Skip to main content

Table 2 Benchmark results of the RL-biased-G: last epoch

From: Conditional reduction of the loss value versus reinforcement learning for biassing a de-novo drug design generator

Objective

Validity

Novelty

Uniqueness

intDiv

Desirability

1

41.09

45.94

39.93

0.79

74.40 = 561

2

99.75

61.83

1.07

0.67

24.24 = 16

3

52.97

25.82

24.70

0.79

59.46 = 201

4

99.96

96.69

0.17

0.65

84.23 = 13