Nope. They only will be playing three games. Considering that Alpha Go, playing as Master, has won 60 straight games you're probably right about a sweep.Laskos wrote:Therefore, the most flagrant flaws in AlphaGo can be corrected, and for the next march with Ke Jie I expect a 5-0 result.
World #1 Go Player Ke Jie accepts Google Alpha Go Match..
Moderators: hgm, Rebel, chrisw
-
- Posts: 2851
- Joined: Wed Mar 08, 2006 10:01 pm
- Location: Irvine, CA, USA
Re: World #1 Go Player Ke Jie accepts Google Alpha Go Match.
Deasil is the right way to go.
-
- Posts: 10948
- Joined: Wed Jul 26, 2006 10:21 pm
- Full name: Kai Laskos
Re: World #1 Go Player Ke Jie accepts Google Alpha Go Match.
Tomorrow, 23rd of May in the morning, is the first game. If there will be no commercial AlphaGo, we at least will soon have the new Zen which improved 1000 Elo points since its last release, or 5-6 stones. Should be stronger than AlphaGo of one year ago on personal computer, on top pro level.Dirt wrote:Nope. They only will be playing three games. Considering that Alpha Go, playing as Master, has won 60 straight games you're probably right about a sweep.Laskos wrote:Therefore, the most flagrant flaws in AlphaGo can be corrected, and for the next march with Ke Jie I expect a 5-0 result.
-
- Posts: 213
- Joined: Thu Dec 16, 2010 4:39 pm
Re: World #1 Go Player Ke Jie accepts Google Alpha Go Match.
Really? How the Zen team managed this feat? Any breakthrough in the programming model? I'm assuming they are not using reinforcement learning.
Is their (new?) method transposable to chess? Any literature (research article) or other juicy detail about the origin of this jump?
A lot of questions, LOL
Is their (new?) method transposable to chess? Any literature (research article) or other juicy detail about the origin of this jump?
A lot of questions, LOL
Per ardua ad astra
-
- Posts: 10948
- Joined: Wed Jul 26, 2006 10:21 pm
- Full name: Kai Laskos
Re: World #1 Go Player Ke Jie accepts Google Alpha Go Match.
No, the same Deep Learning DCNN with reinforcement learning as in Alpha Go, they seem to apply Nature paper of DeepMind team literally. Taking into account that the previous Deep Learning Zenith 6 (AFAIK no reinforcement learning, no value network) was already 2-3 stones stronger than non- Deep Learning Zen, the total breakthrough in Zen due to new techniques might be close to 1500 ELO points, or 7-9 stones. Only the last month the unofficial Zen progressed 200 ELO points or 1 stone. By now on a home PC it should be at top pro level and stronger than Alpha Go one year ago. Alpha Go must have progressed too, and it is probably better than Zen if ported for home PC, but I don't know if they will release it for public.melajara wrote:Really? How the Zen team managed this feat? Any breakthrough in the programming model? I'm assuming they are not using reinforcement learning.
Is their (new?) method transposable to chess? Any literature (research article) or other juicy detail about the origin of this jump?
A lot of questions, LOL
If in this match against Ke Jie, Alpha Go will be on that huge cluster, it might be 600-800 ELO points stronger than Ke Jie, and it would be no contest.
-
- Posts: 12038
- Joined: Mon Jul 07, 2008 10:50 pm
-
- Posts: 5960
- Joined: Sun Jan 10, 2010 6:15 am
- Location: Maryland USA
Re: World #1 Go Player Ke Jie accepts Google Alpha Go Match.
Yes, it starts tonite U.S. time. Michael Redmond 9 Dan pro estimates a 10% chance of the pro winning even 1 game. He thought that a 2 stone handicap would be fair, but "politically incorrect".Dirt wrote:The match is now scheduled to take place from May 23 to May 27 of 2017. Only three games are in the main match but there will also be matches between professionals alternating moves with computers as well as other activities.
By a curious coincidence (?) just last week the top Shogi engines defeated the World Champion ("Meijin" to be exact) Sato 2 to 0 in comparably serious games, and they weren't even close; my judgment as an amateur 5 Dan is that the engine could easily give a Lance handicap to any human and expect to win So assuming that the computer wins this Go match, Go will have lasted one week longer than Shogi before having fallen to the computers!
Komodo rules!
-
- Posts: 565
- Joined: Thu Nov 13, 2014 12:03 pm
-
- Posts: 2272
- Joined: Mon Sep 29, 2008 1:50 am
Re: World #1 Go Player Ke Jie accepts Google Alpha Go Match.
Only by 0.5 point if I understood correctly. On the other hand AlphaGo is programmed to go for the surest win and not the win that delivers the most points (which may involve taking higher risks). So the fact that AlphaGo "only" won by 0.5 points may be misleading.
Ideas=science. Simplification=engineering.
Without ideas there is nothing to simplify.
Without ideas there is nothing to simplify.
-
- Posts: 10948
- Joined: Wed Jul 26, 2006 10:21 pm
- Full name: Kai Laskos
Re: World #1 Go Player Ke Jie accepts Google Alpha Go Match.
0.5 without komi of 7.5? That was my impression, but I don't know how to count very well. From what I understood, Ke Jie tried to gain points, so that Alpha would need to attack, but mishandled the opening and in fact after the opening Alpha had more (potential) points and Ke had to attack. There was no opportunity to attack Alpha in the whole game, and Alpha, gaining advantage from the opening, just played quietly for the win. Alpha still seems to go to the highest probability, not highest advantage, so potentially it may had 10-15 stones advantage in late midgame, but went for the safer 6-8 points. I will maybe post Crazy Stone Deep Learning (7 amateusr dan) 2 hour analysis, that shows that from this engine (Zenith 6 too) point of view, White (Alpha) had substantial advantage right from the opening after move 20, a thing which seems to be overlooked or muted by 9p pros commenting. It seems that even relatively weak engines compared to those pros (3-4 stones difference) are better at scoring in the initial stages of the game.Michel wrote:Only by 0.5 point if I understood correctly. On the other hand AlphaGo is programmed to go for the surest win and not the win that delivers the most points (which may involve taking higher risks). So the fact that AlphaGo "only" won by 0.5 points may be misleading.
Ke here seems to have had no any chances, a pretty overwhelming win by Alphago.
-
- Posts: 10948
- Joined: Wed Jul 26, 2006 10:21 pm
- Full name: Kai Laskos
Re: World #1 Go Player Ke Jie accepts Google Alpha Go Match.
Crazy Stone Deep Learning (7 dan) gives the following 2 hour analysis (which seems better in the first 100+ moves than 9p pros):Laskos wrote:0.5 without komi of 7.5? That was my impression, but I don't know how to count very well. From what I understood, Ke Jie tried to gain points, so that Alpha would need to attack, but mishandled the opening and in fact after the opening Alpha had more (potential) points and Ke had to attack. There was no opportunity to attack Alpha in the whole game, and Alpha, gaining advantage from the opening, just played quietly for the win. Alpha still seems to go to the highest probability, not highest advantage, so potentially it may had 10-15 stones advantage in late midgame, but went for the safer 6-8 points. I will maybe post Crazy Stone Deep Learning (7 amateusr dan) 2 hour analysis, that shows that from this engine (Zenith 6 too) point of view, White (Alpha) had substantial advantage right from the opening after move 20, a thing which seems to be overlooked or muted by 9p pros commenting. It seems that even relatively weak engines compared to those pros (3-4 stones difference) are better at scoring in the initial stages of the game.Michel wrote:Only by 0.5 point if I understood correctly. On the other hand AlphaGo is programmed to go for the surest win and not the win that delivers the most points (which may involve taking higher risks). So the fact that AlphaGo "only" won by 0.5 points may be misleading.
Ke here seems to have had no any chances, a pretty overwhelming win by Alphago.
Stones advantage AlphaGo (White):
White could have won by 15+ stones, but large stone difference is not what AlphaGo maximizes (unlike most humans), so it decided to cut it in half for better winning chances.
Winning chances (t-value) for AlphaGo:
This quantity is probably maximized by AlphaGo. AlphaGo took lead right from the opening, steadily increasing it.