Towards User Friendly Medication Mapping Using Entity-Boosted Two-Tower Neural Network

Recent advancements in medical entity linking have been applied the area of scientific literature and social media data. However, with adoption telemedicine conversational agents such as Alexa healthcare settings, name inference has become an important task. Medication is task mapping user friendly medication names from a free-form text to concept normalized list. This challenging due differences use terminology health care professionals conversations coming lay public. We begin descriptive phrases (DMP) standard (SMN). Given prescriptions each patient, we want provide them flexibility referring their preferred ways. approach this ranking problem which maps SMN DMP by ordering list medications patient’s prescription obtained pharmacies. Furthermore, leveraged output intermediate layers performed clustering. present Inference Model (MIM) achieving state-of-the-art results. By incorporating entities based attention, further improvement for models.