“What we have done is use natural language processing and machine learning techniques to train this engine to recognize whether this is an outbreak of anthrax in Mongolia versus a reunion of the heavy metal band Anthrax.”
Kamran Kahn, CEO, BlueDot

Coronavirus, or COVID-19 (2019-nCov), started in Wuhan, China in December of 2019, was manifested by a cluster of people with pneumonias that may have originated from live animals at the Hunan Seafood Market. As of middle of February, it is estimated that close to 100,000 people have been infected with close to 2,000 deaths worldwide.

There is a myriad of data and information from a variety of sources such as social media, online blogs, websites, public health channels, and news reports on coronavirus. Machine learning can be used to organize this flow of data and information to more precisely follow and predict areas of disease spread. This data mining strategy using natural language processing needs to differentiate people who may be simply curious when they use search tools for coronavirus from people who have actually contracted the disease. This data inaccuracy was exemplified by the dismal failure of Google Flu Trends in 2013, which overestimated the population afflicted with the flu and promulgated the term “big data hubris”. BlueDot, an AI company with a multidisciplinary team that focuses on machine intelligence and human cognition, uses algorithms and more than 100 datasets to deliver critical insights on infectious disease around the world faster than most government agencies. While sources of data such as social media postings may be difficult or inaccurate, other sources (such as airline ticketing data) may be more reliable. In addition, Hong Kong is also offering an online consultation via the health ministry using an automated artificial intelligence answering tool. Finally, another essential AI application to fight this novel virus is to use this technology for drug discovery. BenevolentAI, an AI company that uses tools such as knowledge graphs, machine learning, and molecular modelling, have discovered compounds that are more likely to be effective than most, with the enzyme adaptor-associated protein kinase 1, or AAK1, as the target in coronavirus.

Tracking and controlling a virus, particularly virulent ones, may be akin to playing real-time strategy games (multiple fronts with incomplete data). Viruses mutate at a rapid pace but one can hope that humans and machines together as a dyad and in synergy can perhaps someday adequately neutralize these viruses.

Anthony Chang, MD, MBA, MPH, MS
Founder, AIMed
Chief Intelligence and Innovation Officer
Medical Director, The Sharon Disney Lund
Medical Intelligence and Innovation Institute (mi3)
Children’s Hospital of Orange County