Since 2013, Torobo, a robotic arm designed to test the limits of artificial intelligence, has had one ambition—to be admitted to the University of Tokyo. However, it has repeatedly failed. This year was no exception.

    In early November Torobo took an exam to prepare for an all-important standardized test. While Torobo's total score of 525 out of 950 was 14 points higher than in 2015，its standard score of 57.1% was slightly lower, because human students did better in the 2016 test.

    Though the score，which is higher than the national average, is enough to get the robot admitted to many other famous Japanese universities,
itfalls short of the minimum required for the University of Tokyo.

    A closer analysis of the results showed that Torobo's newly programmed ability to solve complex physics problems helped increase its
year-over-year physics score from 46. 5% to 59%. The robot also dug deep into its database of information from textbooks and websites to obtain
an impressive 66.3% in world history. The average 60% scores acquired in the two math sections, were not too shabby either.

    But，while Torobo is able to handle knowledge-based questions and ones involving complex mathematical calculations, it has a hard time thinking independently. As a result, the robot is unable to fully understand multiple sentences and phrases to arrive at the logical conclusions required in the English language tests. This weakness was reflected in the weak oral and writing scores, which came in at 36.2% and 50.5%, respectively.

    Fortunately, after four tiring years, the robot's creators have decided to free Torobo from its annual test-taking suffering, the research team will instead focus on improving the robot's ability to pick out specific answers from massive sets of data—a skill that led to its excellent history scores. They believe the expertise(专业技术）will  help Torobo succeed as an industrial robot.