Predicting the Coronavirus Outbreak: How AI Connects the Dots to Warn About Disease Threats
BlueDot’s Al algorithm is being featured in news programs everywhere for its ability to predict new Coronavirus outbreak locations days before official alerts come from the Centers for Disease Control and Prevention (CDC) and the World Health Organization. BlueDot is a Canadian artificial intelligence firm that uses both official and unofficial data reports to produce its predictions. BlueDot’s Al algorithm’s use of large sources of data from airlines, news stories in various languages and reports from disease tracking networks allows it to better produce accurate results ahead of schedule. The algorithm tracks how the diseases are spread, unlike traditional epidemiology that tracks where and when people contract a disease. However, the accuracy varies upon the amount and quality of incoming data.
Data, as previously mentioned, needs a large amount of preprocessing for computers to be able to process it. Preprocessing is often done through a newer technique called deep learning, which is used to make sense of unstructured data. Newer algorithms process this data through artificial neural processing that works similarly to the brain. Currently, these algorithms are trained to process unstructured data therefore the results are only as reliable as their training. Often times, too little data or poor training cause the results of algorithms like BlueDot’s Al algorithm to be skewed. Al has great promise. As the training of these newer algorithms improves so does the predictability of Al. Al increasingly grows as a useful tool for predicting the spread and location of new Coronavirus outbreaks.