Researchers have successfully applied AI in predicting the spread of coronavirus as well as identify potential drugs for the treatment of the disease. The virus outbreak was first reported in Wuhan China and has infected over 60,000 people and killed more than 1,350 people. Countries such as the United States have imposed travel restrictions to prevent the spread of the virus from the country.
Use AI in tracking the spread of coronavirus
Scientists in China are using machine learning techniques to monitor the spread of coronavirus across the country. The experts from Harvard Medical School are mining data on social media posts, web, news reports, official medical data, and doctors’ information to anticipate areas and populations that the virus might be spreading next. The AI program is tracking the use of specific words within a geographical area that might imply symptoms such as respiratory fever, which are associated with the virus. The algorithm uses natural language processing to parse textual data from web sources and machine learning to distinguish various usage scenarios of the keywords. For example, the program can distinguish between someone delivering the news about coronavirus and someone complaining about the symptoms of the disease. Another company called BlueDot applied data mining on a variety of sources, except social media, to identify the outbreak before Chinese health officials confirmed the disease.
Challenges associated with the use of AI in combating 2019-nCoV
Despite the effectiveness of the program, it still faces challenges of people spreading rumors about the disease. In addition, the program has trouble distinguishing the symptoms of coronavirus from those of the common flu which are very similar while researchers have not yet identified all the symptoms of the disease. Another challenge is the inability to track the spread of coronavirus in areas with very few cases such as the United States and other parts of China. The lack of historical data also poses a challenge in identifying cases involving the disease.
Despite the challenges, the program was able to identify the outbreak in an area on December 30 last year through a concentration of flu-like symptoms in a specific area. The World Health Organization confirmed the case and has already declared coronavirus a global health emergency.
Overcoming the challenges
To overcome the lack of comprehensive data, researchers are considering the use of crowdsourced data. Experts are looking at utilizing data provided by volunteers as well as other online platforms. Tech companies are pulling request data for people suspecting of contracting the disease, mobile pings, flight and train itineraries. Some Chinese tech companies have developed apps to alert people whether they have traveled with a person with a confirmed case. Companies such as Baidu and the developers of WeChat, TenCent, have provided anonymous historical data to map the spread of the virus. Other firms such as Megvii have set up detection centers in major ports of transit. Baidu has also adopted a similar approach to detecting cases on the ground.
Apart from incorporating more data sources, experts are also retraining the AI models to monitor a different set of coronavirus symptoms.
Unlike the previous outbreaks of contagious diseases such as severe acute respiratory syndrome (SARS), researchers have more data and tools to model the spread of the epidemic. For the first time in the history of epidemics, big data is playing a major role in combating the spread of a major outbreak. Recent advances in AI have allowed experts and governments to address health care challenges in real-time.
Use of AI in finding coronavirus treatment
Apart from modeling the spread of the coronavirus epidemic, researchers are also using AI to mine through existing medical repositories for possible treatments of the virus. Using deep learning, researchers from South Korean firm Deargen indicate they have discovered various antiretroviral drugs that could potentially treat the virus.
Similarly, a UK-based startup company BenevolentAI applies artificial intelligence to search through a list of approved treatments that could mitigate the spread of coronavirus. The AI algorithm searches for drugs from a repository of medical data and scientific literature based on the chemical properties of the coronavirus. The AI algorithm narrowed down to a drug called Baricitinib used in the treatment of rheumatoid arthritis as a possible candidate for the cure.
Another AI company called Silico Medicine disclosed that its AI algorithm had discovered new molecules that would stop the replication of the coronavirus in the bodies of its victims. Despite the progress made, scientists warn that it could take time. However, the application of AI in coronavirus epidemic could reduce the time researchers spend in analyzing medical sources thus rushing the process. Similarly, the use of AI could complement human efforts in addressing public health concerns in real-time and using a smaller team.