang, Shixiang Gu, and Ben Poole. 2016. Categorical reparameterization with gumbel-softmax.arXiv preprint arXiv:1611.01144(2016)

    2016 arXiv

  4. [12]

    Xunqiang Jiang, Tianrui Jia, Yuan Fang, Chuan Shi, Zhe Lin, and Hui Wang. 2021. Pre-training on large-scale heterogeneous graph. InProceedings of the 27th ACM SIGKDD conference on knowledge discovery & data mining. 756–766

    2021

  5. [13]

    Baoyu Jing, Chanyoung Park, and Hanghang Tong. 2021. Hdmi: High-order deep multiplex infomax. InProceedings of the Web Conference 2021. 2414–2424

    2021

  6. [14]

    Jaehyeong Jo, Jinheon Baek, Seul Lee, Dongki Kim, Minki Kang, and Sung Ju Hwang. 2021. Edge representation learning with hypergraphs.Advances in Neural Information Processing Systems34 (2021), 7534–7546

    2021

  7. [15]

    Jintang Li, Zheng Wei, Jiawang Dan, Jing Zhou, Yuchang Zhu, Ruofan Wu, Baokun Wang, Zhang Zhen, Changhua Meng, Hong Jin, et al. 2023. Hetero 2 Net: Heterophily-aware Representation Learning on Heterogeneous Graphs.arXiv preprint arXiv:2310.11664(2023)

    2023 arXiv

  8. [16]

    Xiang Li, Renyu Zhu, Yao Cheng, Caihua Shan, Siqiang Luo, Dongsheng Li, and Weining Qian. 2022. Finding global homophily in graph neural networks when meeting heterophily. InInternational Conference on Machine Learning. PMLR, 13242–13256

    2022

  9. [17]

    Yixin Liu, Yizhen Zheng, Daokun Zhang, Vincent CS Lee, and Shirui Pan. 2023. Beyond smoothing: Unsupervised graph representation learning with edge het- erophily discriminating. InProceedings of the AAAI conference on artificial intelli- gence, Vol. 37. 4516–4524

    2023

  10. [18]

    Sitao Luan, Chenqing Hua, Qincheng Lu, Jiaqi Zhu, Mingde Zhao, Shuyuan Zhang, Xiao-Wen Chang, and Doina Precup. 2022. Revisiting heterophily for graph neural networks.Advances in neural information processing systems35 (2022), 1362–1375

    2022

  11. [19]

    Chris J Maddison, Andriy Mnih, and Yee Whye Teh. 2016. The concrete distri- bution: A continuous relaxation of discrete random variables.arXiv preprint arXiv:1611.00712(2016)

    2016 arXiv

  12. [20]

    Yujie Mo, Zhihe Lu, Runpeng Yu, Xiaofeng Zhu, and Xinchao Wang. 2024. Revis- iting Self-Supervised Heterogeneous Graph Learning from Spectral Clustering Perspective.Advances in Neural Information Processing Systems37 (2024), 43133– 43163

    2024

  13. [21]

    Yujie Mo, Feiping Nie, Ping Hu, Heng Tao Shen, Zheng Zhang, Xinchao Wang, and Xiaofeng Zhu. [n. d.]. Self-Supervised Heterogeneous Graph Learning: a Homophily and Heterogeneity View. InThe Twelfth International Conference on Learning Representations

  14. [22]

    Shirui Pan, Ruiqi Hu, Guodong Long, Jing Jiang, Lina Yao, and Chengqi Zhang

  15. [23]

    Chanyoung Park, Donghyun Kim, Jiawei Han, and Hwanjo Yu. 2020. Unsu- pervised attributed multiplex network embedding. InProceedings of the AAAI conference on artificial intelligence, Vol. 34. 5371–5378

    2020

  16. [24]

    Joonhyung Park, Jaeyun Song, and Eunho Yang. 2021. Graphens: Neighbor-aware ego network synthesis for class-imbalanced node classification. InInternational conference on learning representations

    2021

  17. [25]

    Michael Schlichtkrull, Thomas N Kipf, Peter Bloem, Rianne Van Den Berg, Ivan Titov, and Max Welling. 2018. Modeling relational data with graph convolu- tional networks. InThe semantic web: 15th international conference, ESWC 2018, Heraklion, Crete, Greece, June 3–7, 2018, proc...

    2018

  18. [26]

    Zhixiang Shen and Zhao Kang. 2024. When Heterophily Meets Heterogeneous Graphs: Latent Graphs Guided Unsupervised Representation Learning.arXiv preprint arXiv:2409.00687(2024)

    2024 arXiv

  19. [27]

    Chuan Shi, Yuanfu Lu, Linmei Hu, Zhiyuan Liu, and Huadong Ma. 2020. RHINE: Relation structure-aware heterogeneous information network embedding.IEEE Transactions on Knowledge and Data Engineering34, 1 (2020), 433–447. KDD ’25, August 3–7, 2025, Toronto, ON, Canada Ziyu Zheng, ...

    2020

  20. [28]

    Mingyue Tang, Carl Yang, and Pan Li. 2022. Graph auto-encoder via neighborhood wasserstein reconstruction.arXiv preprint arXiv:2202.09025(2022)

    2022 arXiv

  21. [29]

    Yijun Tian, Kaiwen Dong, Chunhui Zhang, Chuxu Zhang, and Nitesh V Chawla

  22. [30]

    Xiao Wang, Houye Ji, Chuan Shi, Bai Wang, Yanfang Ye, Peng Cui, and Philip S Yu

  23. [31]

    Xiao Wang, Nian Liu, Hui Han, and Chuan Shi. 2021. Self-supervised heteroge- neous graph neural network with co-contrastive learning. InProceedings of the 27th ACM SIGKDD conference on knowledge discovery & data mining. 1726–1736

    2021

  24. [32]

    Felix Wu, Amauri Souza, Tianyi Zhang, Christopher Fifty, Tao Yu, and Kilian Weinberger. 2019. Simplifying graph convolutional networks. InInternational conference on machine learning. PMLR, 6861–6871

    2019

  25. [33]

    Teng Xiao, Huaisheng Zhu, Zhengyu Chen, and Suhang Wang. 2024. Simple and asymmetric graph contrastive learning without augmentations.Advances in Neural Information Processing Systems36 (2024)

    2024

  26. [34]

    Naganand Yadati, Madhav Nimishakavi, Prateek Yadav, Vikram Nitin, Anand Louis, and Partha Talukdar. 2019. Hypergcn: A new method for training graph convolutional networks on hypergraphs.Advances in neural information process- ing systems32 (2019)

    2019

  27. [35]

    Meng Yan, Haibin Huang, Ying Liu, Juan Zhao, Xiyue Gao, Cai Xu, Ziyu Guan, and Wei Zhao. 2025. TruthSR: Trustworthy Sequential Recommender Systems via User-generated Multimodal Content. InDatabase Systems for Advanced Ap- plications, Makoto Onizuka, Jae-Gil Lee, Yongxin Tong, ...

    2025

  28. [36]

    Xiaocheng Yang, Mingyu Yan, Shirui Pan, Xiaochun Ye, and Dongrui Fan. 2023. Simple and efficient heterogeneous graph neural network. InProceedings of the AAAI conference on artificial intelligence, Vol. 37. 10816–10824

    2023

  29. [37]

    Yaming Yang, Ziyu Guan, Jianxin Li, Wei Zhao, Jiangtao Cui, and Quan Wang

  30. [38]

    Yaming Yang, Ziyu Guan, Zhe Wang, Wei Zhao, Cai Xu, Weigang Lu, and Jianbin Huang. 2022. Self-supervised heterogeneous graph pre-training based on struc- tural clustering.Advances in Neural Information Processing Systems35 (2022), 16962–16974

    2022

  31. [39]

    Yuning You, Tianlong Chen, Yongduo Sui, Ting Chen, Zhangyang Wang, and Yang Shen. 2020. Graph contrastive learning with augmentations.Advances in neural information processing systems33 (2020), 5812–5823

    2020

  32. [40]

    Jianxiang Yu, Qingqing Ge, Xiang Li, and Aoying Zhou. 2024. Heterogeneous Graph Contrastive Learning with Meta-Path Contexts and Adaptively Weighted Negative Samples.IEEE Transactions on Knowledge and Data Engineering(2024)

    2024

  33. [41]

    Chuxu Zhang, Dongjin Song, Chao Huang, Ananthram Swami, and Nitesh V Chawla. 2019. Heterogeneous graph neural network. InProceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining. 793–803

    2019

  34. [42]

    Fanjin Zhang, Xiao Liu, Jie Tang, Yuxiao Dong, Peiran Yao, Jie Zhang, Xiaotao Gu, Yan Wang, Bin Shao, Rui Li, et al. 2019. OAG: Toward linking large-scale heterogeneous entity graphs. InProceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data m...

    2019

  35. [43]

    Jianan Zhao, Xiao Wang, Chuan Shi, Zekuan Liu, and Yanfang Ye. 2020. Net- work schema preserving heterogeneous information network embedding. In International joint conference on artificial intelligence (IJCAI)

    2020

  36. [44]

    Xin Zheng, Yi Wang, Yixin Liu, Ming Li, Miao Zhang, Di Jin, Philip S Yu, and Shirui Pan. 2022. Graph neural networks for graphs with heterophily: A survey. arXiv preprint arXiv:2202.07082(2022)

    2022 arXiv

  37. [45]

    Jiong Zhu, Ryan A Rossi, Anup Rao, Tung Mai, Nedim Lipka, Nesreen K Ahmed, and Danai Koutra. 2021. Graph neural networks with heterophily. InProceedings of the AAAI conference on artificial intelligence, Vol. 35. 11168–11176. Enhancing Homophily-Heterophily Separation: Relatio...

    2021

  38. [2018]

    In Proceedings of the 27th International Joint Conference on Artificial Intelligence

    Adversarially regularized graph autoencoder for graph embedding. In Proceedings of the 27th International Joint Conference on Artificial Intelligence. 2609–2615

  39. [2019]

    InThe World Wide Web Conference

    Heterogeneous graph attention network. InThe World Wide Web Conference. 2022–2032

    2022

  40. [2021]

    TKDE(2021)

    Interpretable and efficient heterogeneous graph convolutional network. TKDE(2021)

    2021

  41. [2023]

    InProceedings of the AAAI Conference on Artificial Intelligence, Vol

    Heterogeneous graph masked autoencoders. InProceedings of the AAAI Conference on Artificial Intelligence, Vol. 37. 9997–10005

Pith tools

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