Baidu AI robot trumps humans in talent reality show following Master’s Go victories

There are games that people naturally find difficult such as Go. There are also things like face recognition which we are intuitively good at. On the heels of Google’s Master, the updated version of AlphaGo, defeating world-renowned Go players, we just lost another battle to Artificial Intelligence (AI).

Xiaodu, Baidu’s AI-powered robot, beat top-notch brainy intellectuals on a face recognition competition in a popular talent reality television show, Super Brain, which aired on Jiangsu TV this past Friday. Super Brain is the third most watched television show in the country in 2016.

Think Super Brain as a weekly mental Olympics. The show aims to find talents with exceptional brainpower from around the world, pushing human limitations in areas such as memory, face recognition, and Rubik’s cubes.

Wang Feng, one of the show’s well known contestants, can remember 2,280 random numbers in an hour. Xiaodu the robot competes with Wang in this episode. They were both provided with 20 photos featuring girls from a local band in their toddler years. The task is to spot the grown-up girls based on the photos.

The tricky part was to identify a pair of twins from the girl band. Wang only managed to answer one right. However, the robot managed to spot both girls, and won the competition.

“What we found is that even the best scientists of today partially understand how the human being functions. We still don’t understand most of it,” said Andrew Ng, Baidu’s Chief Scientist. “So Xiaodu depends very little on our limited knowledge of how the brain works. It depends mostly on looking at lots of data and figuring out for itself how to complete the task.”

(from left to right) Andrew Ng, the host, and Xiaodu on Super Brain
(from left to right) Andrew Ng, the host, and Xiaodu on Super Brain

Xiaodu is part of Baidu Brain which was introduced at the Baidu World Conference in September 2016 by Robin Li, Founder and CEO of Baidu. Baidu Brain consists of four parts: voice recognition and synthesis, image recognition, natural language processing, and user profiling.

(All photos from Baidu)

No Comments Yet

Leave a Reply

© 2022 All Tech Asia. All Rights Reserved.

AllTechAsia is a startup media platform dedicated to providing the hottest news, data service and analysis on the tech and startup scene of Asian markets in English. Contact us:

%d bloggers like this: