Table 3 Metrics calculated for the final selections of high-scoring molecules according to the DRD2 bioactivity predictor (\(> 0.5\)), mono-objective optimization
From: Human-in-the-loop active learning for goal-oriented molecule generation
Metric | No Feedback | Feedback (\(T = 10\)) | |||
|---|---|---|---|---|---|
EPIG (\(\sigma _{\epsilon } = 0.3\)) | Uncertainty (\(\sigma _{\epsilon } = 0.3\)) | Greedy (\(\sigma _{\epsilon } = 0.3\)) | Random (\(\sigma _{\epsilon } = 0.3\)) | ||
Number of molecules | 121.00 ± 1.41 | 97.63 ± 6.04 | 88.00 ± 12.39 | 85.11 ± 22.53 | 98.43 ± 11.13 |
MAE Oracle-Pred. \(\downarrow\) | \(0.61 \pm 0.02\) | \(0.23 \pm 0.05\) ** | 0.14 ± 0.05 ** | \(0.31 \pm 0.04\) ** | \(0.15 \pm 0.04\) ** |
Internal Diversity \(\uparrow\) | \(0.70 \pm 0.01\) | \(0.60 \pm 0.03\) ** | \(0.60 \pm 0.03\) ** | 0.65 ± 0.06 * | \(0.57 \pm 0.02\) ** |
SA \(\downarrow\) | 3.36 ± 0.09 | \(3.58 \pm 0.44\) | \(3.63 \pm 0.31\) * | \(3.63 \pm 0.34\) * | \(3.91 \pm 0.57\) * |
QED \(\uparrow\) | \(0.41 \pm 0.03\) | 0.60 ± 0.08 ** | \(0.54 \pm 0.10\) ** | \(0.51 \pm 0.08\) ** | \(0.50 \pm 0.06\) ** |
Novelty \(\uparrow\) | 1.0 \(\pm 0.0\) | 1.0 \(\pm 0.0\) | 1.0 \(\pm 0.0\) | 1.0 \(\pm 0.0\) | 1.0 \(\pm 0.0\) |
Uniqueness \(\uparrow\) | 1.0 ± 0.0 | 1.0 ± 0.0 | 1.0 ± 0.0 | 1.0 ± 0.0 | 1.0 ± 0.0 |
Frag Gen-Train \(\uparrow\) | 0.95 \(\pm 0.01\) | \(0.90 \pm 0.10\) | \(0.85 \pm 0.20\) | \(0.64 \pm 0.21\) ** | \(0.90 \pm 0.18\) |
SNN Gen-Train \(\uparrow\) | \(0.41 \pm 0.01\) | \(0.49 \pm 0.02\) ** | \(0.52 \pm 0.02\) ** | \(0.46 \pm 0.05\) * | 0.52 ± 0.03 ** |
FCD Gen-Train \(\downarrow\) | 39.23 ± 2.03 | \(37.17 \pm 3.34\) | \(\textbf{36.19} \pm \textbf{3.85}\) | \(40.96 \pm 5.46\) | \(38.46 \pm 3.97\) |
Frag Gen-Queries \(\uparrow\) | - | 0.97 ± 0.10 | \(0.91 \pm 0.12\) | \(0.67 \pm 0.17\) | 0.97 ± 0.05 |
SNN Gen-Queries \(\uparrow\) | - | 0.54 ± 0.03 | \(0.50 \pm 0.01\) | \(0.49 \pm 0.07\) | 0.54 ± 0.02 |
FCD Gen-Queries \(\downarrow\) | - | \(11.99 \pm 2.61\) | \(14.80 \pm 4.90\) | \(23.77 \pm 12.06\) | 10.25 ± 1.59 |