Graph Classification On Dd

Metrics

Accuracy

Results

Performance results of various models on this benchmark

Model Name	Accuracy	Paper Title	Repository
1-NMFPool	76.0%	A Non-Negative Factorization approach to node pooling in Graph Convolutional Neural Networks	-
GraphStar	79.60%	Graph Star Net for Generalized Multi-Task Learning	-
DGCNN	79.37%	An End-to-End Deep Learning Architecture for Graph Classification
DGCNN	77.21%	DGCNN: Disordered Graph Convolutional Neural Network Based on the Gaussian Mixture Model	-
GCN	78.151±3.465	Semi-Supervised Classification with Graph Convolutional Networks	-
EigenGCN-3	78.6%	Graph Convolutional Networks with EigenPooling	-
NDP	72%	Hierarchical Representation Learning in Graph Neural Networks with Node Decimation Pooling	-
GATv2	75.966±2.191	How Attentive are Graph Attention Networks?	-
TokenGT	73.950±3.361	Pure Transformers are Powerful Graph Learners	-
GFN-light	78.62%	Are Powerful Graph Neural Nets Necessary? A Dissection on Graph Classification	-
DGK	73.50%	Deep Graph Kernels	-
PNA	78.992±4.407	Principal Neighbourhood Aggregation for Graph Nets	-
S2V (with 2 DiffPool)	82.07%	Hierarchical Graph Representation Learning with Differentiable Pooling	-
GAT	73.109±3.413	Graph Attention Networks	-
SEAL-SAGE	80.88%	Semi-Supervised Graph Classification: A Hierarchical Graph Perspective	-
TREE-G	76.2%	TREE-G: Decision Trees Contesting Graph Neural Networks	-
GMT	78.72%	Accurate Learning of Graph Representations with Graph Multiset Pooling	-
U2GNN (Unsupervised)	95.67%	Universal Graph Transformer Self-Attention Networks	-
LDP + distance	77.5%	A simple yet effective baseline for non-attributed graph classification	-
Graph-JEPA	78.64%	Graph-level Representation Learning with Joint-Embedding Predictive Architectures	-

0 of 52 row(s) selected.