About BianQue

Based on the six characteristics of active health: initiative, preventiveness, precision, personalization, co-construction and sharing, and self-discipline, the Future Technology College of South China University of Technology-Guangdong Key Laboratory of Digital Twins has opened up the Chinese field of life ProactiveHealthGPT, the base of the spatial active health large-scale model, including:

  • BianQue, a large-scale living space health model that has been fine-tuned by tens of millions of Chinese health dialogue data instructions
  • has passed a million scale In the field of psychological counseling, Chinese long-text instructions and multiple rounds of empathy dialogue data combined instructions to fine-tune the mental health model SoulChat (SoulChat) The world accelerates the research and application of large models in active health fields such as chronic diseases and psychological counseling. This project is BianQue, a large model of living space health .

Recently updated

BianQue Health Big Data BianQueCorpus

After research, we found that in the field of health, users usually do not clearly describe their problems in a round of interaction, and the current common open source medical question-and-answer models (for example: ChatDoctor, Materia Medica (HuaTuo, formerly known as Huatuo), DoctorGLM, MedicalGPT-zh) focus on solving the problem of single-round user description, while ignoring the situation that "the user description may be insufficient". Even ChatGPT, which is currently popular, has similar problems: if users do not force ChatGPT to use a question-and-answer format through text descriptions, ChatGPT also prefers to describe users, and quickly gives it what it thinks is appropriate. Suggestions and solutions. However, when a doctor actually talks with a user, there will often be "multiple rounds of continuous inquiries by the doctor based on the user's current description". And at the end, the doctor gives comprehensive advice based on the information provided by the user, as shown in the figure below. We define the process of the doctor's continuous questioning as a chain of questioning (CoQ, Chain of Questioning) . When the model is in the stage of the questioning chain, the next question is usually determined by the dialogue context history.

We combine the current open source Chinese medical question and answer datasets ( MedDialog-CN , IMCS-V2 , CHIP-MDCFNPC , MedDG , cMedQA2 , Chinese-medical-dialogue-data ) to analyze the single round/ The multi-round feature and doctor inquiry feature, combined with the laboratory's long-term self-built living space health dialogue big data, has built Bianque health big data BianQueCorpus with a scale of tens of millions. The dialog data is unified into an instruction format in the form of "patient: xxx\n doctor: xxx\n patient: xxx\n doctor:", as shown in the figure below.

Visit Official Website


Community Posts
no data
Nothing to display