Table 4 The comparison between condtional generative models RL based models in the task of generating molecules containing scaffold 4
From: Multi-objective de novo drug design with conditional graph generative model
Model | % valid | \(R_{\mathbf{c}}\) | \(EOR_{\mathbf{c}}\) | Diversity (\(I_{\mathbf{c}}\)) |
|---|---|---|---|---|
REINVENT | 0.998 ± 0.000 | 1.000 ± 0.000 | 172 | 0.75 ± 0.01 |
Naive RL | 0.984 ± 0.015 | 0.999 ± 0.001 | 172 | 0.48 ± 0.08 |
RL + Prior | 0.999 ± 0.001 | 1.000 ± 0.000 | 172 | 0.55 ± 0.09 |
Graph | 0.982 ± 0.001 | 0.88 ± 0.01 | 151 | 0.815 ± 0.001 |
SMILES | 0.969 ± 0.002 | 0.88 ± 0.00 | 151 | 0.819 ± 0.000 |