AI Against COVID-19

16 Apr, 2020
Vicent Botti, professor and researcher of UPV (Universidad Politecnica de Valencia), describes artificial intelligence solutions in the prevention, diagnosis and treatment in diseases such as COVID-19

ARTIFICIAL INTELLIGENCE AGAINST COVID-19

World Health Organization states that a success in the public health response relies on four critic factors: understanding of transmissibility and high-risk population; establishing the natural history of infection, for instance the incubation period and the mortality rate; identification and characterization of the origin of the epidemic; and the development of epidemiological models to suggest preventive and control measures. Artificial intelligence can provide solutions to each of these factors in order to help healthcare professionals in the prevention, diagnosis and treatment of diseases such as COVID-19. This solutions can be applied at the different stages of the epidemic: before, during and after.

Before the epidemic. AI can detect early outbreaks of epidemic disease outbreaks. In december 2019, BlueDot system detected that the Wuhan coronavirus was a new strain and how it was going to spread around the world. This system combines public health and medical experience through an advanced data analysis and machine learning to give solutions for an early detection of infectious diseases. 

After the epidemic. Data science, machine learning and deep learning systems can help us understand what has happened and obtain epidemiological models that allow us to define effective prevention and control strategies that facilitate making the best decisions. For this purpose not only health data is required but also data such as mobile phone mobility, home energy consumption, banks access, use of public transport, etc. 

During the epidemic. AI can assist in the prevention, diagnosis, and treatment of COVID-19. Alibaba has developed a system that can detect the coronavirus on chest CT scans with 96% accuracy. DeepMind has shared its AI results detailing the structure of six proteins linked to SARS-CoV-2, the coronavirus that causes COVID-19. We could go on listing multiple AI applications. But, in my opinion, at this stage there are two AI solutions that could have a special impact

First one is to use optimisation methods that enable an efficient use of available human and material resources. This would allow assigning patients to the most appropriate hospitals, distributing medical supplies to the neediest hospitals at all times... 

Second, is the use of AI techniques to analyze the movement of people in the geographic areas with data of mobile networks. Assuming the case of a person who goes to the hospital and tests positive; healthcare professionals will ask about the people with whom this person has been in contact. But this person, the previous days, has traveled by public transport, has gone shopping, etc. and has been in contact with other people. How to identify them to warn them so that they can follow confinement strategies or carry out tests to limit the spread of the virus? In my opinion, what would have the greatest impact is an AI system capable of predicting the risk of contagion.

Click here for reference (Spanish)