List your company
ResearchFormerly WizWand

Molecular Property Prediction on BBBP (test)

0.92ROC-AUCProtoW-L2
0.45720.577350.69750.81765Jun 8, 2020Jun 3, 2021May 30, 2022May 25, 2023May 20, 2024May 15, 2025May 11, 2026
Updated 5mo ago

Evaluation results

MethodLinks
ProtoW-L2Contrastive regularization=false, Distance metric=L22020.060.92
ProtoW-L2Contrastive regularization=true, Distance metric=L22020.060.92
ProtoW-DotContrastive regularization=false, Distance metric=Dot2020.060.919
ProtoW-DotContrastive regularization=true, Distance metric=Dot2020.060.919
ProtoS-L2Contrastive regularization=true, Distance metric=L22020.060.918
D-MPNNPooling method=summation2020.060.915
D-MPNN+TopK PoolPooling method=TopK Pool2020.060.912
Fingerprint+MLP2020.060.911
D-MPNN+SAG PoolPooling method=SAG Pool2020.060.901
GINPooling method=summation2020.060.9
MolSight S6prediction_mode=Yes/No parsed prediction2026.050.893
GATPooling method=summation2020.060.888
GIT-MolInput Modality=Graph + SMILES2023.080.739
GraphLoG2025.040.725
GPF-plusPre-training Strategy=GCL2022.090.7218
Mole-BERT2023.080.719
GIT-MolInput Modality=SMILES2023.080.719
MoleBERTBackbone=MoleBERT, Training=Pre-trained2025.040.719
GPFPre-training Strategy=GCL2022.090.7111
GIT-MolInput Modality=Graph2023.080.711
GraphMVP2023.080.708
MGSSL (DFS)Backbone=GIN, Evaluation Protocol=Fine-tuning2021.100.705
KV-PLM2023.080.705
MoMu2023.080.705
GraphMAE+GFSEBackbone=GraphMAE, SE=GFSE2025.040.705
SUPThardPre-training Strategy=ContextPred2024.020.7018
SUPTsoftPre-training Strategy=ContextPred2024.020.7006
SUPTsoftPre-training Strategy=EdgePred2024.020.6993
SUPTsoftPre-training Strategy=GCL2024.020.6979
MGSSL (BFS)Backbone=GIN, Evaluation Protocol=Fine-tuning2021.100.697
SUPThardPre-training Strategy=GCL2024.020.6968
FTPre-training Strategy=ContextPred2022.090.6965
FTPre-training Strategy=ContextPred2024.020.6965
GPFPre-training Strategy=EdgePred2022.090.6957
SUPThardPre-training Strategy=EdgePred2024.020.6951
GraphCL2025.040.695
GraphMAEBackbone=GraphMAE, Training=Pre-trained2025.040.695
FTPre-training Strategy=GCL2022.090.6949
FTPre-training Strategy=GCL2024.020.6949
GPPT (w/o ol)Pre-training Strategy=EdgePred2022.090.6943
GPPTw/o olPre-training Strategy=EdgePred2024.020.6943
GPFPre-training Strategy=EdgePred2024.020.6936
GPF-plusPre-training Strategy=EdgePred2024.020.6931
GraphPromptPre-training Strategy=EdgePred2022.090.6929
GraphPromptPre-training Strategy=EdgePred2024.020.6929
GPF-plusPre-training Strategy=ContextPred2022.090.6915
GINE+GFSEBackbone=GINE, SE=GFSE2025.040.691
GPF-plusPre-training Strategy=EdgePred2022.090.6906
GPFPre-training Strategy=ContextPred2024.020.6904
MoleBERT+GFSEBackbone=MoleBERT, SE=GFSE2025.040.689
GPF-plusPre-training Strategy=ContextPred2024.020.6859
GPFPre-training Strategy=ContextPred2022.090.6848
SUPThardPre-training Strategy=AttrMasking2024.020.6839
SUPTsoftPre-training Strategy=AttrMasking2024.020.6837
GPFPre-training Strategy=GCL2024.020.6811
InfomaxBackbone=GIN, Evaluation Protocol=Fine-tuning2021.100.681
GPFPre-training Strategy=AttrMasking2022.090.6809
GroverBackbone=GIN, Evaluation Protocol=Fine-tuning2021.100.68
GPS+GFSEBackbone=GPS, SE=GFSE2025.040.68
SSP2025.040.679
GPS+RWSEBackbone=GPS, SE=RWSE2025.040.679
GPS+LapPEBackbone=GPS, SE=LapPE2025.040.679
GraphCL2023.080.678
GPS+GPSEBackbone=GPS, SE=GPSE2025.040.678
GPF-plusPre-training Strategy=GCL2024.020.6777
GPF-plusPre-training Strategy=AttrMasking2022.090.6771
GINEBackbone=GINE, Training=from-scratch2025.040.677
SUPTsoftPre-training Strategy=Infomax2024.020.6763
FTPre-training Strategy=Infomax2022.090.6755
FTPre-training Strategy=Infomax2024.020.6755
GPT-GNNBackbone=GIN, Evaluation Protocol=Fine-tuning2021.100.675
GPFPre-training Strategy=AttrMasking2024.020.6749
GPF-plusPre-training Strategy=AttrMasking2024.020.6732
SUPThardPre-training Strategy=Infomax2024.020.6729
GPF-plusPre-training Strategy=Infomax2022.090.6717
GPSBackbone=GPS, Training=from-scratch2025.040.671
GPF-plusPre-training Strategy=Infomax2024.020.6703
GCCBackbone=GIN, Evaluation Protocol=Fine-tuning2021.100.669
GPFPre-training Strategy=Infomax2022.090.6683
GINE+RWSEBackbone=GINE, SE=RWSE2025.040.667
GPFPre-training Strategy=Infomax2024.020.6658
FTPre-training Strategy=EdgePred2022.090.6656
FTPre-training Strategy=EdgePred2024.020.6656
MoleBERT+RWSEBackbone=MoleBERT, SE=RWSE2025.040.665
GINE+GPSEBackbone=GINE, SE=GPSE2025.040.664
GraphMAE+RWSEBackbone=GraphMAE, SE=RWSE2025.040.664
FTPre-training Strategy=AttrMasking2022.090.6633
FTPre-training Strategy=AttrMasking2024.020.6633
GINE+LapPEBackbone=GINE, SE=LapPE2025.040.658
Attribute maskingBackbone=GIN, Evaluation Protocol=Fine-tuning2021.100.656
No pretrainBackbone=GIN, Evaluation Protocol=Fine-tuning2021.100.655
GPPTPre-training Strategy=EdgePred2022.090.6413
GPPTPre-training Strategy=EdgePred2024.020.6413
Qwen2-VL-2Bprediction_mode=Yes/No parsed prediction2026.050.475