It is currently Fri Mar 29, 2024 1:49 am

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 123 posts ]  Go to page Previous  1 ... 3, 4, 5, 6, 7  Next

Who would win?
AlphaGo win against both Ke Jie and Chinese team 87%  87%  [ 46 ]
AlphaGo lost against Ke Jie but win against Chinese team 4%  4%  [ 2 ]
AlphaGo win against Ke Jie but lost against Chinese team 6%  6%  [ 3 ]
AlphaGo lost on both event 4%  4%  [ 2 ]
Total votes : 53
Author Message
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #101 Posted: Wed May 24, 2017 9:58 pm 
Lives in gote

Posts: 470
Liked others: 62
Was liked: 278
Rank: UK 2d Dec15
KGS: mathmo 4d
IGS: mathmo 4d
the chinese commentators (gu li, etc) have been impressed by ke jie and are actually quite optimistic

Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #102 Posted: Wed May 24, 2017 10:06 pm 
Lives in gote

Posts: 470
Liked others: 62
Was liked: 278
Rank: UK 2d Dec15
KGS: mathmo 4d
IGS: mathmo 4d
the feeling is maybe the top right was wrong for ke jie, but actually demis said it was perfect play for the first 30 moves. Next, they criticised Ke Jie for playing on the left and not protecting on the lower side. Alphago played some very strange and crude moves, but then that it seemed W was in serious trouble. But it was a huge complicated mess with the life and death of 6-10 groups unclear.

Then Ke Jie played a brilliant move on the upper side preparing for the ko.

Then on, they can't find a way out for alphago from the complications due to the huge ko on the lower left that could kill B's group, and so they've been optimistic for quite a while though it is too complicated to say. And they still think Ke is playing very well, finding better moves than they've noticed.


Last edited by dhu163 on Wed May 24, 2017 10:15 pm, edited 1 time in total.
Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #103 Posted: Wed May 24, 2017 10:10 pm 
Lives in gote

Posts: 470
Liked others: 62
Was liked: 278
Rank: UK 2d Dec15
KGS: mathmo 4d
IGS: mathmo 4d
Gu Li just said he thinks alphago's winning percentage is probably higher for Ke Jie!

Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #104 Posted: Wed May 24, 2017 10:13 pm 
Lives in gote
User avatar

Posts: 392
Liked others: 23
Was liked: 43
Rank: NR
long ago, i predicted that Ke Jie would win on points in the second round, to keep the audience enthralled, only to be crushed by an uppercut in the third. let's see what happens on Saturday :)

_________________
i shrink, therefore i swarm

Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #105 Posted: Wed May 24, 2017 10:19 pm 
Lives in gote

Posts: 470
Liked others: 62
Was liked: 278
Rank: UK 2d Dec15
KGS: mathmo 4d
IGS: mathmo 4d
unsurprisingly Gu Li now notices good moves for alphago.

he predicts several different options that might happen since often W might be behind by a ko threat.

Or B ends the ko, W captures the top right, seals in the centre B group for a semeai where W loses the lower right. And W tries to kill the entire top side.

Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #106 Posted: Wed May 24, 2017 10:25 pm 
Lives in gote

Posts: 470
Liked others: 62
Was liked: 278
Rank: UK 2d Dec15
KGS: mathmo 4d
IGS: mathmo 4d
now he thinks ke jie made a big mistake with too small a ko threat

Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #107 Posted: Wed May 24, 2017 10:40 pm 
Lives in gote

Posts: 384
Liked others: 22
Was liked: 98
Rank: KGS 2d
I very much appreciate that Ke Jie is trying all different strategies in these games. AlphaGo may not have thought the mid-game was all that complicated, but it was pretty exciting to see for human players.

Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #108 Posted: Wed May 24, 2017 10:41 pm 
Lives in gote
User avatar

Posts: 392
Liked others: 23
Was liked: 43
Rank: NR
Michael and Andrew agreed that it looks like this Alphago is making less reliance on Monte-Carlo than last year - but i'm not so sure... the biggest change in her is the introduction of TPUs, and it's possible that this enables her convolution inputs to scope bigger quantities of pixels - possibly even the whole board. that would make a huge benefit - but it's also possible that the TPUs can do with hardware what used to be done with software insofar as search is concerned, so it could be that her Monte-Carlo search is much broader than before as well. Stephanie mentioned her deep reading... however, my prediction for Round 2 has just gone wrong so i could be wrong about this too :)
https://books.google.com.au/books?id=AT ... lo&f=false
PS oh, no, wait... they were using TPUs last year as well
https://cloudplatform.googleblog.com/20 ... -chip.html
i guess the photo in the article is the sort of thing they mean when they say Alphago is running on a single machine - try balancing that lot on your laptop knees :)

_________________
i shrink, therefore i swarm


Last edited by djhbrown on Wed May 24, 2017 11:22 pm, edited 2 times in total.
Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #109 Posted: Wed May 24, 2017 10:42 pm 
Lives in gote

Posts: 470
Liked others: 62
Was liked: 278
Rank: UK 2d Dec15
KGS: mathmo 4d
IGS: mathmo 4d
yes, ke jie resigned

but the game was a spectacular one

Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #110 Posted: Wed May 24, 2017 10:55 pm 
Lives in gote
User avatar

Posts: 392
Liked others: 23
Was liked: 43
Rank: NR
their post-game analysis looks even more dramatic than the game; hope someone is recording it, with sound

_________________
i shrink, therefore i swarm

Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #111 Posted: Thu May 25, 2017 12:01 am 
Lives in gote

Posts: 470
Liked others: 62
Was liked: 278
Rank: UK 2d Dec15
KGS: mathmo 4d
IGS: mathmo 4d
hassabis did comment that they've never seen alphago put under so much pressure, the first 15 moves alphago thought were perfect and up to move 100 it was extremely close. Ke Jie said he thought he was winning around move 100 but said he got too nervous and hasty.

Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #112 Posted: Thu May 25, 2017 12:20 am 
Lives in sente

Posts: 727
Liked others: 44
Was liked: 218
GD Posts: 10
Ke Jie request to play as white in the next game (he cite white has more winrate) so that he can give exciting match and Demis Hassabis said yes

Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #113 Posted: Thu May 25, 2017 3:13 am 
Judan

Posts: 6725
Location: Cambridge, UK
Liked others: 436
Was liked: 3719
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
For discussion of the games themselves without off-topic chit-chat: viewtopic.php?f=13&t=14250.


This post by Uberdude was liked by: Bonobo
Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #114 Posted: Thu May 25, 2017 7:35 am 
Lives with ko

Posts: 129
Liked others: 20
Was liked: 17
Not much to see here. Sooner than expected, but still. AlphaGo is to Go what was DeepBlue to Chess. Hopefully, we'll see engine development speed up, hopefully we'll see engine tournaments on a (semi) professional level (like e.g. TCEC in chess).

Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #115 Posted: Thu May 25, 2017 11:09 am 
Lives in gote

Posts: 676
Liked others: 6
Was liked: 31
KGS: 2d
Sneegurd wrote:
Not much to see here. Sooner than expected, but still. AlphaGo is to Go what was DeepBlue to Chess. Hopefully, we'll see engine development speed up, hopefully we'll see engine tournaments on a (semi) professional level (like e.g. TCEC in chess).


...hopefully we'll see better conditions for humans, e.g. 2-3 days for a game like in the old days, so that a player can really study a position or a database of 100.000 AG-self-games or at least some games before a competition.

Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #116 Posted: Thu May 25, 2017 2:27 pm 
Lives with ko
User avatar

Posts: 160
Location: Italy
Liked others: 238
Was liked: 60
Rank: SDK
Sneegurd wrote:
AlphaGo is to Go what was DeepBlue to Chess.

I liked better the following sentence (given by Demis Hassabis I think)
Quote:
where Deep Blue have been the end, AlphaGo is the beginning.

In the first game, Ke Jie played so many of the new moves learned by AlphaGo games and many Pro players confirmed that the AI impact on Go theory have been huge in just two years, compared to the millennia that Go existed.
I am only a beginner, but I enjoy this game and I am happy to share this opportunity.
Galation

_________________
When you play Weiqi you are joining millions of people across four thousand years of time.
Jonathan Hop - So You Want to Play Go?

Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #117 Posted: Fri May 26, 2017 6:33 am 
Lives in gote
User avatar

Posts: 310
Location: Deutschland
Liked others: 272
Was liked: 126
Rank: EGF 4 kyu
Sneegurd wrote:
Not much to see here. Sooner than expected, but still. AlphaGo is to Go what was DeepBlue to Chess. Hopefully, we'll see engine development speed up, hopefully we'll see engine tournaments on a (semi) professional level (like e.g. TCEC in chess).


I don't think AlphaGo is comparable to DeepBlue. I think AG is far more generalised and the difference in ability between AlphaGo and top humans is way, way more marked. After all, DeepBlue won but it was nowhere near a whitewash - it took ten years for chess AIs to progress to that point and I think AG is already there.

Pippen wrote:
...hopefully we'll see better conditions for humans, e.g. 2-3 days for a game like in the old days, so that a player can really study a position or a database of 100.000 AG-self-games or at least some games before a competition.


I agree with the desire for longer time for the humans but did humans "in the old days" have the option to do research during that time? I think they could only use their own mental faculty - not databases. I also wonder whether the modern style would really benefit from extremely long time controls.

Personally, I have absolutely zero interest in a game where a human player is using a database and playing against an AI. To my mind, that would be the most un-Go-like thing you could have. I'd hate to see a future where sequences and moves have to be "legitimised" by databases or bots before humans are prepared to try them out.


This post by Charlie was liked by: Waylon
Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #118 Posted: Fri May 26, 2017 9:02 am 
Lives in sente

Posts: 727
Liked others: 44
Was liked: 218
GD Posts: 10
Stephanie Yin calls AlphaGo slack move 'Announcement Move' and Hajin Lee like this name very much. Pro team, however...

Image


This post by pookpooi was liked by 11 people: Baywa, Bill Spight, Bonobo, daal, ez4u, fwiffo, gamesorry, jptavan, mhlepore, Waylon, xiayun
Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #119 Posted: Sat May 27, 2017 2:28 am 
Lives in sente

Posts: 727
Liked others: 44
Was liked: 218
GD Posts: 10
Image

Top
 Profile  
 
Offline
 Post subject: Re: Predict AlphaGo on Future of Go Summit in Wuzhen, China
Post #120 Posted: Sat May 27, 2017 6:34 am 
Oza
User avatar

Posts: 2221
Location: Germany
Liked others: 8262
Was liked: 924
Rank: OGS 9k
OGS: trohde
Universal go server handle: trohde
(Emphasis added)

DeepMind wrote:
While AlphaGo is stepping back from competitive play, it’s certainly not the end of our work with the Go community, to which we owe a huge debt of gratitude for their encouragement and motivation over the past few years.

We plan to publish one final academic paper later this year that will detail the extensive set of improvements we made to the algorithms’ efficiency and potential to be generalised across a broader set of problems. Just like our first AlphaGo paper, we hope that other developers will pick up the baton, and use these new advances to build their own set of strong Go programs.

We’re also working on a teaching tool - one of the top requests we’ve received throughout this week. The tool will show AlphaGo’s analysis of Go positions, providing an insight into how the program thinks, and hopefully giving all players and fans the opportunity to see the game through the lens of AlphaGo. We’re particularly honoured that our first collaborator in this effort will be the great Ke Jie, who has agreed to work with us on a study of his match with AlphaGo. We’re excited to hear his insights into these amazing games, and to have the chance to share some of AlphaGo’s own analysis too.
[..]
We’re now publishing a special set of 50 AlphaGo vs AlphaGo games, played at full length time controls, which we believe contain many new and interesting ideas and strategies.


https://deepmind.com/blog/alphagos-next-move/

_________________
“The only difference between me and a madman is that I’m not mad.” — Salvador Dali ★ Play a slooooow correspondence game with me on OGS? :)


This post by Bonobo was liked by: wolfking
Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 123 posts ]  Go to page Previous  1 ... 3, 4, 5, 6, 7  Next

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group