Graph Property Prediction On Ogbg Molhiv

评估指标

Test ROC-AUC

Validation ROC-AUC

评测结果

各个模型在此基准测试上的表现结果

			Paper Title	Repository
HyperFusion	0.8475 ± 0.0003	0.8275 ± 0.0008	-	-
PAS+FPs	0.8420 ± 0.0015	0.8238 ± 0.0028	-	-
HIG	0.8403 ± 0.0021	0.8176 ± 0.0034	-	-
DeepAUC	0.8352 ± 0.0054	0.8238 ± 0.0061	Large-scale Robust Deep AUC Maximization: A New Surrogate Loss and Empirical Studies on Medical Image Classification
FingerPrint+GMAN	0.8244 ± 0.0033	0.8329 ± 0.0039	-	-
Neural FingerPrints	0.8232 ± 0.0047	0.8331 ± 0.0054	Molecular Representation Learning by Leveraging Chemical Information	-
Graphormer + FPs	0.8225 ± 0.0001	0.8396 ± 0.0001	Do Transformers Really Perform Bad for Graph Representation?
Molecular FP + Random Forest	0.8208 ± 0.0037	0.8036 ± 0.0059	-	-
GPTrans-B	0.8126 ± 0.0032	-	Graph Propagation Transformer for Graph Representation Learning
CIN	0.8094 ± 0.0057	0.8277 ± 0.0099	Weisfeiler and Lehman Go Cellular: CW Networks
GSAT	0.8067 ± 0.0950	0.8347 ± 0.0031	Interpretable and Generalizable Graph Learning via Stochastic Attention Mechanism
EGT	0.806 ± 0.0065	-	Global Self-Attention as a Replacement for Graph Convolution
MorganFP+Rand. Forest	0.8060 ± 0.0010	0.8420 ± 0.0030	-	-
CIN-small	0.8055 ± 0.0104	0.8310 ± 0.0102	Weisfeiler and Lehman Go Cellular: CW Networks
Graphormer	0.8051 ± 0.0053	0.8310 ± 0.0089	Do Transformers Really Perform Bad for Graph Representation?
Graphormer (pre-trained on PCQM4M)	0.8051 ± 0.0053	0.8310 ± 0.0089	Do Transformers Really Perform Bad for Graph Representation?
GatedGCN+	0.8040 ± 0.0164	0.8329 ± 0.0158	Can Classic GNNs Be Strong Baselines for Graph-level Tasks? Simple Architectures Meet Excellence
directional GSN	0.8039 ± 0.0090	0.8473 ± 0.0096	Improving Graph Neural Network Expressivity via Subgraph Isomorphism Counting
P-WL	0.8039 ± 0.0040	0.8279 ± 0.0059	A Persistent Weisfeiler–Lehman Procedure for Graph Classification	-
Nested GIN+virtual node (ens)	0.7986 ± 0.0105	0.8080 ± 0.0278	Nested Graph Neural Networks

0 of 43 row(s) selected.