Computer system created by AI2 achieved 49% accuracy in geometry portion of SAT test
Scientists at the Allen Institute for Artificial Intelligence (AI2) have developed a new computer system that can solve geometry problems just like an average high school student taking SAT. The artificial intelligence system correctly answered 49% geometry questions and roughly scored 500 out of 800 in the entire math portion of SAT, which is on par with the average score in 2015 test.
Buy Now: Sony PlaysStation VR In Stock Here
They are calling this system GeoS and it is certainly a groundbreaking achievement but not up to that level right now where it can take over human intelligence. The AI2 collaborated with University of Washington Computer Science department for the creation of this computer system. The system is perfectly capable of interpreting diagram and geometric shapes and using natural language to read the questions.
“Unlike the Turing Test, standardized tests such as the SAT provide us today with a way to measure machines ability to reason and to compare its abilities with that of a human,” said Oren Etzioni, CEO of AI2. “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.”
GeoS uses text and diagram to interpret the questions, then matches it with logical possible solution and send it over to geometric solver to solve. Lastly, it compares the answer with multiple choice options given in the paper. That is exactly how GeoS works.
“We’re excited about GeoS performance on real-world tasks,” said Ali Farhadi, a senior research manager at AI2. “Our biggest challenge was converting the question to a computer-understandable language. One needs to go beyond standard pattern matching approaches for problems like solving geometry questions that require in-depth understating of text, diagram and reasoning.”
Don't Miss: iPhone 8: Everything You Need to Know
Right now, GeoS can answer plane geometry questions but soon it will get upgraded and scaled up to solve the entire math section of SAT and AI2 professionals are expecting to get it done in the next three years. They are also planning to build systems which will be able to tackle science related tests, meaning human-like common sense and knowledge need to be incorporated into them.