mlsanityData to 14 Oct 2026

Toward Data-centric Directed Graph Learning: An Entropy-driven Approach

arXiv 2505.00983 ↗First posted 2 May 2025cs.LG cs.AI cs.DBcited by 0

Not on the map. Why?

Abstract (arXiv)

The directed graph (digraph), as a generalization of undirected graphs, exhibits superior representation capability in modeling complex topology systems and has garnered considerable attention in recent years. Despite the notable efforts made by existing DiGraph Neural Networks (DiGNNs) to leverage directed edges, they still fail to comprehensively delve into the abundant data knowledge concealed in the digraphs. This data-level limitation results in model-level sub-optimal predictive performance and underscores the necessity of further exploring the potential correlations between the directed edges (topology) and node profiles (feature and labels) from a data-centric perspective, thereby empowering model-centric neural networks with stronger encoding capabilities. In this paper, we propose Entropy-driven Digraph knowlEdge distillatioN (EDEN), which can serve as a data-centric digraph learning paradigm or a model-agnostic hot-and-plug data-centric Knowledge Distillation (KD) module. The core idea is to achieve data-centric ML, guided by our proposed hierarchical encoding theory for structured data. Specifically, EDEN first utilizes directed structural measurements from a topology perspective to construct a coarse-grained Hierarchical Knowledge Tree (HKT). Subsequently, EDEN quantifies the mutual information of node profiles to refine knowledge flow in the HKT, enabling data-centric KD supervision within model training. As a general framework, EDEN can also naturally extend to undirected scenarios and demonstrate satisfactory performance. In our experiments, EDEN has been widely evaluated on 14 (di)graph datasets (homophily and heterophily) and across 4 downstream tasks. The results demonstrate that EDEN attains SOTA performance and exhibits strong improvement for prevalent (Di)GNNs.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 54 of its 83 references that are in the map. First posted on arXiv 2 May 2025.

Builds on

54 of 83 references.

References · 54

  1. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
  2. 2017Graph Attention Networkscited by 8,923
  3. 2019Optuna: A Next-generation Hyperparameter Optimization Frameworkcited by 2,501
  4. 2020Open Graph Benchmark: Datasets for Machine Learning on Graphscited by 2,163
  5. 2019Simplifying Graph Convolutional Networkscited by 1,477
  6. 2018Deep Graph Infomaxcited by 1,106
  7. 2018Pitfalls of Graph Neural Network Evaluationcited by 1,007
  8. 2018Predict then Propagate: Graph Neural Networks meet Personalized PageRankcited by 993
  9. 2020Simple and Deep Graph Convolutional Networkscited by 860
  10. 2021How Attentive are Graph Attention Networks?cited by 835
  11. 2018Link Prediction Based on Graph Neural Networkscited by 832
  12. 2019GraphSAINT: Graph Sampling Based Inductive Learning Methodcited by 618
38 more
  1. 2021Datasets: A Community Library for Natural Language Processingcited by 530
  2. 2023A critical look at the evaluation of GNNs under heterophily: Are we really making progress?cited by 343
  3. 2017Deep Gaussian Embedding of Graphs: Unsupervised Inductive Learning via Rankingcited by 335
  4. 2018Learning Word Vectors for 157 Languagescited by 280
  5. 2020Wiki-CS: A Wikipedia-Based Benchmark for Graph Neural Networkscited by 260
  6. 2020Graph Representation Learning via Graphical Mutual Information Maximizationcited by 217
  7. 2021Is Homophily a Necessity for Graph Neural Networks?cited by 182
  8. 2022Graph Neural Networks for Graphs with Heterophily: A Surveycited by 175
  9. 2022Revisiting Heterophily For Graph Neural Networkscited by 169
  10. 2022NAGphormer: A Tokenized Graph Transformer for Node Classification in Large Graphscited by 137
  11. 2021MagNet: A Neural Network for Directed Graphscited by 107
  12. 2023Edge Directionality Improves Learning on Heterophilic Graphscited by 90
  13. 2019Hierarchical Graph Pooling with Structure Learningcited by 81
  14. 2020Directed Graph Convolutional Networkcited by 73
  15. 2023Transformers Meet Directed Graphscited by 40
  16. 2023Graph Neural Networks with Learnable and Optimal Polynomial Basescited by 36
  17. 2020On Self-Distilling Graph Neural Networkcited by 35
  18. 2023SE-GSL: A General and Effective Graph Structure Learning Framework through Structural Entropy Optimizationcited by 35
  19. 2022MSGNN: A Spectral Graph Neural Network Based on a Novel Magnetic Signed Laplaciancited by 28
  20. 2022Bring Your Own View: Graph Neural Networks for Link Prediction with Personalized Subgraph Selectioncited by 26
  21. 2021A New Perspective on the Effects of Spectrum in Graph Neural Networkscited by 24
  22. 2022FreeKD: Free-direction Knowledge Distillation for Graph Neural Networkscited by 23
  23. 2022Directed Graph Auto-Encoderscited by 22
  24. 2023Data-centric Graph Learning: A Surveycited by 22
  25. 2020Iterative Graph Self-Distillationcited by 21
  26. 2022Compact Graph Structure Learning via Mutual Information Compressioncited by 20
  27. 2023HoloNets: Spectral Convolutions do extend to Directed Graphscited by 20
  28. 2023Data-Centric Learning from Unlabeled Graphs with Diffusion Modelcited by 19
  29. 2023Breaking the Entanglement of Homophily and Heterophily in Semi-supervised Node Classificationcited by 19
  30. 2021ROD: Reception-aware Online Distillation for Sparse Graphscited by 18
  31. 2022PyTorch Geometric Signed Directed: A Software Package on Graph Neural Networks for Signed and Directed Graphscited by 18
  32. 2023The Devil is in the Data: Learning Fair Graph Neural Networks via Partial Knowledge Distillationcited by 16
  33. 2024Rethinking Node-wise Propagation for Large-scale Graph Learningcited by 16
  34. 2021DIGRAC: Digraph Clustering Based on Flow Imbalancecited by 15
  35. 2023Towards Data-centric Graph Machine Learning: Review and Outlookcited by 15
  36. 2022NAFS: A Simple yet Tough-to-beat Baseline for Graph Representation Learningcited by 11
  37. 2022A Magnetic Framelet-Based Convolutional Neural Network for Directed Graphscited by 11
  38. 2024LightDiC: A Simple yet Effective Approach for Large-scale Digraph Representation Learningcited by 11

50 most cited of 54 All 54 (JSON)

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2022Geometric Knowledge Distillation: Topology Compression for Graph Neural Networks
  2. 2023Knowledge Distillation on Graphs: A Survey
  3. 2023Edge-free but Structure-aware: Prototype-Guided Knowledge Distillation from GNNs to MLPs
  4. 2026InfGraND: An Influence-Guided GNN-to-MLP Knowledge Distillation
  5. 2025Enhancing Graph Neural Networks: A Mutual Learning Approach
  6. 2025Generating Directed Graphs with Dual Attention and Asymmetric Encoding
  7. 2025State Space Models over Directed Graphs
  8. 2025TinyGraphEstimator: Adapting Lightweight Language Models for Graph Structure Inference