HomeScience & TechnologyCan AI score your next test?

Can AI score your next test?

Artificial Intelligence (AI) could give a boost to online education by providing automated feedback to students.


See also: AI identifies areas on the Moon for landing and exploration

This spring, Philips Pham was among more than 12.000 people in 148 countries who took an online course called Code in Place. The course, which teaches the basics of computer programming, is taught by Stanford University.

Four weeks later, Mr. Pham, a 23-year-old student living in the southern tip of Sweden, took his first test, trying to write a program that could pull waves of tiny blue diamonds into a black-and-white grid. Several days later, he received a detailed review of his code.

See also: Cancer and AI: Researchers Develop Tool for Early Diagnosis of Prostate Cancer

"It seems you have made a small mistake," the reviewer noted. "Maybe you run to the wall after drawing the third wave."

The feedback was exactly what Mr Pham needed. And it came from a machine.

During this online class, a new type of artificial intelligence (AI) provided feedback to Mr. Pham and thousands of other students who took the same test. Built by a team researchers of Stanford, this automated system shows a new future for online education, which can easily reach thousands of people, but does not always provide the guidance many students need.

Dr. Chelsea Finn and her team designed this system exclusively for the Stanford programming class. But they used techniques that could automate the feedback students receive in other situations, such as classes beyond programming.

Oren Etzioni, CEO of the Allen Institute of Artificial Intelligence and former professor of computer science at University of Washington, warned that these techniques are far from the comments of human trainers. Comments and advice from teachers, teaching assistants and teachers are always preferable to an automated review.

See also: AI: How does it help us better understand the universe?

However, Dr. Etzioni described Stanford's work as a "step in an important direction," with automated feedback being better than nothing.

The new automated system is a way to reach more students than the instructors themselves could approach. And if he can clearly identify problems in the student code, showing the specific coding mistakes they make and how often they do, he could help them. trainers to better understand which students need help and how to help them.

Source of information:

Teo Ehc
Be the limited edition.