pith. sign in

arxiv: 2606.32016 · v1 · pith:LV5OCR7Xnew · submitted 2026-06-30 · 💻 cs.LG

FedLAB: Traceable Semantic Codebooks for Federated Multimodal Graph Foundation Learning

Zekai Chen , Kairui Yang , Xuaner Chen , Xunkai Li , Xun Wu , Rong-Hua Li , Guoren Wang This is my paper

Pith reviewed 2026-07-01 06:08 UTC · model grok-4.3

classification 💻 cs.LG
keywords federated learningmultimodal graphssemantic codebookstraceabilitygraph foundation modelsbarycenter pre-trainingdata privacy
0
0 comments X

The pith

Typed hierarchical codebooks organize modality evidence, node semantics, and topology context for traceable federated multimodal graph learning without sharing raw data.

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

The paper establishes a framework that structures knowledge from distributed multimodal graphs into typed hierarchical codebooks covering three aspects: modality evidence, node semantics, and topology context. These codebooks undergo refinement via federated semantic barycenter pre-training while all raw contents and local structures remain on client devices. The resulting system delivers both performance gains on downstream tasks and a built-in interface for tracing how different knowledge types contribute to predictions. A sympathetic reader would care because it solves the tension between transferable foundation-model learning and the need for semantic interpretability under strict privacy isolation.

Core claim

FedLAB organizes multimodal graph knowledge into typed hierarchical codebooks for modality evidence, node semantics, and topology context. It refines these trace units through federated semantic barycenter pre-training while keeping raw multimodal contents and graph structures local. Experiments on 10 benchmarks and 6 downstream tasks show improvements of up to 7.53% over state-of-the-art baselines while preserving a native semantic trace interface.

What carries the argument

typed hierarchical codebooks for modality evidence, node semantics, and topology context that serve as traceable units refined by federated semantic barycenter pre-training

If this is right

  • Enables transferable representation learning across decentralized clients for diverse graph-centric and modality-centric tasks.
  • Delivers up to 7.53% higher accuracy than prior federated baselines on the evaluated benchmarks.
  • Exposes how modality evidence, node semantics, and topology context jointly support each prediction through the native trace interface.
  • Maintains strict data isolation by exchanging only refined codebook updates rather than raw contents or local graph structures.

Where Pith is reading between the lines

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

  • The same codebook organization could be tested on non-graph multimodal data such as distributed sensor streams or document collections.
  • Native traces might support post-hoc auditing of model decisions in domains that require explanation under privacy regulations.
  • Exchanging only codebook refinements could lower communication volume compared with full model or embedding synchronization in other federated settings.

Load-bearing premise

That typed hierarchical codebooks for modality evidence, node semantics, and topology context can be refined through federated semantic barycenter pre-training in a way that jointly supports predictions and exposes intrinsic semantic traceability without any sharing of raw multimodal contents or local graph structures.

What would settle it

An experiment in which the hierarchical codebook structure is replaced by standard parameter or prototype exchange and both the performance margin and the ability to trace modality-node-topology contributions disappear.

Figures

Figures reproduced from arXiv: 2606.32016 by Guoren Wang, Kairui Yang, Rong-Hua Li, Xuaner Chen, Xunkai Li, Xun Wu, Zekai Chen.

Figure 1
Figure 1. Figure 1: A traceable modality-evidence interface. Fed [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: A traceable topology-context interface. FedLAB [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Framework overview of FedLAB. The server maintains a shared backbone and typed trace codebooks, while clients [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Robustness analysis under hyper-parameter varia [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 6
Figure 6. Figure 6: Convergence comparison across communication [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗
Figure 8
Figure 8. Figure 8: Modality evidence traceability. 0 25 50 Removed (%) 0 25 50 Drop (%) Grocery 0 25 50 Removed (%) Bili Music 0 25 50 Removed (%) Bili Cartoon 0 25 50 Removed (%) Bili Food 1 5 10 Top-K 0 40 80 SCC@K (%) Toys 1 5 10 Top-K DY 1 5 10 Top-K QB 1 5 10 Top-K KU FedLAB FedMVP FedBook Random [PITH_FULL_IMAGE:figures/full_fig_p007_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Semantic traceability validation on topology trace [PITH_FULL_IMAGE:figures/full_fig_p007_9.png] view at source ↗
read the original abstract

Multimodal graph foundation models aim to learn reusable knowledge from graphs enriched with text, images, attributes, and relational topology, thereby supporting diverse graph-centric and modality-centric tasks. In practice, however, such multimodal graphs are often distributed across decentralized clients, where raw contents and local structures cannot be centrally shared due to privacy constraints. This motivates federated multimodal graph foundation learning, which requires not only transferable representation learning but also intrinsic semantic traceability under strict data isolation. Existing methods usually exchange or store knowledge through parameters, prototypes, embeddings, or compact codebooks, which support optimization and transfer but do not explicitly expose how modality evidence, node semantics, and topology context jointly support predictions. To bridge this gap, we propose FedLAB, a traceable semantic codebook framework that organizes multimodal graph knowledge into typed hierarchical codebooks for modality evidence, node semantics, and topology context. FedLAB further refines these trace units through federated semantic barycenter pre-training while keeping raw multimodal contents and graph structures local. Extensive experiments on 10 benchmarks and 6 downstream tasks show that FedLAB improves over state-of-the-art baselines by up to 7.53\%, while preserving a native semantic trace interface.

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 / 2 minor

Summary. The paper proposes FedLAB, a framework for federated multimodal graph foundation learning. It organizes multimodal graph knowledge into typed hierarchical codebooks for modality evidence, node semantics, and topology context. These codebooks are refined via federated semantic barycenter pre-training while keeping raw contents and structures local. The method claims to support both transferable predictions and intrinsic semantic traceability. Experiments across 10 benchmarks and 6 downstream tasks report improvements of up to 7.53% over state-of-the-art baselines.

Significance. If the performance gains and traceability properties hold under rigorous validation, the work could meaningfully advance privacy-preserving multimodal graph learning by combining optimization with explicit semantic tracing. The hierarchical codebook design addresses a gap in existing parameter/prototype-based federated methods by exposing joint contributions from modalities and topology.

major comments (2)
  1. [Abstract / Experiments] Abstract and experimental sections: the central performance claim of up to 7.53% improvement is reported without any description of experimental controls, statistical testing, ablation studies, number of clients, or multiple-comparison correction. This directly affects evaluability of the headline result.
  2. [Proposed Method] Method (typed hierarchical codebooks and federated semantic barycenter pre-training): the manuscript must demonstrate that the refinement process jointly supports predictions and traceability without introducing circularity or requiring raw data sharing; the current description leaves the weakest assumption unverified in detail.
minor comments (2)
  1. [Method] Notation for codebook types (modality evidence, node semantics, topology context) should be introduced with explicit symbols or diagrams for clarity.
  2. [Abstract] The abstract could briefly state the privacy model (e.g., what is communicated between clients and server) to strengthen the isolation claim.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback. We address each major comment below and note planned revisions for improved rigor and clarity.

read point-by-point responses

  1. Referee: [Abstract / Experiments] Abstract and experimental sections: the central performance claim of up to 7.53% improvement is reported without any description of experimental controls, statistical testing, ablation studies, number of clients, or multiple-comparison correction. This directly affects evaluability of the headline result.

    Authors: We agree the abstract is a high-level summary and omits these details. The experimental section describes the 10 benchmarks and 6 tasks with SOTA comparisons, but to strengthen evaluability we will add explicit reporting of client counts in the federated setup, standard deviations over multiple runs, statistical significance tests with multiple-comparison correction, and expanded ablations in the revision. The 7.53% is the largest observed gain across all reported experiments. revision: yes

  2. Referee: [Proposed Method] Method (typed hierarchical codebooks and federated semantic barycenter pre-training): the manuscript must demonstrate that the refinement process jointly supports predictions and traceability without introducing circularity or requiring raw data sharing; the current description leaves the weakest assumption unverified in detail.

    Authors: Only aggregated barycenter updates are exchanged; raw multimodal contents and local graphs remain private. The typed hierarchy supplies traceability as an inspection interface separate from the prediction optimization objective. Pre-training aligns semantics for both goals without circular dependency. We will add a dedicated subsection with a formal separation argument and verification experiments in the revision. revision: partial

Circularity Check

0 steps flagged

No significant circularity identified

full rationale

The paper proposes a new methodological framework (typed hierarchical codebooks refined via federated semantic barycenter pre-training) rather than deriving headline performance metrics from quantities already fitted inside the paper. No equations, self-citations, or ansatzes are visible in the abstract or description that reduce the claimed improvements or traceability interface to inputs by construction. The derivation is therefore self-contained as an architectural proposal supported by external benchmark experiments.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 1 invented entities

Abstract-only review; no equations, hyperparameters, or background assumptions are stated, so the ledger records only the high-level invented construct mentioned in the proposal.

invented entities (1)
  • typed hierarchical codebooks no independent evidence
    purpose: Organize multimodal graph knowledge into modality evidence, node semantics, and topology context units that support traceable predictions
    Central new construct introduced to provide intrinsic semantic traceability under data isolation

pith-pipeline@v0.9.1-grok · 5760 in / 1390 out tokens · 28144 ms · 2026-07-01T06:08:16.628092+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

243 extracted references · 52 canonical work pages · 19 internal anchors

  1. [1]

    Artificial intelligence and statistics , pages=

    Communication-efficient learning of deep networks from decentralized data , author=. Artificial intelligence and statistics , pages=. 2017 , organization=

    2017

  2. [2]

    IEEE Transactions on Parallel and Distributed Systems , volume=

    Fedgraph: Federated graph learning with intelligent sampling , author=. IEEE Transactions on Parallel and Distributed Systems , volume=. 2021 , publisher=

    2021

  3. [4]

    Subgraph federated learning over heterogeneous graphs , author=. Proc. FedGraph (CIKM) , pages=

  4. [5]

    Advances in neural information processing systems , volume=

    FedGCN: Convergence-communication tradeoffs in federated training of graph convolutional networks , author=. Advances in neural information processing systems , volume=

  5. [7]

    Proceedings of the 28th ACM SIGKDD conference on knowledge discovery and data mining , pages=

    Graphmae: Self-supervised masked graph autoencoders , author=. Proceedings of the 28th ACM SIGKDD conference on knowledge discovery and data mining , pages=

  6. [8]

    International conference on learning representations , volume=

    One for all: Towards training one graph model for all classification tasks , author=. International conference on learning representations , volume=

  7. [9]

    Advances in Neural Information Processing Systems , volume=

    Towards effective federated graph foundation model via mitigating knowledge entanglement , author=. Advances in Neural Information Processing Systems , volume=

  8. [11]

    Proceedings of the ACM on Web Conference 2025 , pages=

    Graphclip: Enhancing transferability in graph foundation models for text-attributed graphs , author=. Proceedings of the ACM on Web Conference 2025 , pages=

    2025

  9. [12]

    Proceedings of the ACM on Web Conference 2025 , pages=

    Unigraph2: Learning a unified embedding space to bind multimodal graphs , author=. Proceedings of the ACM on Web Conference 2025 , pages=

    2025

  10. [13]

    Advances in neural information processing systems , volume=

    Gnnexplainer: Generating explanations for graph neural networks , author=. Advances in neural information processing systems , volume=

  11. [14]

    Advances in neural information processing systems , volume=

    Parameterized explainer for graph neural network , author=. Advances in neural information processing systems , volume=

  12. [15]

    IEEE Transactions on Knowledge and Data Engineering , volume=

    Graphlime: Local interpretable model explanations for graph neural networks , author=. IEEE Transactions on Knowledge and Data Engineering , volume=. 2022 , publisher=

    2022

  13. [16]

    2012 , publisher=

    Clifford algebra to geometric calculus: a unified language for mathematics and physics , author=. 2012 , publisher=

    2012

  14. [17]

    Clifford algebras and their applications in mathematical physics , pages=

    Clifford algebras and spinors , author=. Clifford algebras and their applications in mathematical physics , pages=. 2001 , publisher=

    2001

  15. [18]

    2009 , publisher=

    Geometric algebra for computer science (revised edition): An object-oriented approach to geometry , author=. 2009 , publisher=

    2009

  16. [19]

    International Conference on Machine Learning, ICML , year=

    Learning transferable visual models from natural language supervision , author=. International Conference on Machine Learning, ICML , year=

  17. [20]

    OPT: Open Pre-trained Transformer Language Models

    Opt: Open pre-trained transformer language models , author=. arXiv preprint arXiv:2205.01068 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  18. [21]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR , year=

    High-resolution image synthesis with latent diffusion models , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR , year=

  19. [22]

    Proceedings of the Conference on Empirical Methods in Natural Language Processing and the International Joint Conference on Natural Language Processing, EMNLP-IJCNLP , year=

    Justifying recommendations using distantly-labeled reviews and fine-grained aspects , author=. Proceedings of the Conference on Empirical Methods in Natural Language Processing and the International Joint Conference on Natural Language Processing, EMNLP-IJCNLP , year=

  20. [23]

    Advances in Neural Information Processing Systems, NeurIPS, Datasets and Benchmarks Track, NeurIPS DB Track , year=

    RedCaps: Web-curated image-text data created by the people, for the people , author=. Advances in Neural Information Processing Systems, NeurIPS, Datasets and Benchmarks Track, NeurIPS DB Track , year=

  21. [24]

    Bridging Language and Items for Retrieval and Recommendation: Benchmarking LLMs as Semantic Encoders

    Bridging language and items for retrieval and recommendation , author=. arXiv preprint arXiv:2403.03952 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  22. [25]

    Proceedings of the European Conference on Computer Vision Workshops, ECCV , year=

    How to read paintings: semantic art understanding with multi-modal retrieval , author=. Proceedings of the European Conference on Computer Vision Workshops, ECCV , year=

  23. [26]

    Proceedings of the IEEE International Conference on Computer Vision, ICCV , year=

    Flickr30k entities: Collecting region-to-phrase correspondences for richer image-to-sentence models , author=. Proceedings of the IEEE International Conference on Computer Vision, ICCV , year=

  24. [27]

    Proceedings of the ACM Conference on Recommender Systems, RecSys , year=

    Item recommendation on monotonic behavior chains , author=. Proceedings of the ACM Conference on Recommender Systems, RecSys , year=

  25. [28]

    Proceedings of the Annual Meeting of the Association for Computational Linguistics, ACL , year=

    Fine-grained spoiler detection from large-scale review corpora , author=. Proceedings of the Annual Meeting of the Association for Computational Linguistics, ACL , year=

  26. [29]

    arXiv preprint arXiv:2310.07478 , year=

    Multimodal graph learning for generative tasks , author=. arXiv preprint arXiv:2310.07478 , year=

    work page arXiv

  27. [30]

    Proceedings of the AAAI Conference on Artificial Intelligence, AAAI , year=

    Modality-Independent Graph Neural Networks with Global Transformers for Multimodal Recommendation , author=. Proceedings of the AAAI Conference on Artificial Intelligence, AAAI , year=

  28. [31]

    Proceedings of the AAAI Conference on Artificial Intelligence, AAAI , year=

    NTSFormer: A Self-Teaching Graph Transformer for Multimodal Isolated Cold-Start Node Classification , author=. Proceedings of the AAAI Conference on Artificial Intelligence, AAAI , year=

  29. [32]

    Proceedings of the Annual Meeting of the Association for Computational Linguistics, ACL , year=

    MMGCN: Multimodal Fusion via Deep Graph Convolution Network for Emotion Recognition in Conversation , author=. Proceedings of the Annual Meeting of the Association for Computational Linguistics, ACL , year=

  30. [33]

    arXiv preprint arXiv:2506.10282 , year=

    Graph-MLLM: Harnessing Multimodal Large Language Models for Multimodal Graph Learning , author=. arXiv preprint arXiv:2506.10282 , year=

    work page arXiv

  31. [34]

    MLaGA: Multimodal Large Language and Graph Assistant

    Mlaga: Multimodal large language and graph assistant , author=. arXiv preprint arXiv:2506.02568 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  32. [35]

    arXiv preprint arXiv:2511.20222 , year=

    Decoupling and Damping: Structurally-Regularized Gradient Matching for Multimodal Graph Condensation , author=. arXiv preprint arXiv:2511.20222 , year=

    work page arXiv

  33. [36]

    Proceedings of the Computer Vision and Pattern Recognition Conference, CVPR , pages=

    Graphgpt-o: Synergistic multimodal comprehension and generation on graphs , author=. Proceedings of the Computer Vision and Pattern Recognition Conference, CVPR , pages=

  34. [37]

    Proceedings of the Computer Vision and Pattern Recognition Conference, CVPR , year=

    Mosaic of modalities: A comprehensive benchmark for multimodal graph learning , author=. Proceedings of the Computer Vision and Pattern Recognition Conference, CVPR , year=

  35. [38]

    Proceedings of the International Conference on Machine Learning, ICML , year=

    Graph4MM: Weaving multimodal learning with structural information , author=. Proceedings of the International Conference on Machine Learning, ICML , year=

  36. [39]

    Information Processing & Management , volume=

    Mgat: Multimodal graph attention network for recommendation , author=. Information Processing & Management , volume=. 2020 , publisher=

    2020

  37. [40]

    Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD , year=

    When graph meets multimodal: benchmarking and meditating on multimodal attributed graph learning , author=. Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD , year=

  38. [41]

    AI for Supply Chain: Today and Future

    C-MAG: Cascade Multimodal Attributed Graphs for Supply Chain Link Prediction , author=. Proceedings of the 1st Workshop on "AI for Supply Chain: Today and Future" in ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD , year=

  39. [42]

    Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD , year=

    Cross-Contrastive Clustering for Multimodal Attributed Graphs with Dual Graph Filtering , author=. Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD , year=

  40. [43]

    Proceedings of the ACM International Conference on Multimedia, MM , year=

    Disentangling homophily and heterophily in multimodal graph clustering , author=. Proceedings of the ACM International Conference on Multimedia, MM , year=

  41. [44]

    Advances in Neural Information Processing Systems, NeurIPS , year=

    Instructg2i: Synthesizing images from multimodal attributed graphs , author=. Advances in Neural Information Processing Systems, NeurIPS , year=

  42. [45]

    arXiv preprint arXiv:2506.09738 , year=

    Towards Multi-modal Graph Large Language Model , author=. arXiv preprint arXiv:2506.09738 , year=

    work page arXiv

  43. [46]

    Proceedings of the AAAI Conference on Artificial Intelligence, AAAI , year=

    Multimodal graph neural architecture search under distribution shifts , author=. Proceedings of the AAAI Conference on Artificial Intelligence, AAAI , year=

  44. [47]

    International Conference on Machine Learning, ICML , year=

    Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models , author=. International Conference on Machine Learning, ICML , year=

  45. [48]

    Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, KDD , pages=

    Optuna: A next-generation hyperparameter optimization framework , author=. Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, KDD , pages=

  46. [49]

    Nature Communications , volume=

    Biomedical knowledge graph learning for drug repurposing by extending guilt-by-association to multiple layers , author=. Nature Communications , volume=. 2023 , publisher=

    2023

  47. [50]

    IEEE Journal of Translational Engineering in Health and Medicine , year=

    A Graph Convolutional Network Based on Univariate Neurodegeneration Biomarker for Alzheimer’s Disease Diagnosis , author=. IEEE Journal of Translational Engineering in Health and Medicine , year=

  48. [51]

    Bioinformatics , volume=

    Similarity measures-based graph co-contrastive learning for drug--disease association prediction , author=. Bioinformatics , volume=. 2023 , publisher=

    2023

  49. [52]

    Intelligent Systems with Applications , pages=

    Predicting Systemic Risk in Financial Systems Using Deep Graph Learning , author=. Intelligent Systems with Applications , pages=. 2023 , publisher=

    2023

  50. [53]

    Pacific-Asia Conference on Knowledge Discovery and Data Mining , pages=

    Anti-Money Laundering in Cryptocurrency via Multi-Relational Graph Neural Network , author=. Pacific-Asia Conference on Knowledge Discovery and Data Mining , pages=. 2023 , organization=

    2023

  51. [54]

    2023 IEEE 6th Information Technology, Networking, Electronic and Automation Control Conference , volume=

    Default Risk Assessment of Internet Financial Enterprises Based on Graph Neural Network , author=. 2023 IEEE 6th Information Technology, Networking, Electronic and Automation Control Conference , volume=. 2023 , organization=

    2023

  52. [55]

    Advances in neural information processing systems, NeurIPS , year=

    Link prediction based on graph neural networks , author=. Advances in neural information processing systems, NeurIPS , year=

  53. [56]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , year=

    Line graph neural networks for link prediction , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , year=

  54. [57]

    Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD , pages =

    Besta, Maciej and Grob, Raphael and Miglioli, Cesare and Bernold, Nicola and Kwasniewski, Grzegorz and Gjini, Gabriel and Kanakagiri, Raghavendra and Ashkboos, Saleh and Gianinazzi, Lukas and Dryden, Nikoli and Hoefler, Torsten , title=. Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD , pages =. 2022 , address=

    2022

  55. [58]

    Proceedings of the Sixteenth ACM International Conference on Web Search and Data Mining, WSDM , pages=

    Bring your own view: Graph neural networks for link prediction with personalized subgraph selection , author=. Proceedings of the Sixteenth ACM International Conference on Web Search and Data Mining, WSDM , pages=

  56. [59]

    arXiv preprint arXiv:1911.05954 , year=

    Hierarchical graph pooling with structure learning , author=. arXiv preprint arXiv:1911.05954 , year=

    work page arXiv 1911

  57. [60]

    Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, KDD , pages=

    Graph convolutional networks with eigenpooling , author=. Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, KDD , pages=

  58. [61]

    International Conference on Machine Learning, ICML , pages=

    A new perspective on the effects of spectrum in graph neural networks , author=. International Conference on Machine Learning, ICML , pages=. 2022 , organization=

    2022

  59. [62]

    Advances in neural information processing systems, NeurIPS , year=

    Template based graph neural network with optimal transport distances , author=. Advances in neural information processing systems, NeurIPS , year=

  60. [63]

    International Conference on Learning Representations, ICLR , year =

    Adversarial Attacks on Graph Neural Networks via Meta Learning , author=. International Conference on Learning Representations, ICLR , year =

  61. [64]

    and Salakhutdinov, Ruslan , title =

    Yang, Zhilin and Cohen, William W. and Salakhutdinov, Ruslan , title =. 2016 , booktitle=

    2016

  62. [65]

    Pitfalls of Graph Neural Network Evaluation

    Pitfalls of graph neural network evaluation , author=. arXiv preprint arXiv:1811.05868 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  63. [66]

    Journal of Complex Networks , volume =

    Rozemberczki, Benedek and Allen, Carl and Sarkar, Rik , title = ". Journal of Complex Networks , volume =. 2021 , month =

    2021

  64. [67]

    Advances in neural information processing systems, NeurIPS , volume=

    Open graph benchmark: Datasets for machine learning on graphs , author=. Advances in neural information processing systems, NeurIPS , volume=

  65. [68]

    arXiv preprint arXiv:2302.13522 , year=

    IGB: Addressing The Gaps In Labeling, Features, Heterogeneity, and Size of Public Graph Datasets for Deep Learning Research , author=. arXiv preprint arXiv:2302.13522 , year=

    work page arXiv

  66. [69]

    International conference on learning representations, ICLR , year=

    Graphsaint: Graph sampling based inductive learning method , author=. International conference on learning representations, ICLR , year=

  67. [70]

    Journal of statistical mechanics: theory and experiment , volume=

    Fast unfolding of communities in large networks , author=. Journal of statistical mechanics: theory and experiment , volume=. 2008 , publisher=

    2008

  68. [71]

    SIAM Journal on scientific Computing , volume=

    A fast and high quality multilevel scheme for partitioning irregular graphs , author=. SIAM Journal on scientific Computing , volume=. 1998 , publisher=

    1998

  69. [72]

    IEEE transactions on neural networks and learning systems , volume=

    A comprehensive survey on graph neural networks , author=. IEEE transactions on neural networks and learning systems , volume=. 2020 , publisher=

    2020

  70. [73]

    ACM Transactions on Intelligent Systems and Technology (TIST) , volume=

    Graph Neural Networks: Taxonomy, Advances, and Trends , author=. ACM Transactions on Intelligent Systems and Technology (TIST) , volume=. 2022 , publisher=

    2022

  71. [74]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

    Graph neural networks in network neuroscience , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=. 2022 , publisher=

    2022

  72. [75]

    IEEE Transactions on Neural Networks and Learning Systems , year=

    Graph-based semi-supervised learning: A comprehensive review , author=. IEEE Transactions on Neural Networks and Learning Systems , year=

  73. [76]

    International Conference on Learning Representations, ICLR , year=

    Is homophily a necessity for graph neural networks? , author=. International Conference on Learning Representations, ICLR , year=

  74. [77]

    Advances in neural information processing systems, NeurIPS , year=

    Revisiting heterophily for graph neural networks , author=. Advances in neural information processing systems, NeurIPS , year=

  75. [78]

    Graph Neural Networks for Graphs with Heterophily: A Survey

    Graph neural networks for graphs with heterophily: A survey , author=. arXiv preprint arXiv:2202.07082 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  76. [79]

    International Conference on Learning Representations, ICLR , year=

    A critical look at the evaluation of GNNs under heterophily: are we really making progress? , author=. International Conference on Learning Representations, ICLR , year=

  77. [80]

    2022 , journal=

    Zhang, Wentao and Yin, Ziqi and Sheng, Zeang and Li, Yang and Ouyang, Wen and Li, Xiaosen and Tao, Yangyu and Yang, Zhi and Cui, Bin , title =. 2022 , journal=

    2022

  78. [81]

    IEEE Transactions on Knowledge and Data Engineering , year=

    Adaptive hypergraph auto-encoder for relational data clustering , author=. IEEE Transactions on Knowledge and Data Engineering , year=

  79. [82]

    International Conference on Learning Representations, ICLR , year =

    Semi-supervised classification with graph convolutional networks , author=. International Conference on Learning Representations, ICLR , year =

  80. [83]

    Advances in neural information processing systems, NeurIPS , year=

    Inductive representation learning on large graphs , author=. Advances in neural information processing systems, NeurIPS , year=

Showing first 80 references.