An accurate and interpretable deep learning model for yield prediction using hybrid molecular representations

A deep learning framework integrating hybrid molecular fingerprints is established for intelligent design of ionic liquid catalysts for CO 2 cycloaddition reaction under mild conditions.