Medical AI
These days, Google is making headlines as its artificial intelligence (AI) AlphaGo beated top pro Go player Lee Se-dol 2:0 in a highly publicized five-game Go series. The internet search giant is expanding its AI business by taking over four robotics companies including DeepMind which designed AlphaGo.
But in Korea, AI is an underdeveloped and poorly invested sector. “Korean companies have not made much progress in AI research. They still have a long way to go in terms of AI commercialization,” said Jin Jeong-yeol, director of the Kohyoung Technology. “Even though some small and medium-sized Korean companies apply for the patents of robotics technology, they are poor at connecting their technology to other business such as drones and self-driving cars and failing to effectively cope with market changes,” said an official of the Robotics and Automation Evaluation Department at the Office of Patent Administration of Korea.
But it is in hospitals that AI is applied most actively at home and abroad. In hospitals, AI is used for various purposes such as analyzing medical videos to pinpoint diseases and suggesting treatment plans optimized for some diseases. In Korea, various venture startups are developing health care technologies based on AI.
Vuno was founded by researchers who conducted AI projects at the Samsung Advanced Institute of Technology. The company is developing Vuno-Med where medical data meet the “deep learning” which is a core technology of AlphaGo playing Go game with Lee Se-dol.
Deep learning, a kind of machine learning, is a technology that trains the computer to enable it to identify things by imitating a human information processing system where the human brain does after finding patterns in numerous data. Buno-Med to which this technology is applied helps doctors diagnose better by making most of computed tomography (CT) and magnetic resonance imaging (MRI) data and detecting lung diseases quickly.
At the moment, Buno is engaging in work to prove the accuracy of Buno-Med’s lung disease diagnoses in concert with a research team at Seoul Asan Hospital. Their joint research over the past year found out that Buno-Med’s accuracy stood at 97 percent.
Lunit, founded by KAIST-educated young engineers fascinated at AI, dived into developing deep learning-based medical image reading technology. This company successfully attracted an investment of two billion won from Softbank Ventures among others after gaining worldwide recognitions including claiming fifth place after beating Google in an international image recognition contest held by Microsoft. The company is developing technology to diagnose breast cancer and tuberculosis in their early stages by having AI learn a huge amount of medical visual data in partnership with major hospitals including Seoul Samsung Hospital.
Another venture start-up is applying AI technology to the treatment of patients. Neofect is developing technology to empower patients to implement rehabilitation training programs without medical staffs’ helps by connecting AI technology to rehabilitation medical equipment. The company was founded jointly by its CEO Ban Ho-young who completed the MBA program of Virginia University and Choi Yong-keun, its CTO who studied with Ban at the KAIST in Korea.
The “Rafael Glove” developed by this company is medical equipment through which a stroke-stricken patients with their limbs paralyzed can receive rehabilitation treatment while playing a VR game with gloves on. The gloves have sensors.
When used first, Rafael Glove turns the ranges of a patient’s movements into the degrees of angles and saves them as data. During trainings, AI controls the levels of the training by analyzing data measured during the training in real time. In the next training session, the software suggests proper goals and plans to the patients in accordance with his or her conditions. The company is planning on developing home rehabilitation equipment which can save rehabilitation patients’ trips to hospitals.
New drug development also began to involve AI, too. Standardigm founded by AI specialist and CEO Kim Jin-han and directors Song Sang-ok and Yun So-jeong, both of whom are system biologists, is developing technology that can significantly change new drug development processes by mixing the two sectors of AI and system biology. The technology aims to reduce a new drug development process that takes over ten years and over one trillion won.
They are developing computer modeling technology that predicts drug’s working mechanism by studying the huge volume of medical and biological information. This new technology is expected to assemble drugs to combat diseases and optimize the selection of drug candidate groups and patient groups for clinical tests for new drug development.