Heterophilic Graphs
Chameleon
Structure
1Graphs
2,277Nodes
36,101Edges
5Classes
2325Node Feature Dim
0.23Homophily
Coverage
404Results
1Variants
57Papers
- Graph Transductive Sharpening: Leveraging Unlabeled Predictions in Node Classification Graph Transductive Sharpening: Leveraging Unlabeled Predictions in Node Classification
- Latent-Hysteresis Graph ODEs: Modeling Coupled Topology-Feature Evolution via Continuous Phase Transitions Latent-Hysteresis Graph ODEs: Modeling Coupled Topology-Feature Evolution via Continuous Phase Transitions
- ICLR 2026: Learning Posterior Predictive Distributions for Node Classification from Synthetic Graph Priors Learning Posterior Predictive Distributions for Node Classification from Synthetic Graph Priors ICLR 2026
- WWW 2026: Frequency-Corrupt Based Graph Self-Supervised Learning Frequency-Corrupt Based Graph Self-Supervised Learning WWW 2026
- Geometric Evolution Graph Convolutional Networks: Enhancing Graph Representation Learning via Ricci Flow Geometric Evolution Graph Convolutional Networks: Enhancing Graph Representation Learning via Ricci Flow
- $P^2$GNN: Two Prototype Sets to boost GNN Performance $P^2$GNN: Two Prototype Sets to boost GNN Performance
- SPGCL: Simple yet Powerful Graph Contrastive Learning via SVD-Guided Structural Perturbation SPGCL: Simple yet Powerful Graph Contrastive Learning via SVD-Guided Structural Perturbation
- GADPN: Graph Adaptive Denoising and Perturbation Networks via Singular Value Decomposition GADPN: Graph Adaptive Denoising and Perturbation Networks via Singular Value Decomposition
- mHC-GNN: Manifold-Constrained Hyper-Connections for Graph Neural Networks mHC-GNN: Manifold-Constrained Hyper-Connections for Graph Neural Networks
- LoG 2025: Can TabPFN Compete with GNNs for Node Classification via Graph Tabularization? Can TabPFN Compete with GNNs for Node Classification via Graph Tabularization? LoG 2025
- Gauge-Equivariant Graph Networks via Self-Interference Cancellation Gauge-Equivariant Graph Networks via Self-Interference Cancellation
- NeurIPS 2025: One Prompt Fits All: Universal Graph Adaptation for Pretrained Models One Prompt Fits All: Universal Graph Adaptation for Pretrained Models NeurIPS 2025
- Graph Homophily Booster: Rethinking the Role of Discrete Features on Heterophilic Graphs Graph Homophily Booster: Rethinking the Role of Discrete Features on Heterophilic Graphs
- Local Virtual Nodes for Alleviating Over-Squashing in Graph Neural Networks Local Virtual Nodes for Alleviating Over-Squashing in Graph Neural Networks
- KDD 2025: Discrepancy-Aware Graph Mask Auto-Encoder Discrepancy-Aware Graph Mask Auto-Encoder KDD 2025
- Equivariance Everywhere All At Once: A Recipe for Graph Foundation Models Equivariance Everywhere All At Once: A Recipe for Graph Foundation Models
- Directed Homophily-Aware Graph Neural Network Directed Homophily-Aware Graph Neural Network
- ICML 2025: Aggregation Buffer: Revisiting DropEdge with a New Parameter Block Aggregation Buffer: Revisiting DropEdge with a New Parameter Block ICML 2025
- IJCAI 2025: What Can We Learn From MIMO Graph Convolutions? What Can We Learn From MIMO Graph Convolutions? IJCAI 2025
- IJCAI 2025: Rethinking Graph Contrastive Learning through Relative Similarity Preservation Rethinking Graph Contrastive Learning through Relative Similarity Preservation IJCAI 2025
- Efficient Learning on Large Graphs using a Densifying Regularity Lemma Efficient Learning on Large Graphs using a Densifying Regularity Lemma
- TMLR 2025: Diffusion on Graph: Augmentation of Graph Structure for Node Classification Diffusion on Graph: Augmentation of Graph Structure for Node Classification TMLR 2025
- KDD 2025: GCoT: Chain-of-Thought Prompt Learning for Graphs GCoT: Chain-of-Thought Prompt Learning for Graphs KDD 2025
- NeurIPS 2024: Graph Neural Networks Need Cluster-Normalize-Activate Modules Graph Neural Networks Need Cluster-Normalize-Activate Modules NeurIPS 2024
- Scale-aware Message Passing For Graph Node Classification Scale-aware Message Passing For Graph Node Classification
- TMLR 2024: SCNode: Spatial and Contextual Coordinates for Graph Representation Learning SCNode: Spatial and Contextual Coordinates for Graph Representation Learning TMLR 2024
- Scalable Graph Compressed Convolutions Scalable Graph Compressed Convolutions
- TMLR 2024: GraphFM: A generalist graph transformer that learns transferable representations across diverse domains GraphFM: A generalist graph transformer that learns transferable representations across diverse domains TMLR 2024
- ICLR 2024: The Effectiveness of Curvature-Based Rewiring and the Role of Hyperparameters in GNNs Revisited The Effectiveness of Curvature-Based Rewiring and the Role of Hyperparameters in GNNs Revisited ICLR 2024
- KDD 2024: The Heterophilic Snowflake Hypothesis: Training and Empowering GNNs for Heterophilic Graphs The Heterophilic Snowflake Hypothesis: Training and Empowering GNNs for Heterophilic Graphs KDD 2024
- ICLR 2024: Node Identifiers: Compact, Discrete Representations for Efficient Graph Learning Node Identifiers: Compact, Discrete Representations for Efficient Graph Learning ICLR 2024
- ICLR 2024: Unleashing the Potential of Fractional Calculus in Graph Neural Networks with FROND Unleashing the Potential of Fractional Calculus in Graph Neural Networks with FROND ICLR 2024
- Graph Neural Aggregation-diffusion with Metastability Graph Neural Aggregation-diffusion with Metastability
- ICML 2024: Understanding Heterophily for Graph Neural Networks Understanding Heterophily for Graph Neural Networks ICML 2024
- Breaking the Entanglement of Homophily and Heterophily in Semi-supervised Node Classification Breaking the Entanglement of Homophily and Heterophily in Semi-supervised Node Classification
- DeepRicci: Self-supervised Graph Structure-Feature Co-Refinement for Alleviating Over-squashing DeepRicci: Self-supervised Graph Structure-Feature Co-Refinement for Alleviating Over-squashing
- Spectral-Aware Augmentation for Enhanced Graph Representation Learning Spectral-Aware Augmentation for Enhanced Graph Representation Learning
- AAAI 2023: Feature Transportation Improves Graph Neural Networks Feature Transportation Improves Graph Neural Networks AAAI 2023
- Shared Growth of Graph Neural Networks via Prompted Free-direction Knowledge Distillation Shared Growth of Graph Neural Networks via Prompted Free-direction Knowledge Distillation
- IJCAI 2023: Causal-Based Supervision of Attention in Graph Neural Network: A Better and Simpler Choice towards Powerful Attention Causal-Based Supervision of Attention in Graph Neural Network: A Better and Simpler Choice towards Powerful Attention IJCAI 2023
- ICML 2023: Feature Expansion for Graph Neural Networks Feature Expansion for Graph Neural Networks ICML 2023
- ICML 2023: Leveraging Label Non-Uniformity for Node Classification in Graph Neural Networks Leveraging Label Non-Uniformity for Node Classification in Graph Neural Networks ICML 2023
- LON-GNN: Spectral GNNs with Learnable Orthonormal Basis LON-GNN: Spectral GNNs with Learnable Orthonormal Basis
- ICML 2022: TIDE: Time Derivative Diffusion for Deep Learning on Graphs TIDE: Time Derivative Diffusion for Deep Learning on Graphs ICML 2022
- ICML 2022: Revisiting Over-smoothing and Over-squashing using Ollivier's Ricci Curvature Revisiting Over-smoothing and Over-squashing using Ollivier's Ricci Curvature ICML 2022
- AAAI 2022: Beyond Smoothing: Unsupervised Graph Representation Learning with Edge Heterophily Discriminating Beyond Smoothing: Unsupervised Graph Representation Learning with Edge Heterophily Discriminating AAAI 2022
- From Node Interaction to Hop Interaction: New Effective and Scalable Graph Learning Paradigm From Node Interaction to Hop Interaction: New Effective and Scalable Graph Learning Paradigm
- NeurIPS 2022: Hierarchical Graph Transformer with Adaptive Node Sampling Hierarchical Graph Transformer with Adaptive Node Sampling NeurIPS 2022
- ICLR 2022: Gradient Gating for Deep Multi-Rate Learning on Graphs Gradient Gating for Deep Multi-Rate Learning on Graphs ICLR 2022
- GPNet: Simplifying Graph Neural Networks via Multi-channel Geometric Polynomials GPNet: Simplifying Graph Neural Networks via Multi-channel Geometric Polynomials
- ICML 2022: Optimization-Induced Graph Implicit Nonlinear Diffusion Optimization-Induced Graph Implicit Nonlinear Diffusion ICML 2022
- AAAI 2022: Restructuring Graphs for Higher Homophily via Adaptive Spectral Clustering Restructuring Graphs for Higher Homophily via Adaptive Spectral Clustering AAAI 2022
- Discovering the Representation Bottleneck of Graph Neural Networks Discovering the Representation Bottleneck of Graph Neural Networks
- Augmentation-Free Graph Contrastive Learning with Performance Guarantee Augmentation-Free Graph Contrastive Learning with Performance Guarantee
- AAAI 2021: Block Modeling-Guided Graph Convolutional Neural Networks Block Modeling-Guided Graph Convolutional Neural Networks AAAI 2021
- Simplifying approach to Node Classification in Graph Neural Networks Simplifying approach to Node Classification in Graph Neural Networks
- Non-Local Graph Neural Networks Non-Local Graph Neural Networks
5-class node classification Wikipedia Chameleon page node classification (5 classes, heterophilic). Uses the Geom-GCN/PyG node features and graph. Evaluated by Accuracy.
Dataset details
Task
Wikipedia Chameleon page node classification (5 classes, heterophilic). The primary catalog metric is Accuracy.
Data
Wikipedia Chameleon page node classification (5 classes, heterophilic). The cataloged artifact reports 1 graphs, 5 target classes or tasks.
Features
Features and edges follow the Geom-GCN/PyG artifact: WebKB uses webpage bag-of-words and hyperlinks, Actor uses film-page keywords and co-occurrence links, and WikipediaNetwork uses page features and hyperlinks.
Splits and evaluation
The catalog records these protocols or variants: Standard split, Filtered split. Evaluation uses Accuracy (higher is better). Exact masks or folds must come from the cited release.
Quirks and pitfalls
These small or repurposed graphs have contested preprocessing and split conventions. Results across Geom-GCN, PyG, and corrected heterophily artifacts are not automatically comparable; class imbalance and high variance require multiple splits.
Sources
Geom-GCN 10 fixed splits.
SOTA Progression
new SOTA
matches SOTA at that time
other proposed result
other baseline result
Leaderboard
| # | Model | Accuracy ↑ | Accuracy (micro F1 per L1108) | F1 Score | Paper |
|---|---|---|---|---|---|
| 1 | Hetero-S (GAT 8-layer) | 0.8693 | — | — | |
| 2 | Dir-GNN+CNA | 0.8586
±
0.0180 | — | — | |
| 3 | GAT | 0.8560 | — | — | |
| 4 | SCNode | 0.8408
±
0.0155 | — | — | |
| 5 | FC-GSSL | 0.8231
±
0.0051 | — | — | |
| 6 | GraphPAE | 0.8051
±
0.0125 | — | — | |
| 7 | FaberNet | 0.8033
±
0.0119 | — | — | |
| 8 | IBG-NN | 0.8015
±
0.0113 | — | — | |
| 9 | DHGNN | 0.8011
±
0.0173 | — | — | |
| 10 | ScaleNet | 0.8010
±
0.0150 | — | — | |
| 11 | A + D + R | 0.7991 | — | — | |
| 12 | LargeScaleNet | 0.7990
±
0.0160 | — | — | |
| 13 | NDDGNN | 0.7979
±
0.0104 | — | — | |
| 14 | DirGNN | 0.7974
±
0.0140 | — | — | |
| 15 | Dir-GNN | 0.7970
±
0.0130 | — | — | |
| 16 | GraphMAE | 0.7950
±
0.0057 | — | — | |
| 17 | CoCN | 0.7935
±
0.0142 | — | — | |
| 18 | CoCN exp. | 0.7935 | — | — | |
| 19 | CoCN vanilla | 0.7917
±
0.0017 | — | — | |
| 20 | GraphMAE2 | 0.7913
±
0.0019 | — | — | |
| 21 | FSGNN (Homo/Hetero) 3-hop | 0.7893
±
0.0103 | — | — | |
| 22 | Sp^2GCL | 0.7888
±
0.0104 | — | — | |
| 23 | GPNet | 0.7861
±
0.0020 | — | — | |
| 24 | GPR-GNN | 0.7830
±
0.0060 | — | — | |
| 25 | FSGNN | 0.7827
±
0.0128 | — | — | |
| 26 | DGMAE | 0.7550
±
0.0117 | — | — | |
| 27 | CCA-SSG | 0.7500
±
0.0075 | — | — | |
| 28 | ACM-GCN | 0.7476
±
0.0220 | — | — | |
| 29 | UDGNN_GCN | 0.7453
±
0.0120 | — | — | |
| 30 | MaskGAE | 0.7450
±
0.0087 | — | — | |
| 31 | JacobiConv | 0.7420
±
0.0103 | — | — | |
| 32 | JKNet | 0.7419 | — | — | |
| 33 | FSGNN (Homo/Hetero) 8-hop | 0.7394
±
0.0202 | — | — | |
| 34 | MVGRL | 0.7319
±
0.0042 | — | — | |
| 35 | LON-GNN | 0.7300
±
0.0220 | — | — | |
| 36 | HGODE | 0.7256
±
0.0124 | — | — | |
| 37 | AUG-MAE | 0.7219
±
0.0155 | — | — | |
| 38 | NWR-GAE | 0.7204
±
0.0259 | — | — | |
| 39 | FROND | 0.7162
±
0.0161 | — | — | |
| 40 | AERO-GNN | 0.7158
±
0.0240 | — | — | |
| 41 | F-GREAD-BS | 0.7145
±
0.0198 | — | — | |
| 42 | G2-GraphSAGE | 0.7140
±
0.0238 | — | — | |
| 43 | G^2 | 0.7140
±
0.0238 | — | — | |
| 44 | GREAD-BS | 0.7140
±
0.0130 | — | — | |
| 45 | VR-GNN | 0.7121
±
0.0117 | — | — | |
| 46 | HopGNN | 0.7121
±
0.0145 | — | — | |
| 47 | GloGNN | 0.7121
±
0.0180 | — | — | |
| 48 | GGCN | 0.7114
±
0.0180 | — | — | |
| 49 | GRAFF | 0.7108
±
0.0175 | — | — | |
| 50 | DoG | 0.7060
±
0.0140 | — | — |
SOTA performance
set a new SOTA at the time
matched the SOTA at the time