AI beats Chinese doctors in a tumor diagnosis competition

AI has trumped humans yet again - this time in diagnosing tumor accurately and quickly. The BioMind AI system, designed especially for neuroimaging recognition, has managed to beat a team of 15 top Chinese doctors by a margin of two to one.

When diagnosing tumors of the brain in the 225 test cases, BioMind managed to complete the task in just 15 minutes, as opposed to the doctors who took 30 minutes to complete the same task. BioMind correctly detected a tumor with a staggering accuracy of 87%, as compared to the doctors who managed only 66%.

What is BioMind?

BioMind is an artificially intelligent system developed at the Artificial Intelligence Research Centre for Neurological Disorders at the Beijing Tiantan Hospital, in collaboration with the Capital Medical University.

For training the system, it was fed thousands of images of different nervous system-related diseases, especially related to the brain, that the hospital had archived for over 10 years. Such a large data set allowed the AI to diagnose the common and not-so-common neurological diseases with a staggering accuracy of approximately 90% in comparison to a senior doctor at the hospital.

AI not something to be afraid of, says hospital vice president

Wang Yongjun, the executive vice president of the Tiantan Hospital which conducted this competition, expressed no concerns over this comprehensive victory of the AI over humans. He said the contest was not intended to pit humans against AI, but to help the doctors interact better with the technology.

One of the doctors who lost against BioMind, Dr. Lin Yi, said she welcomes Artificial Intelligence as a friend and thinks it will reduce the doctors’ workload. Not just that, it will also push them to keep learning and improve their skills.

The tumor diagnosis competition demonstrated a well-trained AI’s capability to outperform humans yet again. Only recently, a team of AI algorithms had beaten the world’s best human Dota 2 players in the multiplayer mode, proving they are capable of working in teams. The recent events only fuel further the debate of whether AI would replace human effort, or would augment human expertise to accomplish the tasks more efficiently.