Can we use artificial intelligence technologies such as Machine Learning and Deep Learning to help in finding medicine or vaccine for COVID-19. Any ideas?
you question two parts ans, 1) not really solution since the big.. AI required a lot of data, the data we have mostly the counts such deaths and infected which will not serve no purpose
2) using AI are have come close to find the RNA of the virus, since it is just one more form SARAS which crippled the nations about decade and half back, but the problem is the vector which is again different from the previous once which required an different approach and today they were able to find the what it effects, while it effects you dont have right compound to apply directly need to be test and i will take own time
AI requires big data for modelling possible alternative scenarios versus a given problem. In the case of COVID, an AI engine should be fueled with historical, actual, geographic, gender, age of infected-recovered-deaths, duration from infection to symptoms, other population and lots of additional data. It will also depend on what type of AI engine will be used. Neural networks will be very suitable for learning purposes, once the optimal model has been designed. Combination with Neural networks and Fuzzy logic will allow data customization. This is indeed a very complex, but interesting field for an adequate AI engine to be developed. However, as already explained above, all depends on a trustfull data input.
As an AI is a continous learning system, we already should use existing data to start the learning process. Major dynamics are already to be classified as past historical data, actual and future data are quite important and deviations from standard values will be expected after the reverse lockdown.
Yes absolutely we can empower our COVID research process by using AI . Infact clinical trials that happen use AI and COVID is no exception. Since it has started recently and data collection is still underway we might have to wait for a while.
In fact, waiting till when? We should start collecting now all actual available and of course step by step new incoming data. Waiting is definitely the worst option.
As for similar models in other scientific disciplines where I'm directly involved in (energy consumption and efficiency linked to process optimization), my best results were and are always achieved when collecting historical data up to the time of the change and then follow up with new data. This option saves time and provides quick and best results.