Document-aware Information Extractor for Chinese Medical Dialogue

Electronic medical records (EMRs) are one of the methods to help doctors effectively manage and analyze patient medical records. These EMRs not only help doctors save a lot of time to analyze medical records, but also reduce the hospital's demand for doctors and reduce hospital expenditure costs. Therefore, we proposed the document-aware information extractor (DIE) to effectively extract the information about the patient's physical condition in the conversation between the doctor and the patient. In this paper, we proposed a encoder-decoder model to extract the medical items amongst the doctor-patient dialogue for further usage of EMRs generation. The experimental result shows that our model achieves better results compared to the baseline models, which indicates the model effectiveness.