Illustration of multi-modal learning on drug development. The core idea is to develop an AI model to simultaneously cope with different data modals (i.e. image and text) for drug recognition.