| Template based Graph Neural Network with Optimal Transport Distances | ✓ Link | 88.1% | | TFGW ADJ (L=2) | 2022-05-31 |
| An end-to-end attention-based approach for learning on graphs | ✓ Link | 87.835±0.644 | | ESA (Edge set attention, no positional encodings) | 2024-02-16 |
| Learning metrics for persistence-based summaries and applications for graph classification | ✓ Link | 87.2% | | WKPI-kmeans | 2019-04-27 |
| Optimal Transport for structured data with application on graphs | ✓ Link | 86.42% | | FGW wl h=4 sp | 2018-05-23 |
| Weisfeiler and Leman Go Neural: Higher-order Graph Neural Networks | ✓ Link | 86.1% | | WL-OA Kernel | 2018-10-04 |
| On Valid Optimal Assignment Kernels and Applications to Graph Classification | | 86.1% | | WL-OA | 2016-06-03 |
| Optimal Transport for structured data with application on graphs | ✓ Link | 85.82% | | FGW wl h=2 sp | 2018-05-23 |
| Wasserstein Weisfeiler-Lehman Graph Kernels | ✓ Link | 85.75% | | WWL | 2019-06-04 |
| Weisfeiler and Leman go sparse: Towards scalable higher-order graph embeddings | ✓ Link | 85.5% | | δ-2-LWL | 2019-04-02 |
| Learning Universal Adversarial Perturbations with Generative Models | ✓ Link | 85.50% | | DUGNN | 2017-08-17 |
| CIN++: Enhancing Topological Message Passing | ✓ Link | 85.3% | | CIN++ | 2023-06-06 |
| Graph Kernels: A Survey | | 85.12% | | CORE-WL | 2019-04-27 |
| Recipe for a General, Powerful, Scalable Graph Transformer | ✓ Link | 85.110±1.423 | | GraphGPS | 2022-05-25 |
| Graph Attention Networks | ✓ Link | 85.109±1.107 | | GAT | 2017-10-30 |
| Weisfeiler and Lehman Go Paths: Learning Topological Features via Path Complexes | | 85.1% | | PIN | 2023-08-13 |
| Principal Neighbourhood Aggregation for Graph Nets | ✓ Link | 84.964±1.391 | | PNA | 2020-04-12 |
| A New Perspective on the Effects of Spectrum in Graph Neural Networks | ✓ Link | 84.87% | | Norm-GN | 2021-12-14 |
| How Powerful are Graph Neural Networks? | ✓ Link | 84.818±0.936 | | GIN | 2018-10-01 |
| Propagation kernels: efficient graph kernels from propagated information | ✓ Link | 84.5% | | Propagation kernels (pk) | 2019-02-01 |
| Cell Attention Networks | ✓ Link | 84.5% | | CAN | 2022-09-16 |
| DropGNN: Random Dropouts Increase the Expressiveness of Graph Neural Networks | ✓ Link | 84.331±1.564 | | DropGIN | 2021-11-11 |
| Semi-Supervised Classification with Graph Convolutional Networks | ✓ Link | 84.185±0.644 | | GCN | 2016-09-09 |
| Gaussian-Induced Convolution for Graphs | | 84.08% | | GIC | 2018-11-11 |
| Mutual Information Maximization in Graph Neural Networks | ✓ Link | 83.85% | | sGIN | 2019-05-21 |
| Dynamic Edge-Conditioned Filters in Convolutional Neural Networks on Graphs | ✓ Link | 83.8% | | ECC (5 scores) | 2017-04-10 |
| Are Powerful Graph Neural Nets Necessary? A Dissection on Graph Classification | ✓ Link | 83.65% | | GFN | 2019-05-11 |
| Spectral Multigraph Networks for Discovering and Fusing Relationships in Molecules | ✓ Link | 83.4% | | Multigraph ChebNet | 2018-11-23 |
| Provably Powerful Graph Networks | ✓ Link | 83.19% | | PPGN | 2019-05-27 |
| Graph Capsule Convolutional Neural Networks | ✓ Link | 82.72% | | GCAPS-CNN | 2018-05-21 |
| How Powerful are Graph Neural Networks? | ✓ Link | 82.7% | | GIN-0 | 2018-10-01 |
| How Attentive are Graph Attention Networks? | ✓ Link | 82.384±1.700 | | GATv2 | 2021-05-30 |
| Improving Attention Mechanism in Graph Neural Networks via Cardinality Preservation | ✓ Link | 82.28% | | GAT-GC (f-Scaled) | 2019-07-04 |
| DAGCN: Dual Attention Graph Convolutional Networks | ✓ Link | 81.68% | | DAGCN | 2019-04-04 |
| Relation order histograms as a network embedding tool | ✓ Link | 81.63% | | NERO | 2021-06-09 |
| Are Powerful Graph Neural Nets Necessary? A Dissection on Graph Classification | ✓ Link | 81.43% | | GFN-light | 2019-05-11 |
| Variational Recurrent Neural Networks for Graph Classification | ✓ Link | 80.7% | | VRGC | 2019-02-07 |
| Graph isomorphism UNet | ✓ Link | 80.2% | | GIUNet | 2023-08-23 |
| A Fair Comparison of Graph Neural Networks for Graph Classification | ✓ Link | 80% | | GIN | 2019-12-20 |
| Pre-training Graph Neural Networks on Molecules by Using Subgraph-Conditioned Graph Information Bottleneck | ✓ Link | 79.75±0.82 | | S-CGIB | 2025-02-20 |
| Hierarchical Graph Pooling with Structure Learning | ✓ Link | 78.45% | | HGP-SL | 2019-11-14 |
| Capsule Graph Neural Network | ✓ Link | 78.35% | | CapsGNN | 2019-05-01 |
| Transitivity-Preserving Graph Representation Learning for Bridging Local Connectivity and Role-based Similarity | ✓ Link | 77.55 ±0.16% | | UGT | 2023-08-18 |
| Strengthening structural baselines for graph classification using Local Topological Profile | ✓ Link | 77.1 ± 3.7 | 77.1 ± 3.7 | Local Topological Profile (LTP) | 2023-05-01 |
| Do Transformers Really Perform Bad for Graph Representation? | ✓ Link | 77.032±1.393 | | Graphormer | 2021-06-09 |
| Wasserstein Embedding for Graph Learning | ✓ Link | 76.8% | | WEGL | 2020-06-16 |
| Pure Transformers are Powerful Graph Learners | ✓ Link | 76.740±2.054 | | TokenGT | 2022-07-06 |
| A Fair Comparison of Graph Neural Networks for Graph Classification | ✓ Link | 76.4% | | DGCNN | 2019-12-20 |
| Learning Convolutional Neural Networks for Graphs | ✓ Link | 76.34% | | PSCN | 2016-05-17 |
| Weisfeiler and Leman Go Neural: Higher-order Graph Neural Networks | ✓ Link | 76.2% | | k-GNN | 2018-10-04 |
| TREE-G: Decision Trees Contesting Graph Neural Networks | ✓ Link | 75.9% | | TREE-G | 2022-07-06 |
| A Simple Baseline Algorithm for Graph Classification | ✓ Link | 75.2% | | SF + RFC | 2018-10-22 |
| Fea2Fea: Exploring Structural Feature Correlations via Graph Neural Networks | ✓ Link | 74.9% | | Fea2Fea-s3 | 2021-06-24 |
| Relational Reasoning Over Spatial-Temporal Graphs for Video Summarization | | 74.48% | | SAEPool_g | 2022-04-06 |
| Self-Attention Graph Pooling | ✓ Link | 74.06% | | SAGPool_g | 2019-04-17 |
| Hierarchical Representation Learning in Graph Neural Networks with Node Decimation Pooling | ✓ Link | 73.5% | | NDP | 2019-10-24 |
| A Novel Higher-order Weisfeiler-Lehman Graph Convolution | ✓ Link | 73.5 | | 2-WL-GNN | 2020-07-01 |
| graph2vec: Learning Distributed Representations of Graphs | ✓ Link | 73.22% ± 1.81% | | graph2vec | 2017-07-17 |
| A simple yet effective baseline for non-attributed graph classification | ✓ Link | 73.0% | | LDP | 2018-11-08 |
| Optimal Transport for structured data with application on graphs | ✓ Link | 72.82% | | FGW raw sp | 2018-05-23 |
| Multi-hop Attention-based Graph Pooling: A Personalized PageRank Perspective | ✓ Link | 72.32% | | MAGPool | 2024-03-04 |
| ASAP: Adaptive Structure Aware Pooling for Learning Hierarchical Graph Representations | ✓ Link | 71.48 | | ASAP | 2019-11-18 |
| Subgraph Networks with Application to Structural Feature Space Expansion | | 70.26% | | Deep WL SGN(0,1,2) | 2019-03-21 |
| An End-to-End Deep Learning Architecture for Graph Classification | ✓ Link | 69.00% | | DGCNN (sum) | 2018-04-29 |
| DDGK: Learning Graph Representations for Deep Divergence Graph Kernels | ✓ Link | 68.1% | | DDGK | 2019-04-21 |
| Graph Classification using Structural Attention | ✓ Link | 67.71% | | GAM | 2018-07-19 |
| Self-Attention Graph Pooling | ✓ Link | 67.45% | | SAGPool_h | 2019-04-17 |
| A Non-Negative Factorization approach to node pooling in Graph Convolutional Neural Networks | | 66.2% | | 1-NMFPool | 2019-09-07 |
| Capsule Neural Networks for Graph Classification using Explicit Tensorial Graph Representations | | 65.9% | | BC + Capsules | 2019-02-22 |
| SPI-GCN: A Simple Permutation-Invariant Graph Convolutional Network | | 64.11% | | SPI-GCN | 2019-04-08 |