Since a team of computer scientists from the Massachusetts Institute of Technology (MIT) were able to design a system that can outperform human intelligence when looking for hidden patterns in big chunks of data, humans may soon become obsolete in big data analysis.
Until now, data analysis and data mining were processes that required the intervention of human element to be successful. Looking for relevant patterns in heaps of unfamiliar data required human intuition in order to provide positive results.
But MIT scientists claim that their “Data Science Machine” is faster and often more reliable than humans when it comes to analyzing big data and extracting patterns that may have some type of predictive power.
The MIT team said that their machine participated in three competitions against humans and it finished ahead of 615 human teams. Overall, more than 900 human teams took part in the competitions.
In the first two competitions, the Data Science machine had an accuracy of 94 percent to 96 percent as compared with the human winners of the competitions. In the third competition, the accuracy dropped to 87 percent.
But while human competitors needed several months to come up with their final prediction algorithms, the MIT machine needed only two to 12 hours.
“We view the Data Science Machine as a natural complement to human intelligence,”
said Max Kanter, one of the researchers who designed the big data system.
Kanter noted that there are huge amounts of data out there that no one analyzed yet. He also said that he hoped the new machine may at least help humans get started.
A paper on the machine was co-authored by Kanter and his mentor Kalyan Veeramachaneni of the MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL).
CSAIL researchers are now interested in coupling machine-learning techniques with big data analysis. A machine may soon tell what the chances for a particular student to fail his or her English Literature exam are or how much energy potential certain wind-farm sites may have. Yet some researchers are concerned that humans may soon become obsolete in big data analysis.
The Data Science Machine already predicted with great accuracy what the chances for a student to drop online classes were. But the system needed at least two algorithms to make predictions. It analyzed how much time the student spent on the online courses, and how much time the student had to submit an assigment before an incoming deadline.
Margo Seltzer, a computer scientist from Harvard University, said that the new machine does not only solve problems, but it also offers a different perspective on the problems.
Image Source: Pixabay