The robot scored 456 points in the exam, well above the national average, according to its research team at the Tsinghua-iFlyTek Joint Lab of Tsinghua University and China's leading AI enterprise iFlyTek Co., Ltd.
Xiaoyi studied nearly one million medical images, 53 medical books, two million medical records, and 400,000 medical literatures and medical reports before sitting for the test, said Wu Ji, director of the joint lab.
For the critics who say that the robot only works by relying on its formidable memory ability, Wu explained that it does not make sense, because medical examination questions vary every year.
Wu said its study, reasoning, and judgment abilities also played a part in Xiaoyi's success.
However, the robot is not likely to replace human doctors, noted Tao Xiaodong, a project manager at iFlyTek, adding that the robot is not able to flexibly use the knowledge it possesses, as there are often unforeseen circumstances during actual diagnosis and treatment.
The robot will be applied nationwide from March next year in medical education and training, and will be used to assist doctors in clinical diagnosis and treatment to improve diagnosis accuracy and shorten treatment time.
At present, the research team has established cooperative relationship with the National Medical Examination Center to jointly explore further applications of AI in medical examinations.
This article was originally published in Huanqiu.