| Higher-order Graph Convolutional Network with Flower-Petals Laplacians on Simplicial Complexes | ✓ Link | 94.99±0.65 | 5-HiGCN | 2023-09-22 |
| Graph Neural Reaction Diffusion Models | | 93.72 ± 4.59 | RDGNN-I | 2024-06-16 |
| Diffusion-Jump GNNs: Homophiliation via Learnable Metric Filters | ✓ Link | 92.54±3.70 | DJ-GNN | 2023-06-29 |
| GraphRARE: Reinforcement Learning Enhanced Graph Neural Network with Relative Entropy | | 90.00±2.97 | H2GCN-RARE (λ=1.0) | 2023-12-15 |
| Neural Sheaf Diffusion: A Topological Perspective on Heterophily and Oversmoothing in GNNs | ✓ Link | 89.41 ± 4.74 | O(d)-NSD | 2022-02-09 |
| Neural Sheaf Diffusion: A Topological Perspective on Heterophily and Oversmoothing in GNNs | ✓ Link | 89.21 ± 3.84 | Gen-NSD | 2022-02-09 |
| Sign is Not a Remedy: Multiset-to-Multiset Message Passing for Learning on Heterophilic Graphs | ✓ Link | 89.01 ± 4.1 | M2M-GNN | 2024-05-31 |
| Heterophilous Distribution Propagation for Graph Neural Networks | | 88.82 ± 3.40 | HDP | 2024-05-31 |
| The Heterophilic Snowflake Hypothesis: Training and Empowering GNNs for Heterophilic Graphs | ✓ Link | 88.77 | MGNN + Hetero-S (6 layers) | 2024-06-18 |
| Sheaf Neural Networks with Connection Laplacians | ✓ Link | 88.73±4.47 | Conn-NSD | 2022-06-17 |
| Neural Sheaf Diffusion: A Topological Perspective on Heterophily and Oversmoothing in GNNs | ✓ Link | 88.63 ± 2.75 | Diag-NSD | 2022-02-09 |
| Self-attention Dual Embedding for Graphs with Heterophily | | 88.63±4.54 | SADE-GCN | 2023-05-28 |
| Improving Graph Neural Networks with Simple Architecture Design | ✓ Link | 88.43±3.22 | FSGNN (3-hop) | 2021-05-17 |
| Revisiting Heterophily For Graph Neural Networks | ✓ Link | 88.43 ± 3.22 | ACM-GCN | 2022-10-14 |
| Revisiting Heterophily For Graph Neural Networks | ✓ Link | 88.43 ± 2.39 | ACM-GCN+ | 2022-10-14 |
| Revisiting Heterophily For Graph Neural Networks | ✓ Link | 88.43 ± 3.66 | ACMII-GCN++ | 2022-10-14 |
| Refining Latent Homophilic Structures over Heterophilic Graphs for Robust Graph Convolution Networks | | 88.32±2.3 | LHS | 2023-12-27 |
| Revisiting Heterophily For Graph Neural Networks | ✓ Link | 88.24 ± 3.16 | ACM-GCN++ | 2022-10-14 |
| Finding Global Homophily in Graph Neural Networks When Meeting Heterophily | ✓ Link | 88.04±3.22 | GloGNN++ | 2022-05-15 |
| Revisiting Heterophily For Graph Neural Networks | ✓ Link | 88.04 ± 3.66 | ACMII-GCN+ | 2022-10-14 |
| Ordered GNN: Ordering Message Passing to Deal with Heterophily and Over-smoothing | ✓ Link | 88.04±3.63 | Ordered GNN | 2023-02-03 |
| UniG-Encoder: A Universal Feature Encoder for Graph and Hypergraph Node Classification | ✓ Link | 88.03±4.42 | UniG-Encoder | 2023-08-03 |
| Improving Graph Neural Networks by Learning Continuous Edge Directions | ✓ Link | 87.84±3.70 | CoED | 2024-10-18 |
| UniGAP: A Universal and Adaptive Graph Upsampling Approach to Mitigate Over-Smoothing in Node Classification Tasks | ✓ Link | 87.73 ± 4.8 | H2GCN + UniGAP | 2024-07-28 |
| Graph Neural Aggregation-diffusion with Metastability | | 87.7±3.7 | GRADE-GAT | 2024-03-29 |
| GCNH: A Simple Method For Representation Learning On Heterophilous Graphs | ✓ Link | 87.65±3.59 | GCNH | 2023-04-21 |
| Universal Deep GNNs: Rethinking Residual Connection in GNNs from a Path Decomposition Perspective for Preventing the Over-smoothing | | 87.64±3.74 | UDGNN (GCN) | 2022-05-30 |
| Revisiting Heterophily For Graph Neural Networks | ✓ Link | 87.45 ± 3.74 | ACMII-GCN | 2022-10-14 |
| Transfer Entropy in Graph Convolutional Neural Networks | ✓ Link | 87.45 ± 3.70 | TE-GCNN | 2024-06-08 |
| Non-Local Graph Neural Networks | ✓ Link | 87.3 ± 4.3 | NLMLP | 2020-05-29 |
| Finding Global Homophily in Graph Neural Networks When Meeting Heterophily | ✓ Link | 87.06±3.53 | GloGNN | 2022-05-15 |
| Beyond Low-frequency Information in Graph Convolutional Networks | ✓ Link | 86.98 ± 3.78 | WRGAT | 2021-01-04 |
| Label-Wise Graph Convolutional Network for Heterophilic Graphs | ✓ Link | 86.9±2.2 | LW-GCN | 2021-10-15 |
| Two Sides of the Same Coin: Heterophily and Oversmoothing in Graph Convolutional Neural Networks | ✓ Link | 86.86 ± 3.29 | GGCN | 2021-02-12 |
| Simple Truncated SVD based Model for Node Classification on Heterophilic Graphs | | 86.67±4.22 | HLP Concat | 2021-06-24 |
| CN-Motifs Perceptive Graph Neural Networks | | 86.63 ± 3.57 | CNMPGNN | 2021-11-15 |
| Revisiting Heterophily For Graph Neural Networks | ✓ Link | 86.47 ± 3.77 | ACM-SGC-1 | 2022-10-14 |
| Revisiting Heterophily For Graph Neural Networks | ✓ Link | 86.47 ± 3.77 | ACM-SGC-2 | 2022-10-14 |
| FDGATII : Fast Dynamic Graph Attention with Initial Residual and Identity Mapping | ✓ Link | 86.2745 | FDGATII | 2021-10-21 |
| Mamba-Based Graph Convolutional Networks: Tackling Over-smoothing with Selective State Space | | 86.27±2.16 | MbaGCN | 2025-01-26 |
| Learn from Heterophily: Heterophilous Information-enhanced Graph Neural Network | ✓ Link | 85.88 ± 3.18 | HiGNN | 2024-03-26 |
| CAT: A Causally Graph Attention Network for Trimming Heterophilic Graph | ✓ Link | 85.6±2.1 | CATv3-sup | 2023-12-14 |
| Tree Decomposed Graph Neural Network | ✓ Link | 85.57 ± 3.78 (0, 3-5) | TDGNN-w | 2021-08-25 |
| Make Heterophily Graphs Better Fit GNN: A Graph Rewiring Approach | | 85.01±5.51 | H2GCN DHGR | 2022-09-17 |
| Beyond Homophily in Graph Neural Networks: Current Limitations and Effective Designs | ✓ Link | 84.31 ± 3.70 | H2GCN-1 | 2020-06-20 |
| Beyond Homophily with Graph Echo State Networks | | 83.3±3.8 | Graph ESN | 2022-10-27 |
| Beyond Homophily in Graph Neural Networks: Current Limitations and Effective Designs | ✓ Link | 83.14 ± 4.26 | H2GCN-2 | 2020-06-20 |
| Adaptive Universal Generalized PageRank Graph Neural Network | ✓ Link | 82.55 ± 6.23 | GPRGCN | 2020-06-14 |
| Transitivity-Preserving Graph Representation Learning for Bridging Local Connectivity and Role-based Similarity | ✓ Link | 81.6 ±8.24 | UGT | 2023-08-18 |
| Breaking the Entanglement of Homophily and Heterophily in Semi-supervised Node Classification | | 81.6±3.5 | ADPA | 2023-12-07 |
| Simple and Deep Graph Convolutional Networks | ✓ Link | 80.39 ± 3.40 | GCNII | 2020-07-04 |
| DeltaGNN: Graph Neural Network with Information Flow Control | ✓ Link | 80.00±0.88 | DeltaGNN linear | 2025-01-10 |
| Beyond Low-frequency Information in Graph Convolutional Networks | ✓ Link | 79.61 ± 1.58 | FAGCN | 2021-01-04 |
| DiffWire: Inductive Graph Rewiring via the Lovász Bound | ✓ Link | 79.05 | CT-Layer | 2022-06-15 |
| MixHop: Higher-Order Graph Convolutional Architectures via Sparsified Neighborhood Mixing | ✓ Link | 75.88 ± 4.90 | MixHop | 2019-04-30 |
| Large Scale Learning on Non-Homophilous Graphs: New Benchmarks and Strong Simple Methods | ✓ Link | 75.49 ± 5.72 | LINKX | 2021-10-27 |
| DiffWire: Inductive Graph Rewiring via the Lovász Bound | ✓ Link | 69.25 | CT-Layer (PE) | 2022-06-15 |
| Geom-GCN: Geometric Graph Convolutional Networks | ✓ Link | 64.12 | Geom-GCN-P | 2020-02-13 |
| Non-Local Graph Neural Networks | ✓ Link | 60.2 ± 5.3 | NLGCN | 2020-05-29 |
| Geom-GCN: Geometric Graph Convolutional Networks | ✓ Link | 58.24 | Geom-GCN-I | 2020-02-13 |
| Non-Local Graph Neural Networks | ✓ Link | 56.9 ± 7.3 | NLGAT | 2020-05-29 |
| Geom-GCN: Geometric Graph Convolutional Networks | ✓ Link | 56.67 | Geom-GCN-S | 2020-02-13 |
| Understanding over-squashing and bottlenecks on graphs via curvature | ✓ Link | 55.51±0.27 | SDRF | 2021-11-29 |