Article

Find, Search, Reprint, Submit Articles For Free

Chinese Professor Trains AI Programs to Spot Beauties and Criminals


file1.jpeg

(YicaiGlobal) Dec. 21 — After training a program to distinguish between criminals and non-criminals with a success rate of 86 percent, Prof. Wu Xiaolin of Shanghai Jiaotong University has launched a service using artificial intelligence to identify good-looking women, according to tabloid news site Dzh.mop.com.

Wu said the current facial recognition system can identify people’s biological characteristics and the next task is extremely attractive and challenging — whether AI can speculate on the social characteristics of human beings based on facial recognition.

Once the program managed to differentiate between shots of criminals and ordinary people, Wu turned his research to women. He is trying to get AI to judge whether they look attractive and compare the results with the opinions of male students. His program’s aesthetic standards are highly consistent with those of male students.

His research team used 3,954 photos of Chinese women, including 2,000 ‘commended’ photos and 1,954 ‘derided’ photos grouped by Wu’s male graduate students. After initial trials, the programs can now identify the photos that got the thumbs up from those that got the thumbs down with an accuracy rate of 80 percent.

Compared with identifying criminals, spotting beauties is tougher for AI programs because the appreciation of beauty is a complex personal judgment integrating the individualities and social values of both the observer and the observed.

 

Once the program managed to differentiate between shots of criminals and ordinary people, Wu turned his research to women. He is trying to get AI to judge whether they look attractive and compare the results with the opinions of male students. His program’s aesthetic standards are highly consistent with those of male students.

His research team used 3,954 photos of Chinese women, including 2,000 ‘commended’ photos and 1,954 ‘derided’ photos grouped by Wu’s male graduate students. After initial trials, the programs can now identify the photos that got the thumbs up from those that got the thumbs down with an accuracy rate of 80 percent.

 

Once the program managed to differentiate between shots of criminals and ordinary people, Wu turned his research to women. He is trying to get AI to judge whether they look attractive and compare the results with the opinions of male students. His program’s aesthetic standards are highly consistent with those of male students.

His research team used 3,954 photos of Chinese women, including 2,000 ‘commended’ photos and 1,954 ‘derided’ photos grouped by Wu’s male graduate students. After initial trials, the programs can now identify the photos that got the thumbs up from those that got the thumbs down with an accuracy rate of 80 percent.

Domain info

Other articles