The Dr-KGQA System for Automatically Answering Medication Related Questions in Chinese

As the public medical resources becomes more and more scarce due to the aging of a large population in China, online medical consultation become a popular alternative. However, existing online medical consultation providers or forums are either labor-intensive, thus expensive, or free but unreliable with respect to the quality of answers. With the aim of providing a convenient, instant and reliable tool for people in need of simple medication related consultations, we developed a question and answering system (Dr-KGQA) to automatically answer the questions from users in an interactive manner, with the help of our powerful Chinese medical knowledge graph (Med-KG). Our system employs a pipeline of deep learning models including named entity recognition (NER) and relation matching. We conduct a series of experiments to demonstrate the performance of our system, on a human annotated dataset we collect. We have piloted our system on a family doctor platform in Chongqing, China, and the results show that our system is promising for real-world applications.