pith. machine review for the scientific record. sign in

arxiv: 2604.26301 · v1 · submitted 2026-04-29 · 💻 cs.LG

Recognition: unknown

Cheeger--Hodge Contrastive Learning for Structurally Robust Graph Representation Learning

Mengyang Zhao , Longlong Li , Cunquan Qu
Authors on Pith no claims yet

Pith reviewed 2026-05-07 13:34 UTC · model grok-4.3

classification 💻 cs.LG
keywords Cheeger-Hodge signaturegraph contrastive learningstructural robustnessalgebraic connectivityHodge Laplaciangraph representationsperturbation stability
0
0 comments X

The pith

Aligning graph encoders to a stable Cheeger-Hodge signature yields representations robust to structural perturbations.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Graph contrastive learning often produces brittle embeddings under structural changes because it relies on augmentation choices to define learned invariances. This paper proposes Cheeger-Hodge Contrastive Learning that instead aligns encoder outputs to a joint signature combining algebraic connectivity from λ₂ with the low-frequency spectrum of the 1-Hodge Laplacian across augmented graph views. The signature is constructed to remain consistent under local perturbations while encoding both global connectivity and higher-order structure. If the alignment succeeds, the resulting embeddings resist small structural alterations and deliver stronger performance and generalization on graph tasks.

Core claim

CHCL aligns encoder representations with the Cheeger-Hodge joint signature across augmented views to learn graph embeddings robust to local structural perturbations. The signature combines a Cheeger-inspired connectivity measure derived from algebraic connectivity λ₂ with the low-frequency spectrum of the 1-Hodge Laplacian, thereby capturing both global connectivity and higher-order structural information while remaining stable under perturbations.

What carries the argument

The Cheeger-Hodge joint signature, formed by combining algebraic connectivity λ₂ with the low-frequency 1-Hodge Laplacian spectrum, that acts as a perturbation-stable target for contrastive alignment.

If this is right

  • Learned embeddings maintain accuracy under edge additions, deletions, or noise in the input graphs.
  • Performance gains appear in both standard benchmarks and transfer settings for node and graph classification.
  • The method reduces dependence on specific augmentation strategies to achieve invariance.
  • Generalization improves when models trained on one set of graphs are applied to structurally altered versions.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same stable signature could serve as a regularizer in supervised graph neural network training to encourage structural consistency.
  • Real-world graphs with measurement noise or missing links might benefit from pre-alignment to this signature before downstream tasks.
  • Extending the target signature to additional Hodge Laplacian frequencies could capture richer topological features without changing the contrastive setup.

Load-bearing premise

The Cheeger-Hodge joint signature is inherently stable under local structural perturbations and that forcing alignment to it produces representations whose robustness follows directly from this alignment.

What would settle it

Apply small random edge perturbations to graphs from the evaluation benchmarks and check whether λ₂ and the low-frequency 1-Hodge eigenvalues remain nearly unchanged; large shifts would show the signature is not stable enough for the alignment to confer robustness.

Figures

Figures reproduced from arXiv: 2604.26301 by Cunquan Qu, Longlong Li, Mengyang Zhao.

Figure 1
Figure 1. Figure 1: Schematic of CHCL and its capabilities on graph representation learning tasks. CHCL theory embeds Cheeger and view at source ↗
Figure 2
Figure 2. Figure 2: Ablation study comparing CHCL with three ablated view at source ↗
Figure 3
Figure 3. Figure 3: Sensitivity analysis of key hyperparameters in CHCL. view at source ↗
Figure 4
Figure 4. Figure 4: Performance comparison of CHCL and baseline methods under varying levels of edge dropping and feature masking view at source ↗
Figure 5
Figure 5. Figure 5: Interpretability of the structural semantics learned by CHCL. view at source ↗
read the original abstract

Graph Contrastive Learning (GCL) has emerged as a prominent framework for unsupervised graph representation learning. However, relying on augmentation design alone to define the invariances learned by GCL can be brittle under structural perturbations. To address this issue, we propose Cheeger--Hodge Contrastive Learning (CHCL), a framework that aligns a perturbation-stable Cheeger--Hodge joint signature across augmented views for robust graph representation learning. The proposed signature combines a Cheeger-inspired connectivity signature derived from the algebraic connectivity \(\lambda_2\) with the low-frequency spectrum of the 1-Hodge Laplacian, thereby capturing both global connectivity and higher-order structural information. By aligning encoder representations with the proposed Cheeger--Hodge joint signature across augmented views, CHCL learns graph embeddings that are robust to local structural perturbations. Extensive experiments on standard benchmarks, transfer settings demonstrate that CHCL consistently improves performance, robustness, and generalization.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript proposes Cheeger--Hodge Contrastive Learning (CHCL) for unsupervised graph representation learning. It constructs a joint signature from the algebraic connectivity λ₂ and the low-frequency spectrum of the 1-Hodge Laplacian, then aligns encoder representations to this signature across augmented graph views to obtain embeddings that are robust to local structural perturbations. The abstract asserts that this yields consistent improvements in performance, robustness, and generalization on standard benchmarks and transfer settings.

Significance. If the stability of the proposed signature under the method's augmentations can be established and the empirical gains verified with quantitative detail, the work would provide a concrete link between spectral graph invariants and contrastive objectives, potentially improving robustness in graph representation learning beyond augmentation heuristics alone. The reliance on well-studied quantities (λ₂ and Hodge spectrum) is a methodological strength that could support future theoretical analysis.

major comments (2)
  1. Abstract: The central claim that CHCL 'consistently improves performance, robustness, and generalization' is asserted without any quantitative results, error bars, ablation studies, or specific metrics, leaving the empirical support for the robustness claim unverified.
  2. Abstract: The assertion that the Cheeger--Hodge joint signature is 'perturbation-stable' and that alignment to it produces robustness is not accompanied by a derivation, eigenvalue perturbation bound, or measurement of signature drift (pre- vs. post-augmentation), so the transfer of stability from signature to learned embeddings remains an unproven assumption.
minor comments (1)
  1. The abstract sentence 'Extensive experiments on standard benchmarks, transfer settings demonstrate...' is grammatically incomplete and should be revised for clarity.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive feedback on our manuscript. We have revised the abstract to incorporate quantitative highlights from our experiments and to better ground the stability claims. We address each major comment below.

read point-by-point responses

  1. Referee: Abstract: The central claim that CHCL 'consistently improves performance, robustness, and generalization' is asserted without any quantitative results, error bars, ablation studies, or specific metrics, leaving the empirical support for the robustness claim unverified.

    Authors: We agree that the abstract would be strengthened by including concrete quantitative support. In the revised version, we have updated the abstract to report average performance gains (e.g., +2.3% node classification accuracy and +1.8% graph classification accuracy across benchmarks), robustness improvements under structural perturbations (measured via accuracy retention rates), and references to the full experimental tables that contain error bars, statistical significance tests, and ablation studies. revision: yes

  2. Referee: Abstract: The assertion that the Cheeger--Hodge joint signature is 'perturbation-stable' and that alignment to it produces robustness is not accompanied by a derivation, eigenvalue perturbation bound, or measurement of signature drift (pre- vs. post-augmentation), so the transfer of stability from signature to learned embeddings remains an unproven assumption.

    Authors: The full manuscript already contains empirical measurements of signature drift under the exact augmentations used in CHCL, showing that the joint Cheeger-Hodge signature exhibits substantially lower relative change than alternative graph invariants. We have added a short paragraph to the revised abstract summarizing these measurements and inserted a new subsection that cites existing perturbation bounds for algebraic connectivity and the Hodge Laplacian spectrum. A complete end-to-end theoretical derivation of how signature stability transfers to the learned embeddings is beyond the current scope but is now explicitly flagged as future work; the contrastive alignment objective and the empirical results provide the primary support for the robustness claims. revision: partial

Circularity Check

0 steps flagged

No circularity detected; derivation remains self-contained

full rationale

The Cheeger--Hodge joint signature is constructed directly from two independently defined spectral objects (algebraic connectivity λ₂ of the graph Laplacian and the low-frequency part of the 1-Hodge Laplacian spectrum). These quantities pre-exist the paper and are not defined in terms of the contrastive alignment or the robustness property being claimed. The subsequent alignment step is a standard application of the contrastive learning objective to this external target signature; no equation reduces the output robustness to a fitted parameter or to a self-citation chain. No load-bearing premise relies on prior work by the same authors to forbid alternatives or to import an ansatz. The paper therefore supplies an independent construction rather than a self-referential loop.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 1 invented entities

Abstract supplies no explicit free parameters or derivation steps; the central claim rests on the unstated assumption that the joint signature can be computed and aligned in a way that transfers robustness. No invented entities beyond the named signature itself.

axioms (2)
  • standard math Algebraic connectivity λ2 and the spectrum of the 1-Hodge Laplacian are well-defined for undirected graphs and capture global connectivity and higher-order structure respectively.
    Invoked when the abstract states the signature 'combines a Cheeger-inspired connectivity signature derived from λ2 with the low-frequency spectrum of the 1-Hodge Laplacian'.
  • domain assumption Aligning encoder outputs to a fixed structural signature across augmentations produces representations invariant to the perturbations used in augmentation.
    This is the load-bearing modeling assumption that turns the signature into a training target; it is not derived in the abstract.
invented entities (1)
  • Cheeger-Hodge joint signature no independent evidence
    purpose: A perturbation-stable descriptor that fuses global connectivity (λ2) and low-frequency higher-order information for use as a contrastive target.
    Introduced as the core technical object of CHCL; no independent evidence of its stability is supplied in the abstract.

pith-pipeline@v0.9.0 · 5456 in / 1574 out tokens · 39690 ms · 2026-05-07T13:34:48.079149+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

81 extracted references · 6 canonical work pages · 3 internal anchors

  1. [1]

    Kipf and Max Welling

    Thomas N. Kipf and Max Welling. Semi-supervised classification with graph convolutional networks. InInternational Conference on Learning Representations (ICLR), 2017

    2017

  2. [2]

    Multi-scale harmonic encoding for feature-wise graph message passing

    Longlong Li, Mengyang Zhao, Guanghui Wang, and Cunquan Qu. Multi-scale harmonic encoding for feature-wise graph message passing. arXiv preprint arXiv:2505.15015, 2025

    work page arXiv 2025

  3. [3]

    Neural message passing for quantum chemistry

    Justin Gilmer, Samuel S Schoenholz, Patrick F Riley, Oriol Vinyals, and George E Dahl. Neural message passing for quantum chemistry. In International conference on machine learning, pages 1263–1272. Pmlr, 2017

    2017

  4. [4]

    Path complex neural network for molecular property prediction

    Longlong Li, Xiang LIU, Guanghui Wang, Yu Guang Wang, and KELIN XIA. Path complex neural network for molecular property prediction. In ICML 2024 Workshop on Geometry-grounded Representation Learning and Generative Modeling, 2024

    2024

  5. [5]

    Highly accurate protein structure prediction with alphafold.nature, 596(7873):583–589, 2021

    John Jumper, Richard Evans, Alexander Pritzel, Tim Green, Michael Figurnov, Olaf Ronneberger, Kathryn Tunyasuvunakool, Russ Bates, Augustin ˇZ´ıdek, Anna Potapenko, et al. Highly accurate protein structure prediction with alphafold.nature, 596(7873):583–589, 2021

    2021

  6. [6]

    Mrhgnn: enhanced multimodal relational hypergraph neural network for synergistic drug combination forecasting.IEEE transactions on neural networks and learning systems, 2025

    Mengjie Chen, Ming Zhang, Guiying Yan, Guanghui Wang, and Cun- quan Qu. Mrhgnn: enhanced multimodal relational hypergraph neural network for synergistic drug combination forecasting.IEEE transactions on neural networks and learning systems, 2025

    2025

  7. [7]

    How powerful are graph neural networks? InInternational Conference on Learning Representations, 2019

    Keyulu Xu, Weihua Hu, Jure Leskovec, and Stefanie Jegelka. How powerful are graph neural networks? InInternational Conference on Learning Representations, 2019. JOURNAL OF LATEX CLASS FILES, VOL. 18, NO. 9, SEPTEMBER 2020 10

    2019

  8. [8]

    Graph Attention Networks.International Conference on Learning Representations, 2018

    Petar Veli ˇckovi´c, Guillem Cucurull, Arantxa Casanova, Adriana Romero, Pietro Li`o, and Yoshua Bengio. Graph Attention Networks.International Conference on Learning Representations, 2018. accepted as poster

    2018

  9. [9]

    Z. Ying, J. You, C. Morris, et al. Hierarchical graph representation learning with differentiable pooling. InAdvances in Neural Information Processing Systems, volume 31, 2018

    2018

  10. [10]

    Strategies for pre-training graph neural networks

    Weihua Hu, Bowen Liu, Joseph Gomes, Marinka Zitnik, Percy Liang, Vijay Pande, and Jure Leskovec. Strategies for pre-training graph neural networks. InInternational Conference on Learning Representations, 2020

    2020

  11. [11]

    Zitnik, R

    M. Zitnik, R. Sosi ˇc, and J. Leskovec. Prioritizing network communities. Nature Communications, 9(1):2544, 2018

    2018

  12. [12]

    L. Wu, P. Cui, J. Pei, et al. Graph neural networks: foundation, frontiers and applications. InProceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 4840–4841. ACM, 2022

    2022

  13. [13]

    H. Hu, X. Wang, Y . Zhang, et al. A comprehensive survey on contrastive learning.Neurocomputing, 610:128645, 2024

    2024

  14. [14]

    Z. Wu, S. Pan, F. Chen, et al. A comprehensive survey on graph neural networks.IEEE Transactions on Neural Networks and Learning Systems, 32(1):4–24, 2020

    2020

  15. [15]

    T. Chen, S. Kornblith, M. Norouzi, et al. A simple framework for con- trastive learning of visual representations. InInternational Conference on Machine Learning, pages 1597–1607. PMLR, 2020

    2020

  16. [16]

    H. Yang, Y . Wang, X. Zhao, et al. Multi-level graph knowledge contrastive learning.IEEE Transactions on Knowledge and Data Engineering, 2024

    2024

  17. [17]

    B. Li, B. Jing, and H. Tong. Graph communal contrastive learning. InProceedings of the ACM Web Conference 2022, pages 1203–1213. ACM, 2022

    2022

  18. [18]

    S. Li, X. Wang, A. Zhang, et al. Let invariant rationale discovery inspire graph contrastive learning. InInternational Conference on Machine Learning, pages 13052–13065. PMLR, 2022

    2022

  19. [19]

    Graph contrastive learning with augmentations

    Yuning You, Tianlong Chen, Yongduo Sui, Ting Chen, Zhangyang Wang, and Yang Shen. Graph contrastive learning with augmentations. Advances in neural information processing systems, 33:5812–5823, 2020

    2020

  20. [20]

    R. Yuan, Y . Tang, Y . Wu, et al. Clustering enhanced multiplex graph contrastive representation learning.IEEE Transactions on Neural Networks and Learning Systems, 2023

    2023

  21. [21]

    Zhang, H

    Y . Zhang, H. Zhu, Z. Song, et al. Spectral feature augmentation for graph contrastive learning and beyond. InProceedings of the AAAI Conference on Artificial Intelligence, volume 37, pages 11289–11297. AAAI Press, 2023

    2023

  22. [22]

    Y . Wang, X. Pan, S. Song, et al. Implicit semantic data augmentation for deep networks. InAdvances in Neural Information Processing Systems, volume 32, 2019

    2019

  23. [23]

    Community-invariant graph contrastive learning

    Shiyin Tan, Dongyuan Li, Renhe Jiang, Ying Zhang, and Manabu Oku- mura. Community-invariant graph contrastive learning. InInternational Conference on Machine Learning, pages 47579–47606. PMLR, 2024

    2024

  24. [24]

    Zhang, Z

    S. Zhang, Z. Hu, A. Subramonian, et al. Motif-driven contrastive learning of graph representations.IEEE Transactions on Knowledge and Data Engineering, 36(8):4063–4075, 2024

    2024

  25. [25]

    H. Zhao, X. Yang, C. Deng, et al. Unsupervised structure-adaptive graph contrastive learning.IEEE Transactions on Neural Networks and Learning Systems, 35(10):13728–13740, 2023

    2023

  26. [26]

    M. Peng, X. Juan, and Z. Li. Label-guided graph contrastive learning for semi-supervised node classification.Expert Systems with Applications, 239:122385, 2024

    2024

  27. [27]

    Zhang, Y

    H. Zhang, Y . Ren, L. Fu, et al. Multi-scale self-supervised graph con- trastive learning with injective node augmentation.IEEE Transactions on Knowledge and Data Engineering, 36(1):261–274, 2023

    2023

  28. [28]

    J. Zhu, W. Zeng, J. Zhang, et al. Cross-view graph contrastive learning with hypergraph.Information Fusion, 99:101867, 2023

    2023

  29. [29]

    W. Bu, X. Cao, Y . Zheng, et al. Improving augmentation consistency for graph contrastive learning.Pattern Recognition, 148:110182, 2024

    2024

  30. [30]

    R. Liu, R. Yin, Y . Liu, et al. Unbiased and augmentation-free self-supervised graph representation learning.Pattern Recognition, 149:110274, 2024

    2024

  31. [31]

    H. Yan, S. Wang, C. Li, et al. Have our cake and eat it: Augmentation diversity and semantic consistency balanced graph contrastive learning. ACM Transactions on Knowledge Discovery from Data, 19(4):1–25, 2025

    2025

  32. [32]

    Semantic-aware contrastive learning for graph classification.Expert Systems with Applications, page 129976, 2025

    Chengcheng Xu, Tong Han, Tianfeng Wang, Xiao Han, and Zhisong Pan. Semantic-aware contrastive learning for graph classification.Expert Systems with Applications, page 129976, 2025

    2025

  33. [33]

    Y . Liu, L. Shu, C. Chen, et al. Fine-grained semantics enhanced contrastive learning for graphs.IEEE Transactions on Knowledge and Data Engineering, 2024

    2024

  34. [34]

    P. Bao, R. Yan, and S. Pan. Co-augmentation of structure and feature for boosting graph contrastive learning.Information Sciences, 676:120792, 2024

    2024

  35. [35]

    Topogcl: Topological graph contrastive learning

    Yuzhou Chen, Jose Frias, and Yulia R Gel. Topogcl: Topological graph contrastive learning. InProceedings of the AAAI conference on artificial intelligence, volume 38, pages 11453–11461, 2024

    2024

  36. [36]

    Rhomboid tiling for geo- metric graph deep learning

    Yipeng Zhang, Longlong Li, and Kelin Xia. Rhomboid tiling for geo- metric graph deep learning. InForty-second International Conference on Machine Learning, 2025

    2025

  37. [37]

    On the bottleneck of graph neural networks and its practical implications

    Uri Alon and Eran Yahav. On the bottleneck of graph neural networks and its practical implications. InInternational Conference on Learning Representations, 2021

    2021

  38. [38]

    Revisiting over-smoothing and over-squashing using ollivier-ricci curvature

    Khang Nguyen, Hieu Nong, Vinh Nguyen, Nhat Ho, Stanley Osher, and Tan Nguyen. Revisiting over-smoothing and over-squashing using ollivier-ricci curvature. InProceedings of the 40th International Con- ference on Machine Learning, ICML’23. JMLR.org, 2023

    2023

  39. [39]

    American Mathe- matical Soc., 1997

    Fan RK Chung.Spectral graph theory, volume 92. American Mathe- matical Soc., 1997

    1997

  40. [40]

    Cheeger inequalities on simplicial complexes.ANNALI SCUOLA NORMALE SUPERIORE-CLASSE DI SCIENZE, pages 30–30, 2022

    J ¨urgen Jost and Dong Zhang. Cheeger inequalities on simplicial complexes.ANNALI SCUOLA NORMALE SUPERIORE-CLASSE DI SCIENZE, pages 30–30, 2022

    2022

  41. [41]

    Disentangling the spectral properties of the hodge laplacian: not all small eigenvalues are equal

    Vincent P Grande and Michael T Schaub. Disentangling the spectral properties of the hodge laplacian: not all small eigenvalues are equal. In ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 9896–9900. IEEE, 2024

    2024

  42. [42]

    Hamilton, Pietro Li `o, Yoshua Bengio, and R Devon Hjelm

    Petar Veli ˇckovi´c, William Fedus, William L. Hamilton, Pietro Li `o, Yoshua Bengio, and R Devon Hjelm. Deep Graph Infomax. In International Conference on Learning Representations, 2019

    2019

  43. [43]

    Infograph: Unsupervised and semi-supervised graph-level representation learning via mutual information maximization

    Fan-Yun Sun, Jordan Hoffman, Vikas Verma, and Jian Tang. Infograph: Unsupervised and semi-supervised graph-level representation learning via mutual information maximization. InInternational Conference on Learning Representations, 2019

    2019

  44. [44]

    Graph contrastive learning with adaptive augmentation

    Yanqiao Zhu, Yichen Xu, Feng Yu, Qiang Liu, Shu Wu, and Liang Wang. Graph contrastive learning with adaptive augmentation. InProceedings of the web conference 2021, pages 2069–2080, 2021

    2021

  45. [45]

    Autogcl: Automated graph contrastive learning via learnable view generators

    Yihang Yin, Qingzhong Wang, Siyu Huang, Haoyi Xiong, and Xiang Zhang. Autogcl: Automated graph contrastive learning via learnable view generators. InProceedings of the AAAI conference on artificial intelligence, volume 36, pages 8892–8900, 2022

    2022

  46. [46]

    Graph contrastive learning with cohesive subgraph awareness

    Yucheng Wu, Leye Wang, Xiao Han, and Han-Jia Ye. Graph contrastive learning with cohesive subgraph awareness. InProceedings of the ACM Web Conference 2024, pages 629–640, 2024

    2024

  47. [47]

    Spectral augmentation for self-supervised learning on graphs

    Lu Lin and Jinghui Chen. Spectral augmentation for self-supervised learning on graphs. InThe Eleventh International Conference on Learning Representations, 2023

    2023

  48. [48]

    Spectral augmentations for graph contrastive learning

    Amur Ghose, Yingxue Zhang, Jianye Hao, and Mark Coates. Spectral augmentations for graph contrastive learning. InInternational Con- ference on Artificial Intelligence and Statistics, pages 11213–11266. PMLR, 2023

    2023

  49. [49]

    Revisiting graph contrastive learning from the perspective of graph spectrum

    Nian Liu, Xiao Wang, Deyu Bo, Chuan Shi, and Jian Pei. Revisiting graph contrastive learning from the perspective of graph spectrum. Advances in Neural Information Processing Systems, 35:2972–2983, 2022

    2022

  50. [50]

    Algebraic connectivity of graphs.Czechoslovak mathematical journal, 23(2):298–305, 1973

    Miroslav Fiedler. Algebraic connectivity of graphs.Czechoslovak mathematical journal, 23(2):298–305, 1973

    1973

  51. [51]

    Un- derstanding over-squashing and bottlenecks on graphs via curvature.arXiv preprint arXiv:2111.14522,

    Jake Topping, Francesco Di Giovanni, Benjamin Paul Chamberlain, Xiaowen Dong, and Michael M Bronstein. Understanding over- squashing and bottlenecks on graphs via curvature.arXiv preprint arXiv:2111.14522, 2021

    work page arXiv 2021

  52. [52]

    Discrete differential calculus: Graphs, topologies, and gauge theory.Journal of Mathematical Physics, 35(12):6703–6735, 1994

    Aristophanes Dimakis and Folkert M ¨uller-Hoissen. Discrete differential calculus: Graphs, topologies, and gauge theory.Journal of Mathematical Physics, 35(12):6703–6735, 1994

    1994

  53. [53]

    Discrete hodge operators: An algebraic perspective

    Ralf Hiptmair. Discrete hodge operators: An algebraic perspective. Progress In Electromagnetics Research, 32(247269):27, 2001

    2001

  54. [54]

    Discrete hodge theory on graphs: a tutorial.Computing in Science & Engineering, 15(5):42–55, 2012

    James L Johnson and Tom Goldring. Discrete hodge theory on graphs: a tutorial.Computing in Science & Engineering, 15(5):42–55, 2012

    2012

  55. [55]

    Hodge laplacians on graphs.Siam Review, 62(3):685– 715, 2020

    Lek-Heng Lim. Hodge laplacians on graphs.Siam Review, 62(3):685– 715, 2020

    2020

  56. [56]

    Random walks on simplicial complexes and the normalized hodge 1-laplacian.SIAM Review, 62(2):353–391, 2020

    Michael T Schaub, Austin R Benson, Paul Horn, Gabor Lippner, and Ali Jadbabaie. Random walks on simplicial complexes and the normalized hodge 1-laplacian.SIAM Review, 62(2):353–391, 2020

    2020

  57. [57]

    Hodge theory-based biomolecular data analysis.Scientific Reports, 12(1):9699, 2022

    Ronald Koh Joon Wei, Junjie Wee, Valerie Evangelin Laurent, and Kelin Xia. Hodge theory-based biomolecular data analysis.Scientific Reports, 12(1):9699, 2022. JOURNAL OF LATEX CLASS FILES, VOL. 18, NO. 9, SEPTEMBER 2020 11

    2022

  58. [58]

    Heterophily-aware graph attention network.Pattern Recognition, 156:110738, 2024

    Junfu Wang, Yuanfang Guo, Liang Yang, and Yunhong Wang. Heterophily-aware graph attention network.Pattern Recognition, 156:110738, 2024

    2024

  59. [59]

    Hl-hgat: Heterogeneous graph attention network via hodge-laplacian operator.IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025

    Jinghan Huang, Qiufeng Chen, Pengli Zhu, Yijun Bian, Nanguang Chen, Moo K Chung, and Anqi Qiu. Hl-hgat: Heterogeneous graph attention network via hodge-laplacian operator.IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025

    2025

  60. [60]

    Tensor-fused multi-view graph contrastive learning

    Yujia Wu, Junyi Mo, Elynn Chen, and Yuzhou Chen. Tensor-fused multi-view graph contrastive learning. InPacific-Asia Conference on Knowledge Discovery and Data Mining, pages 16–28. Springer, 2025

    2025

  61. [61]

    Eigenvalues and homol- ogy of flag complexes and vector representations of graphs.Geometric & Functional Analysis GAFA, 15(3):555–566, 2005

    Ron Aharoni, Eli Berger, and Roy Meshulam. Eigenvalues and homol- ogy of flag complexes and vector representations of graphs.Geometric & Functional Analysis GAFA, 15(3):555–566, 2005

    2005

  62. [62]

    Hodge laplacian eigenvalues on surfaces with boundary.Annales math ´ematiques du Qu ´ebec, pages 1–20, 2025

    Mikhail Muravyev. Hodge laplacian eigenvalues on surfaces with boundary.Annales math ´ematiques du Qu ´ebec, pages 1–20, 2025

    2025

  63. [63]

    Combinatorial and hodge laplacians: similarities and differ- ences.SIAM Review, 66(3):575–601, 2024

    Emily Ribando-Gros, Rui Wang, Jiahui Chen, Yiying Tong, and Guo- Wei Wei. Combinatorial and hodge laplacians: similarities and differ- ences.SIAM Review, 66(3):575–601, 2024

    2024

  64. [64]

    Spectral detection of simplicial communities via hodge laplacians.Physical Review E, 104(6):064303, 2021

    Sanjukta Krishnagopal and Ginestra Bianconi. Spectral detection of simplicial communities via hodge laplacians.Physical Review E, 104(6):064303, 2021

    2021

  65. [65]

    Kriege, Franka Bause, Kristian Kersting, Petra Mutzel, and Marion Neumann

    Christopher Morris, Nils M. Kriege, Franka Bause, Kristian Kersting, Petra Mutzel, and Marion Neumann. Tudataset: A collection of bench- mark datasets for learning with graphs. InICML 2020 Workshop on Graph Representation Learning and Beyond (GRL+ 2020), 2020

    2020

  66. [66]

    Open graph benchmark: Datasets for machine learning on graphs.Advances in neural information processing systems, 33:22118–22133, 2020

    Weihua Hu, Matthias Fey, Marinka Zitnik, Yuxiao Dong, Hongyu Ren, Bowen Liu, Michele Catasta, and Jure Leskovec. Open graph benchmark: Datasets for machine learning on graphs.Advances in neural information processing systems, 33:22118–22133, 2020

    2020

  67. [67]

    Chembl: a large-scale bioactivity database for drug discovery.Nucleic acids research, 40(D1):D1100– D1107, 2012

    Anna Gaulton, Louisa J Bellis, A Patricia Bento, Jon Chambers, Mark Davies, Anne Hersey, Yvonne Light, Shaun McGlinchey, David Michalovich, Bissan Al-Lazikani, et al. Chembl: a large-scale bioactivity database for drug discovery.Nucleic acids research, 40(D1):D1100– D1107, 2012

    2012

  68. [68]

    Moleculenet: a benchmark for molecular machine learning.Chemical science, 9(2):513–530, 2018

    Zhenqin Wu, Bharath Ramsundar, Evan N Feinberg, Joseph Gomes, Caleb Geniesse, Aneesh S Pappu, Karl Leswing, and Vijay Pande. Moleculenet: a benchmark for molecular machine learning.Chemical science, 9(2):513–530, 2018

    2018

  69. [69]

    Good: A graph out- of-distribution benchmark.Advances in Neural Information Processing Systems, 35:2059–2073, 2022

    Shurui Gui, Xiner Li, Limei Wang, and Shuiwang Ji. Good: A graph out- of-distribution benchmark.Advances in Neural Information Processing Systems, 35:2059–2073, 2022

    2059

  70. [70]

    Adversarial graph augmentation to improve graph contrastive learning.Advances in Neural Information Processing Systems, 34:15920–15933, 2021

    Susheel Suresh, Pan Li, Cong Hao, and Jennifer Neville. Adversarial graph augmentation to improve graph contrastive learning.Advances in Neural Information Processing Systems, 34:15920–15933, 2021

    2021

  71. [71]

    Graph contrastive learning automated

    Yuning You, Tianlong Chen, Yang Shen, and Zhangyang Wang. Graph contrastive learning automated. InInternational conference on machine learning, pages 12121–12132. PMLR, 2021

    2021

  72. [72]

    Unifying graph contrastive learning via graph message augmentation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025

    Ziyan Zhang, Bo Jiang, Jin Tang, and Bin Luo. Unifying graph contrastive learning via graph message augmentation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025

    2025

  73. [73]

    Invariant Risk Minimization

    Martin Arjovsky, L ´eon Bottou, Ishaan Gulrajani, and David Lopez-Paz. Invariant risk minimization.arXiv preprint arXiv:1907.02893, 2019

    work page internal anchor Pith review arXiv 1907

  74. [74]

    Out-of-distribution generalization via risk extrapolation (rex)

    David Krueger, Ethan Caballero, Joern-Henrik Jacobsen, Amy Zhang, Jonathan Binas, Dinghuai Zhang, Remi Le Priol, and Aaron Courville. Out-of-distribution generalization via risk extrapolation (rex). InIn- ternational conference on machine learning, pages 5815–5826. PMLR, 2021

    2021

  75. [75]

    Distributionally Robust Neural Networks for Group Shifts: On the Importance of Regularization for Worst-Case Generalization

    Shiori Sagawa, Pang Wei Koh, Tatsunori B Hashimoto, and Percy Liang. Distributionally robust neural networks for group shifts: On the importance of regularization for worst-case generalization.arXiv preprint arXiv:1911.08731, 2019

    work page internal anchor Pith review arXiv 1911

  76. [76]

    Domain-adversarial training of neural networks.Journal of machine learning research, 17(59):1–35, 2016

    Yaroslav Ganin, Evgeniya Ustinova, Hana Ajakan, Pascal Germain, Hugo Larochelle, Franc ¸ois Laviolette, Mario March, and Victor Lempit- sky. Domain-adversarial training of neural networks.Journal of machine learning research, 17(59):1–35, 2016

    2016

  77. [77]

    Deep coral: Correlation alignment for deep domain adaptation

    Baochen Sun and Kate Saenko. Deep coral: Correlation alignment for deep domain adaptation. InEuropean conference on computer vision, pages 443–450. Springer, 2016

    2016

  78. [78]

    mixup: Beyond Empirical Risk Minimization

    Hongyi Zhang, Moustapha Cisse, Yann N Dauphin, and David Lopez- Paz. mixup: Beyond empirical risk minimization.arXiv preprint arXiv:1710.09412, 2017

    work page internal anchor Pith review arXiv 2017

  79. [79]

    Discovering invariant rationales for graph neural networks.arXiv preprint arXiv:2201.12872, 2022

    Ying-Xin Wu, Xiang Wang, An Zhang, Xiangnan He, and Tat-Seng Chua. Discovering invariant rationales for graph neural networks.arXiv preprint arXiv:2201.12872, 2022

    work page arXiv 2022

  80. [80]

    Spectra of combinatorial laplace operators on simplicial complexes.Advances in Mathematics, 244:303– 336, 2013

    Danijela Horak and J ¨urgen Jost. Spectra of combinatorial laplace operators on simplicial complexes.Advances in Mathematics, 244:303– 336, 2013

    2013

Showing first 80 references.