A chatbot that knows it all
Problem being addressed
Diffculty in seeing a doctor, long queuing time, and inconvenience of making appointments have long been hurdles facing patients when they try to access primary care services.
Based on the chatbot framework, the researchers build an online question-and-answer Healthcare Helper system for answering complex medical questions. It maintains a knowledge graph constructed from medical data collected from the Internet. It also implements a novel text representation and similarity deep learning model to find the most similar question from a large question-and-answer dataset.
Advantages of this solution
The chatbot framework based on a knowledge graph and a text representation and similarity model; the advantage of the knowledge graph lies in that it utilizes structured storage so that it may help easy maintenance and retrieval of domain-specific knowledge, while the advantage of the attention model utilizes deep learning to represent better and comprehend natural language questions. Therefore, the researchers develop a system that combines the advantages of both models by integrating a knowledge graph with a neural-based model.
Solution originally applied in these industries
Possible New Application of the Work
There is big potential of chatbot technologies to play a much more significant role in the medical domain. For example, a software chatbot can be deployed in the real-world to become home healthcare robots or hospital medical inquiry robots. Besides, it is possible to combine the data mining method and predict the potential diseases in a region of the population.
One novelty of the work lies in the utilization of a hybrid question-and-answer model that combines a knowledge graph database and an NLP model. If it cannot find any result, a text similarity model will be used to fund the answers from a large medical question-and-answer dataset. This can be applied in complex online queries to optimise the search results.
Source URL: #############