Mankind has made great advances in technology but has not been able to replace the human element. But slowly this barrier is also being overcome. Data analysis by artificial intelligence is based on fixed parameters. However it was not possible to identify patterns which can influence a foregone outcome. Identifying such patterns requires human intuition which cannot be done by artificial intelligence. MIT researchers have created software for just this kind of big data analysis. Its Data Science Machine beat 615 out of 908 human teams in 3 data science competitions.
Massachusetts Institute of Technology (MIT) scientists are going to take on the question of human intuition head on in big data analysis and let computers search for predictive patterns. MIT scientists are planning to take this big risk in data analysis at the Computer Science and Artificial Intelligence Laboratory (CSAIL). It is known as the Data Science Machine and to test its prowess, got it listed in three data science competition. Pitted against 906 teams, it successfully defeated 615 out of 906 human teams
In the two competitions, data Science Machine managed a score of 94 percent and 96 percent accuracy while in the third competition it managed a score of 87 percent accuracy. While the human teams took months to detect the patterns, the Data Science Machine predictions were accomplished in two to 12 hours.
The foundation of the Data Science Machine is based on a science master thesis developed by Max Kanter. Kanter said that the Data Science Machine must be viewed as a natural complement to human intelligence. There is an infinite volume of Data which is sitting there needing analysis and this machine will give us a tool to accomplish the impossible.
Big data is a gargantuan and multifaceted network and much of its algorithm is automated, human element is still required to search for features which will finally reveal the secret patterns. It is here that human intuition is handy.