We reproduce the human invention of the Chan–Lam reaction by predicting “unseen” reactions with “zero-shot reaction prediction”. We also mimics the later stage of the history of this reaction via “one-shot/few-shot reaction prediction” approaches.
The GCN model predicts reactions based on small-scale datasets, and its accuracy is shown to be significantly higher than that of the transformer model.
Transformer, a sequence-to-sequence deep learning model, is capable of predicting the reaction intermediates of radical cascade cyclization. This study provides a novel approach to help chemists discover the mechanisms of organic reactions.