PROMISE: A pre-trained knowledge-infused multimodal representation learning framework for medication recommendation

Jialun Wu, Xinyao Yu, Kai He, Zeyu Gao, Tieliang Gong

Abstract: Highlights•Presenting an innovative multimodal framework leveraging multi-source domain knowledge.•Patient local state representations are learned through modality-specific learning flows.•A novel paradigm that progresses from detailed health states to overarching perspective.