Exclusive: Ke Jie wants you to know weiqi is fun -- and easy to learn
Vložit
- čas přidán 26. 05. 2017
- Ke Jie, the world's best weiqi (Go) player, cried after losing 0-3 to the artificial intelligence program AlphaGo. He took off his glasses and wiped his eyes. But he's happy to see the world focus on his beloved sport. The star player did not talk much about future. He needed some rest, after the high-profile matches. But Ke's passion for weiqi has not wavered. He wants to tell everyone that weiqi is fun, and contrary to the stereotype, it's easy to learn.
Subscribe to us on CZcams: goo.gl/lP12gA
Watch CGTN Live: • CGTN Live
Download our APP on Apple Store (iOS): itunes.apple.com/us/app/cctvn...
Download our APP on Google Play (Android): play.google.com/store/apps/de...
Follow us on:
Facebook: / chinaglobaltvnetwork
Instagram: cgtn?hl=zh-cn
Twitter: / cgtnofficial
Pinterest: / cgtnofficial
Tumblr: / cctvnews
Weibo: weibo.com/cctvnewsbeijing
![[9X9] #1 starting Go quest! climbing up from low kyu and i don't know any josekies here](/img/n.gif)
Ke Jie carried a really heavy burden being the strongest human Go player and facing AlphaGo formally. It is sad to hear from him that he didn't play good enough. But the matches were nice and he was the right person to represent humans in this challenge. Thanks, Ke Jie!
It's not so much that he didn't play well enough, just that his defeat wasn't as "exciting" (literal) as I suppose he'd hoped
"I just hope people can feel the fun of Weiqi, and not too concerned about win or lose." (c) Ke Jie
Before I thought Kejie was just another arrogant competitor, but after these post-match interviews, I've completely seen more of him which has thoroughly changed that thought. He's a genuine person and down to earth person.
I feel the same way! He actually seems so likeable in interviews
Great last point. Weiqi is about bringing people happiness. Thank you brilliant young man.
Such a cool guy :)
It is like the fastest human runner met a car. The first car may be slow but once the car evolved... there was no way for a human to outrun it.
AI girlfriend ? wtf is that question?
There will be a day were ai won't pass the turing test and you won't ever know if the person you're talking to is a robot or a person. Would you want a robot that could literally be nothing but perfect, would always say the right thing. Would only be after your interests at heart. Sounds pretty great when you actually sit down and think about it.
"but she's not human" If you're not told, you literally wouldn't be able to know when intelligence gets good enough.
But the real question is... Can she (AI) cook?
@@jayvis123111 You mean will pass the Turing test
@@jessstuart7495 no, the real question is. will she force him to cook?
He actually looks like a asian Magnus Carlsen
carlsen looks like a european ke jie.
wise words..
Remain calm, nothing is under control.
The first game was really tough... A rematch would be great!
I'm not so sure. Yes, the game looks close (W+0.5), but ... I ran the game through KataGo. For the first 100 or so moves, according to KataGo, AlphaGo just chipped away, never giving an inch, gaining slowly but surely on every single move. By move 100, AlphaGo was up by 10 points with a 97% win probability. Then, AlphaGo started playing safety moves. AlphaGo's move 128 in particular was a slack move locally (according to KataGo --- I have no clue myself --- I'm just a lowly 1-dan) and gave up almost two points, but it simplified the position and removed possibilities for black, leaving the win probability unchanged at 98%. AlphaGo continued to slowly give back points in exchange for a simpler position --- ultimately leaving black with an extremely close but totally unwinnable game.
So: when the game against AlphaGo looks close --- don't be fooled. "Close" in terms of score, sure. "Close" in the sense of "gosh, I almost won, I just needed to find one more point somewhere" --- no way. Ke Jie was closest to winning at the first move, and it only went downhill after that. Letting the score get really close while keeping its win probability at 99% is just AlphaGo's way of taunting you --- giving you imagined hope while actually removing every last chance of victory. It's like Ke Jie said: AI rules the Go universe; the rest of us just play around in the back yard.
態度一級棒。
He thought steering the game into a high level of complexity was the way to try to win. However that is the last thing you should do against a strong AI. The human should make the game strategically complex but not tactically complex. Even so a well trained neural net is almost unbeatable. The same thing was done with backgammon. Strong programs were developed which used neural nets. These backgammon AI's are almost unbeatable in a match.
I don't think it's quite that predictable. Different AI:s will be good at different things just as different people are good at different things. So it makes sense to try to derive conclusions based on how a paticular AI seems to play. That a rougher game would give a human a chance was a good guess based on how rediculously smooth the AI tended to play as a default.
If it is a job, it is not fun.
I wish the Chinese players would have come out and apologize also for calling Lee Sedol a fraud who was paid to lose to AlphaGo. I am glad Ke Jie now understand the power of AlphaGo, and learn some humility.
No need to apologize, because Alphago that Ke Jie competed with is 100 time stronger than before
For me alpha go isn't a AI, it's only a data base
There’s a lot of math in it to make sure it learns the “ data base” in the most efficient way. No computer could go through all possible moves, which has the number of over 10^80. It can learn it’s “data base” effectively and independently, which is certain level of intelligence.
"Have you ever considered having an ai girlfriend?" What a stupid and sexist question! Have you ever considered being an ai's wife? Has it come to your mind he might be gay?