Table 1. Summary of the metrics to compare performance of the different search approaches over the entire search space. Any approaches with pretraining components were pretrained on 20 000 molecules.
| Search algorithm | Unique new evaluations | Unique molecules with Fcomb > 0 | Rate of discovery of molecules with Fcomb > 0 | Molecules found better than benchmark dataset. (outstanding molecules) |
|---|---|---|---|---|
| BO-Learned | 2096 | 1439 | 69% | 8 |
| BO-Mord | 2675 | 535 | 20% | 16 |
| BO-Prop | 3868 | 2336 | 60% | 1 |
| SUEA | 722 | 394 | 54% | 1 |
| EA | 3614 | 1146 | 32% | 1 |
| Rand | 2887 | 0 | 0% | 0 |