Technology is advancing, taking over, and almost surpassing humanity, as AI intuition beats out humans in data analysis where it concerns both establishing, recognizing and interpreting patterns.
- The Data Science Machine was able to beat out 68% of its human competition
- It managed to analyze and interpret data in just 12 hours, while humans took months
- It was proven to be between 87-96% as accurate as the top teams in three competitions
Researchers at the Massachusetts Institute of Technology (MIT) have designed the ‘Data Science Machine’. It’s a highly effective system that has been widely considered a breakthrough in the field of artificial intelligence. This high praise is due to the fact that it successfully worked without a human in an area where our very own intuition was thought of as critical.
The problem with big-data analysis is rooted in their tremendous content. It takes a long time to understand and create patterns. While computers have been recognized as exceptionally effective at detecting patterns, they lacked the actual understanding of the meaning behind them. However, the team of researchers at MIT believe to have found a solution.
They pinned the Data Science Machine against 906 human teams. While the system was not able to defeat all of them, it did beat out 615, which is an impressive number nonetheless. That amounts to being better than 67.9% of its competition. It was highly successful at determining certain patterns of unfamiliar data, as well as interpreting them. This is where human intuition has been considered to excel.
And while the engineered and programmed system was not able to defeat every single human team in the competition, it showed remarkable speed. While the participants mulled over the data for months, the Data Science Machine achieved the same goal within 12 hours. Sometimes, even more incredible, in just 2 hours.
The system was first tested by entering it in three competitions with human teams that implied its ability at finding predictive patterns in new data sets. In two out of three of those competitions, the Data Science Machine was astoundingly 94% or 96% as accurate as the top participants. For the third competition, it met with a more modest result of 87% as accurate, but the numbers still appear promising.
According to the researchers, there are multiple uses that could be found for their system.
They offered as an example, the situation where it needs to predict when a student would drop out of online classes. By detecting certain patterns, such as how long it takes for them to start their assignments or how active they are, the Data Science Machine will be to tell how likely it is for them to give up entirely.
It seems not to be limited by the lack of the human element, and instead is taking on tasks without the need of a person’s intuition. And, it’s actually excelling at it.
Image source: wtvox.com