LZ v Golaxy

John Fairbairn · Post by **John Fairbairn** » Fri Apr 20, 2018 11:39 am

Your point is taken with regards to overexaggeration causing negative impact on the message itself.

Outside of the realms of Go, the whole PR stunt pulled by Deepmind with the Stockfish thing was pretty outrageous as well, esp considering the weight they had in the community and the public influence and broad exposure... many in the Chess community echo'd the same sentiment that Deepmind didn't really play a fair match when it claimed victory. It seems shenanigans happens on every level... So okay perhaps the Golaxy thing wasn't the worst thing to happen to Go match in 5000 years, but to say that they could have done much better is probably also an understantment. .

I can feel T Mark looking down on this from the galaxy and itching to be able to respond, so I'll do it for him with one of his favourite sayings: "I've told you a million times not to exaggerate."

macelee · Post by **macelee** » Fri Apr 20, 2018 12:21 pm

hydrogenpi7 wrote: These matches were a farce for Golaxy PR. Its not serious and nothing more than a joke

I totally agree with some other members on this forum - you exaggerated here as well.

I am not sure how good PR is defined. But this event definitely reached Go fans very well. Simple statistics: the most popular game in this series had about 490,000 views on the Yikeweiqi platform. In this game alone there were 934 comments posted by the spectators. To put this in context, this week's biggest pro tournament game - the deciding game for the Chinese Tianyuan title - had 996,000 views; yesterday's Changqi Cup game with Ke Jie attracted 638,000 viewers.

As far as I can see, the PR worked very well. Golaxy got plenty of media attention, before its planned big game against Ke Jie. Mr You, the author of Abacus and also behind Golaxy, is almost a celebrity in the Chinese Go circle.

Uberdude · Post by **Uberdude** » Fri Apr 20, 2018 1:28 pm

macelee, were there any more details on the design of Golaxy? I heard it is neural network with something like 19 blocks, so it doesn't sound as novel as suggested.

macelee · Post by **macelee** » Sat Apr 21, 2018 12:12 am

Uberdude wrote:macelee, were there any more details on the design of Golaxy? I heard it is neural network with something like 19 blocks, so it doesn't sound as novel as suggested.

I've not read anything concrete. But people are talking about how Golaxy plays more like a human. Also the developers seem to suggest that Golaxy doesn't play for safety when it leads a lot. Golaxy is also said to support arbitrary board size and arbitrary komi, which is special.

macelee · Post by **macelee** » Mon Apr 23, 2018 10:08 am

It turns out that Golaxy has another aspect that is more similar to human than to other bots. Instead of giving a winning percentage, Golaxy gives its positional judgement in points (e.g. black leads by 6.2 points) as if it can count.

johnsmith · Post by **johnsmith** » Mon Apr 23, 2018 11:37 am

macelee wrote:It turns out that Golaxy has another aspect that is more similar to human than to other bots. Instead of giving a winning percentage, Golaxy gives its positional judgement in points (e.g. black leads by 6.2 points) as if it can count.

So it turns out that if you like territory and points you should play 3-3 in all the corners you can. Which makes sense

Uberdude · Post by **Uberdude** » Wed Apr 25, 2018 7:34 am

Golaxy is playing a series of 30 no-komi games against pros (I guess on yikeweiqi server), macelee has put them in go4go. Humans won 2 of the 27 he has (well done Kang Dongyun and Yun Chanhee!). It beat Park Junghwan, I think the final game will be against Ke Jie?
http://www.go4go.net/go/games/byplayer/1992

Hane · Post by **Hane** » Wed Apr 25, 2018 8:58 am

Golaxy， while adopting AlphaGo‘s basic architecture， has brought innovations into feature combination， model structure， and MCTS calculation methods。 It manages to complete model training with less computing resources or training samples， yet with the potential to achieve the highest possible Go level by AI。

Golaxy tries to play Go in a style that is easier for human players to understand and to learn。 It can play Go on boards of all sizes and all komi settings。 Golaxy will not concede even when leading。

EdLee · Post by **EdLee** » Wed Apr 25, 2018 9:03 am

Golaxy will not concede even when leading。

Seems typo: when losing.

sorin · Post by **sorin** » Wed Apr 25, 2018 9:07 am

Uberdude wrote:Golaxy is playing a series of 30 no-komi games against pros (I guess on yikeweiqi server), macelee has put them in go4go. Humans won 2 of the 27 he has (well done Kang Dongyun and Yun Chanhee!). It beat Park Junghwan, I think the final game will be against Ke Jie?
http://www.go4go.net/go/games/byplayer/1992

I noticed in the Yun Chanhee that Golaxy's opening was very unusual- two moves in 5x7, and two moves in 6x6, while black's moves were all orthodox - I wonder if it was the human operator choosing the first 4 bot moves manually just to make the game more interesting?

sorin · Post by **sorin** » Wed Apr 25, 2018 9:09 am

EdLee wrote:
Golaxy will not concede even when leading。
Seems typo: when losing.

I think it is meant to be: "will not play slack/stupid-looking moves when ahead".

Hane · Post by **Hane** » Wed Apr 25, 2018 9:16 am

source : http://sports.sina.com.cn/go/2018-04-23 ... 5857.shtml

EdLee · Post by **EdLee** » Wed Apr 25, 2018 9:43 am

Thanks, Sorin; makes more sense that way.

Golaxy will not concede even when leading。

An issue here is the definition of concede:
- in terms of the final points difference; or
- in terms of the winrate.

They're implicitly (and selectively) using the former meaning, which is quite human-centric, and a pre-AlphaGo idea.
Whereas AFAIK AlphaGo/AlphaZero-class bots also don't concede in the latter meaning, when ahead. ( AG's winrate monotonically non-decreasing, when AG is ahead. )

macelee · Post by **macelee** » Thu Apr 26, 2018 1:51 am

sorin wrote: I noticed in the Yun Chanhee that Golaxy's opening was very unusual- two moves in 5x7, and two moves in 6x6, while black's moves were all orthodox - I wonder if it was the human operator choosing the first 4 bot moves manually just to make the game more interesting?

I heard that there was some software configuration issues that resulted the rather unusual opening. Also note that there was a very large-scale seki - not sure if that caused difficulties to the AI.

macelee · Post by **macelee** » Thu Apr 26, 2018 4:21 am

macelee wrote:
sorin wrote: Also note that there was a very large-scale seki - not sure if that caused difficulties to the AI.

Golaxy has completed the 30-game series. The total score is 28-2.

Apparently Golaxy has no problem with seki! In fact in game 30 against Zhou Ruiyang, they managed to create a 4-group seki situation. This remind me a recent LG cup game viewtopic.php?p=229331#p229331

Click Here To Show Diagram Code: [go]$$W $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . O . . | $$ | . . . X . . . . . . . O O O X . X O . | $$ | . . X X O . . X . O . X O X X X O . . | $$ | X . O O X X X X O , X . X . X X O . . | $$ | X X . X O O O X O . . X X X O X O . . | $$ | X O X X X . . O O . X . . . O O X O . | $$ | X O O O X O . . . . X O . O . . X O . | $$ | O O O X O . . . . . X O . O X O X . . | $$ | O X X X O O . . . O O X . X O . X . . | $$ | O O . X O . . . . , . . X . O , . . . | $$ | O X X X O . . . . . . . . . . . . . . | $$ | O O X X O . . . . . . . . . . . . . . | $$ | . X . X O O O . . . . . . . . X X . . | $$ | X X X O O X O . . . . . . . . . O . . | $$ | X O X X O X O . . . . . . . . . . . . | $$ | X O O O X X O . . , . . X . . , . . . | $$ | X O . O X O O . . X . . . . X . O . . | $$ | O O X O X O O X . . . . . . . . . . . | $$ | . X X X X O . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

Life In 19x19

LZ v Golaxy

Re: LZ v Golaxy

Re: LZ v Golaxy

Re: LZ v Golaxy

Re: LZ v Golaxy

Re: LZ v Golaxy

Re: LZ v Golaxy

Re: LZ v Golaxy

Re: LZ v Golaxy

Re: LZ v Golaxy

Re:

Re: LZ v Golaxy

Re: LZ v Golaxy

Re: LZ v Golaxy