Skip to main content

Table 6 Results of the constrained optimization for Junction Tree Variational Autoencoder [30] (JT-VAE), Graph Convolutional Policy Network [35] (GCPN) and Mol-CycleGAN

From: Mol-CycleGAN: a generative model for molecular optimization

\(\delta \) JT-VAE GCPN Mol-CycleGAN
Improvement Similarity Success (%) Improvement Similarity Success (%) Improvement Similarity Success (%)
0 1.91 ± 2.04 0.28 ± 0.15 97.5 4.20 ± 1.28 0.32 ± 0.12 100.0 8.30 ± 1.98 0.16 ± 0.09 99.75
0.2 1.68 ± 1.85 0.33 ± 0.13 97.1 4.12 ± 1.19 0.34 ± 0.11 100.0 5.79 ± 2.35 0.30 ± 0.11 93.75
0.4 0.84 ± 1.45 0.51 ± 0.10 83.6 2.49 ± 1.30 0.47 ± 0.08 100.0 2.89 ± 2.08 0.52 ± 0.10 58.75
0.6 0.21 ± 0.75 0.69 ± 0.06 46.4 0.79 ± 0.63 0.68 ± 0.08 100.0 1.22 ± 1.48 0.69 ± 0.07 19.25
  1. The biggest improvements across all methods are italicized