Generation of novel Diels-Alder reactions using a generative adversarial network

Deep learning has enormous potential in the chemical and pharmaceutical fields. Among these, Generative Adversarial Network, as an excellent generative model, has shown its remarkable performance in the field of molecular generation, but it has few applications in organic chemistry. Therefore, we attempt to apply GAN as a generative model for the task of reaction generation to expand the application of GAN in chemistry. In this work, we used the MaskGAN model trained with 14092 Diels-Alder reactions, and we finally generated 1441 novel Diels-Alder reactions that learn reaction rules in-depth, which demonstrates that reaction generation can be used in the field of chemistry, and helps chemists explore novel reactions.