After Go victory, challenges given to 'AlphaMahjong'

AlphaGo后，是否接下来会有Alpha麻将？比较有意思。

---

The world’s top Go player Lee Sedol reviews the match after the fourth match of the Google DeepMind Challenge Match against Google’s artificial intelligence program AlphaGo in Seoul, South Korea, in this handout picture provided by Google and released by News1 on March 13, 2016.

BEIJING – Glib Chinese web users have raised a challenge to the AI program that recently dethroned one of the top human Go players, demanding tongue-in-cheek that the AlphaGo program learn the nation’s real pastime - mahjong.

China’s reaction toward the historic duel between human and artificial intelligence has been mixed. At first, there were questions on why the human player did not come from China, where the game was invented more than 2,500 years ago.

Then, as AlphaGo marked three victories in the five-game match, defiant web users began calling for a challenge in an arena average Chinese are more comfortable in.

“Can AI beat mahjong masters?” was posted several times in the comment section of stories about the AlphaGo-Lee Sedol match on Chinese microblog Sina Weibo, where they attracted a deluge of comments defending humanity’ glory in mahjong.

Mahjong, China’s answer to poker, is usually played by four people. Each turn, players draw tiles from a 144-tile pool, discard or intercept others’ to form sets of tiles that can win. Scientists say compared with Go, mahjong has far fewer permutations for AI calculation, but involves a degree of chance and other factors in favor of humans.

“Sometimes with a glimpse of the other player’s facial expression, I know how he or she is going to play. Can AI do that?” asked one blogger.

“Unlike Go, mahjong is not a quiet game that focuses on calculation. It involves a lot of interactions and teamwork between players,” another blogger commented.

Some netizens described mahjong as a competition on both IQ and EQ. Computers can undoubtedly blow humans out of water in math, but how about their ability of communication and interpreting emotions?

Scientists have advised against exaggerating AI’s prowess after the Go victory. In the gaming field alone, AI researchers agree that computers, in their current stage, may not be able to beat humans in games that feature a high level of irrationality, including understanding of feelings and creativity.

“AlphaGo’s learning and calculation prowess is based on formulas and data, but mahjong involves luck and emotions. To my knowledge, no AI is powerful enough to manage them,” said Ren Yi, an expert on computer game AI and CEO of a Nanjing-based tech company.

Feng Jianfeng, chief scientist for a brain science and AI project launched by the city of Shanghai, also refused to bet on AI’s definitive victory over human mahjong masters, as the game involves more human intelligence than calculation.

“AlphaGo’s structure and arithmetic only emulates some primary functions of human brains, not including emotions, decision making, attention and creativity,” Feng said. “It has just done something computers are very good at. It is a landmark and will greatly boost AI development, but AI still has a long way to go.”

Feng’s team is studying Alzheimer’s disease with the help of AI, but defying public expectations of AI supplementing patients’ declining brain abilities, Feng said what computers could do so far was helping make more accurate diagnoses by navigating through big data.

The disease affects tens of millions of elderly people worldwide and has no effective cure, though medical experts say playing games that activate brains could help with the prevention.

So for those worrying about the disease, the Chinese advice would be, instead of pinning hopes on AI, to sit with friends and enjoy a game of mahjong.