Binary Classification on MoleculeNet BBBP DeepChem (test)
85ROC AUCMGCN
Evaluation results
| Method | Links | |
|---|---|---|
| MGCNBackbone=MGCN, Pre-training Strategy=None2021.06 | 85 | |
| DMP_TF (100M)Backbone=Transformer, Pre-training Strategy=DMP, Pre-training Data Scale=100M molecules2021.06 | 78.4 | |
| DMP_TFBackbone=Transformer, Pre-training Strategy=DMP, Training Objective=MLM + Dual-view consistency2021.06 | 78.1 | |
| DMP_TF+GNNBackbone=Transformer + GNN, Pre-training Strategy=DMP2021.06 | 77.8 | |
| TF (MLM) + GNN (MLM)Backbone=Transformer + GNN, Training Objective=MLM2021.06 | 76.1 | |
| TF (x2)Backbone=Dual Transformer, Training Objective=MLM2021.06 | 75.6 | |
| DMP_GNN (100M)Backbone=GNN, Pre-training Strategy=DMP, Pre-training Data Scale=100M molecules2021.06 | 75.2 | |
| TF (MLM)Backbone=Transformer, Training Objective=MLM2021.06 | 74.9 | |
| DMP_GNNBackbone=GNN, Pre-training Strategy=DMP, Training Objective=MLM + Dual-view consistency2021.06 | 74.7 | |
| GNN (MLM)Backbone=GNN, Training Objective=MLM2021.06 | 74.5 | |
| GNN (x2)Backbone=Dual GNN, Training Objective=MLM2021.06 | 74.1 | |
| MolCLRBackbone=GNN, Pre-training Strategy=Contrastive Learning2021.06 | 73.6 | |
| SVMBackbone=Support Vector Machine, Pre-training Strategy=None2021.06 | 72.9 | |
| RFBackbone=Random Forest, Pre-training Strategy=None2021.06 | 71.4 | |
| D-MPNNBackbone=D-MPNN, Pre-training Strategy=None2021.06 | 71.2 | |
| DMP_TF w/o MLMBackbone=Transformer, Training Objective=Dual-view consistency2021.06 | 71.1 | |
| Hu et al.Backbone=GNN, Pre-training Strategy=Pre-trained2021.06 | 70.8 |