Since the first recording of what we now call Covid-19 infection in Wuhan, Hubei province, China on Dec 31, 2019, the disease has spread worldwide and met with a wide variety of social distancing and quarantine policies. The recent worldwide outbreak of the novel corona-virus (COVID-19) opened up new challenges to the research community. Artificial intelligence methods can be useful to predict the parameters, risks, and effects of such an epidemic. Such predictions can be helpful to control and prevent the spread of such diseases. AI technologies are far from replicating human intelligence, they are proving to be very helpful in tracking the outbreak, diagnosing patients, disinfecting areas, and speeding up the process of finding a cure for COVID-19. Since neural networks can be used to approximate nonlinear functions with a finite set of parameters, they serve as a powerful tool to approximate quarantine effects in combination with the analytical epidemiological models. The downside is that the internal workings of a neural network are difficult to interpret.  As governments and health organizations scramble to contain the spread of corona-virus, they need all the help they can get, including from artificial intelligence (AI).

In paper Neural Network aided quarantine control model estimation of global Covid-19 spread authors attempt to interpret and extrapolate from publicly available data using a mixed first-principles epidemiological equations and data-driven neural network model. Leveraging neural network augmented model, authors focus analysis on four locales: Wuhan, Italy, South Korea and the United States of America, and compare the role played by the quarantine and isolation measures in each of these countries in controlling the effective reproduction number of the virus. Obtained results unequivocally indicate that the countries in which rapid government interventions and strict public health measures for quarantine and isolation were implemented were successful in halting the spread of infection and prevent it from exploding exponentially.

In work Neural network based country wise risk prediction of COVID-19 authors have proposed a Bayesian optimization guided shallow LSTM for predicting the country-specific risk of the COVID-19. They have combined trend data to predict different parameters for the risk classification task, and proposed a shallow Long short-term memory (LSTM) based neural network to predict the risk category of a country. The results show that the proposed pipeline outperforms against state-of-the-art methods for 170 countries data and can be a useful tool for such risk categorization. The tool can be used to predict long-duration outbreak of such an epidemic such that we can take preventive steps earlier.   Authors propose to use the country-specific optimized network for accurate prediction and noted that this is suitable when we have a small and uncertain dataset. Combining the overall optimized LSTMs, authors also note that rather deep neural networks, the majority of the cases a small neural network perform well in the data.

In work Neural Network aided quarantine control model estimation of COVID spread in Wuhan, China scientists used an optimized neural network-augmented SIR model to forecast the quarantine control strength and effective reproduction number in Wuhan.

Neural network based risk prediction of COVID-19
Neural network based risk prediction of COVID-19

By approximating this time varying quarantine strength with a neural network, they train the governing system of augmented SIR differential equations based on a loss function term obtained from the infected and recovered case count data generated by the Chinese National Health Commission. Authors do not consider the cases when the quarantined/isolated population come into contact with the non-infected population and lead to the transmission of disease. Irrespective of these assumptions, authors believe this to be a first study to quantify the effective of quarantine measures implemented in Wuhan, with an interpretable physical model aided by machine learning techniques, involving very few free parameters. By training this governing system, it is possible to not only approximate the maximum plateau value seen in the infected case count 30 days post the implementation of quarantine control, but also recover a monotonically increasing quarantine strength function. At the peak of its quarantine measures authors predict that about 70% of the infected population was effectively isolated and prevented from spreading infection to the non-infected population.