Skip to main content
Fig. 2 | Journal of Cheminformatics

Fig. 2

From: Using GANs with adaptive training data to search for new molecules

Fig. 2

New molecules produced for different replacement strategies with recombination. For control+re (blue), the training data is fixed. For random+re (red), molecules from the generator randomly replace molecules in the training data. For drug+re (green), molecules from the generator only replace training samples if they have a higher drug-likeness score. a Similar to the case without recombination, random and drug replacement strategies outperform control as training progresses. Plot shows average over three training runs for each selection type. b Although drug+re produces less overall new molecules than random+re, it generates more top performers. Plot shows average over three runs for each selection type with error bars showing one standard deviation

Back to article page