Scientists at the Allen Institute for Artificial Intelligence (AI2) have come up with a new computer system that can solve geometry problems alike an average high school student taking SAT. The artificial intelligence system rightly answered 49% geometry questions, scoring roughly 500 out of 800 in the entire math section of SAT, which is on par with average score in 2015 test.

Scientists have called this system GeoS. This is obviously a groundbreaking achievement but hasn’t matched up a level right now where it can take over human intelligence. For the creation of the computer system, the AI2 did partnership with University of Washington Computer Science department. The system can perfectly interpret diagram and geometric shapes and has the ability to use natural language for reading the questions.

Oren Etzioni, CEO of AI2, said that unlike the Turing Test, standardized tests like the SAT has provided them today with a way to calculate machines ability to reason and to compare its abilities with that of a human.

Etzioni said, “Much of what we understand from text and graphics is not explicitly stated, and requires far more knowledge than we appreciate. Creating a system to successfully take these tests is challenging, and we are proud to achieve these unprecedented results”.

With the help of text and diagram, GeoS interprets the questions, then matches it with logical answers and passes it to geometric solver to solve. In the end, it compares the answer with multiple choice options provided in the paper. That’s how GeoS exactly works.