Match 2 15 Minute Summary - Google DeepMind Challenge Match 2016

Google DeepMind · Intermediate ·📰 AI News & Updates ·10y ago

Key Takeaways

The video discusses the Google DeepMind Challenge Match 2016, where AlphaGo, a program developed by DeepMind, competes against Lee Sedol, a legendary Go player, in a 5-game challenge match. The match is summarized by Michael Redmond, a 9-dan professional, and Chris Garlock.

Full Transcript

[Music] hi I'm Chris Garlock and I'm with Michael Redmond nwn professional and uh we are going to do our 15minute highlights of the uh game two in this historic match uh between alphago and Lee sadal uh and exciting game uh really was on the edge of literally actually Michael on the edge of my seat today for about five hours uh great commentary you um and now you have to boil five hours down to 15 minutes so let's get started okay well first of all just to give a general idea of how I felt throughout the game um the game started with um I was really interested to see how alphao would handle black because with black um black has the initiative and uh has sort of the responsibility to make use of that to take control of the game because because of the Comey that black is paying um seven and a half points uh in return for having the first move so black has to take control maybe move the game into a kind of fighting game um usually uh to to make good use of the first move and Alpha go didn't really seem to be doing that in October when it was playing f um so I was looking to see how it would change and I was very impressed because it did and it was even more than I expected so let's see the game he started with staro and I was saying oh start points again but then he played play 34 you were very happy to see this it was a difference because um basically because the three four point um adds variety to the game it it makes it much more complicated so I was wondering if maybe alphago was trying to avoid that at first but now it it's not avoiding it it means that it's willing to play any kind of game basically um now the josei would be to play here or here um actually it's a pretty common uh variation to leave it for to play Tanuki and play once here so this move in itself was not so unusual usually now after this black is going to look at these Stones as dispensable so maybe um if white continues playing moves in that area black might just give them up and so black would play something like this and if white here maybe black would continue with something let's see a chiny style opening and make a moyo on this side of the board while white still needs one more move to finish off all those STS um that would be a kind of a often used opening but instead Alpha go plays once here and then white here uh we would expect some kind of extension but it didn't and played there and after playing this one move to play elsewhere is a bit unusual in fact it's very unusual MH um because it does it um by investing this one move here um it makes it impossible for black to give up that group um whereas if it was just the two stones it would be dispensable after making this exchange and giving white some profit on the left side here um now it's sort of uh it's much more costly for black to throw that group away so if white plays here there's a potential attack on the black Stones uh but apparently alphao wasn't appar wasn't worried about that and actually um didn't do that immediately he chose to play this one which is a more balance oriented move um then Alpha Go played a peep once and that's again an un unusual move but not necessarily bad it does lose some potential to try different moves um but later in the game white will maybe not answer the same way so just making sure that white answers that way it um simplifies the game a little bit so it's not necessarily bad then switch to the corner there uh this is also a standard sequence in itself connect and black cut once this is actually a Joi so it's there's nothing unusual about the play here um but the fact and then he extended the fact that he played the corner here first and then played the extension it's a very good sequence here because uh black has to settle these Stones somehow at some point anyway and this this is an even exchange it's a Joi so it's considered even for both players but the fact that black played this first and and decided the shape here basically before choosing the extension made it easier you might say for black to choose a point to extend like black has a choice between this move or this move or this move which is the original Joi but is sort of um slightly falling out of favor among professionals professional because of the weakness here um but when wet black was had not played all of this black still sort of wanted to play all the way out once this shape is fixed here it's much more reasonable for black to be pulling back like this so it uh make it adds more validity to this this choice that Alpha go made here okay by playing this first so it's a very nice sequence in all uh the game is probably it looks sort of even at this point but alphago is definitely um taking the initiative and deciding what kind of game it's going to be so with white I would already be feeling a bit pressured and I would know that at least it's at the best it's an even game but maybe actually um the the fact that black is sort of taking an initiative here maybe Black's ahead already and I think that's what uh liso said in the press conference was that he felt PR he felt that he was completely close out from the beginning of the game right um and it was the fact that black is putting pressure on white um already from the beginning of the game um and it and not making any mistakes and then um after that uh black we see I I hope to show you that black actually made all of that work and kept his uh lead uh White can's here usually you would expect black to play here in this white was sort of hoping to jump into the three three point which is very popular nowadays it's a very popular Jo that starts with that this sort sort of makes that more difficult for white um and basically because the side here because of the strong white group here the side here is not so big so Black's forcing white to play on that site which is a good strategy and now comes the bombshell right yeah that that was a move that no one I thought it was a mistake when it was played I thought I miss well actually on the server where I was watching to make sure I had the moves right uh the person who um was putting inputting the moves got it wrong he he he he was playing here because which is what what he was expecting right and that person had to to take back that move and I've been that guy so I I you know yeah so um I was looking at the server and it was showing a different move but then I looked at the game room on the TV and oh yeah so this is the right one this this is the move of the game and you know the people doing those servers uh inputting the moves are pretty strong players yeah yeah yeah and then uh black played here and actually I think white should have just played like this yes um and black is playing a lot of innovative exciting moves but it's hard to say if black has gained anything really yet so I think the game would be about even at this point right but when white pwls and black connects here we can see that all of this stuff that black is doing is coming together to and this was the move that black wanted to play to start a fight here and all of these stones that were um traditionally would be called questionable um or actually working in Black's favor when this fight comes comes up so um it's all coming together beautiful SE so that's why I would sort of question this yeah white push black plami covered on the top and this sequence is a kind of a tesuji sequence and it's pretty much forced for both sides and white white is alive there on the side now more or less um black has reduced White's right side and then this move that's a nice light move it's not a move that people would not uh think of it's a nice looking move um and black does have sort of Escape roots for this group here which is weak but it um he has a move somewhere around here which would sort of connect up and and with this Stone here he he has an escape so toward this side so it's not as if black is all that weak so he goes ahead and plays this one um playing this move before this is is also good order of moves um because after this it would not have been forcing and then he this is the weakest point in Black's territory so everything that blacka is doing now is it makes sense and what's more it it was looking adventurous and now it makes sense so that's really brilliant I think yeah when when they played these moves here and and then with the connection just everything as you say comes together uh beautifully and you really had a a lovely feel to the game yeah so now white plays here now white is uh sort of planning to um make an attack on these stones to start that white plays here this sort of uh Cuts this group off from the left side here so uh black has to move out in this direction to s be safe but also white is threatening this one which is actually more Troublesome for black because it creates a weak group here and a weak group here and if this weak group is sort of uh um struggling in the center that means that that would have a bad effect on this half of the board and of course these Stones also so black answered by playing here um and at this point uh he said he play this move which was a very very big territorial move but I think this might have been a problem um because if I was white um I would play somewhere around here or here to make a more direct attack against the black Stones here and I would be let's put it here for inance and say black runs away somehow U black might find a better move than that actually but um just just for instance okay um with a stone in this vicinity it will be very easy for white to jump in here right and for the time being I think this kind of stuff would be bigger although this is a huge move the the value of this move is that makes this a living group but it's not really that weak to start with but it adds security also um after black plays here let's just show that um white can later jump into the corner here so there's the added value there so it's a it's a pretty big move but I think this was actually a chance for white to take the initiative in the center of the board um which would have effects on all sorts of places on the board this is how I would play you know U traditionally it's probably bad form for me to be questioning lisad all's move but that's how I feel in this position and so after black play here now if this this Conn next up to sort of to the moo here now this is going to be a moo the side territory will be well except for the corner here it will be more safe then it will be spreading into the center which would be bad for white so white starts the attack here he started here and here and there but U by instead of connecting the Domin points here black is dodging to the side um and uh hoping to connect up this direction so this is good for black too and and I think then white uh played once here and then invaded um and here now this kind of indirect fight attack against this Zone a direct attack black has to kill it but with an indirect attack black is gaining on the sides so this is a very PR likee move also white played here and jumped and you can see black is sort of playing thickly um a very safe way and also it's an effective attack on white let's just get a few more mov and then this one so I think we can sort of do a because we have to wrap up um I can't really find any mistakes in Black's play and I'm finding moves that I I wouldn't even think of here and they're really brilliant um because they actually he actually put them together to make them work and to sort of um focus on this attack that black made on the right side um um I did I did question White's move here for one thing that allowed that to happen um but that's sort of leadle style I guess but it so I I think we just have to admire what alphago did um more than uh question Lee s's moves right I mean it was it was a brilliant game and and I felt like uh lee sadal fought incredibly strongly you know right through to the very end um and it was just you know an amazing amazing game uh and as you say there was several just sort of brilliant moves uh that we saw coming from alphago this time almost no human Pro would have thought of I think right I think everyone was surprised by this shoulder hit so we'll be looking ahead to game three which will be uh important to see what alphago will come up with next but also uh Lisa do is really under the gun he's 0 and2 now and he's got to win he's got to win uh you know every game from here on out so I don't know about you but uh it's it's going to be tough going it's going to be tough going thank you so much Michael Redmond n down professional I'm Chris Garlock managing editor of the American go ejournal thank you so much for watching and we'll see [Music] you

Original Description

15 minute Summary by Michael Redmond 9 dan professional and Chris Garlock on today's historic second match between DeepMind's program AlphaGo take on the legendary Lee Sedol (9-dan pro), the top Go player of the past decade, in a $1M 5-game challenge match in Seoul. In October 2015, AlphaGo became the first computer program ever to beat a professional Go player by winning 5-0 against the reigning 3-times European Champion Fan Hui (2-dan pro). That work was featured in a front cover article in the science journal Nature in January 2016.
Sign in to unlock AI tutor explanation · ⚡30

Playlist

Uploads from Google DeepMind · Google DeepMind · 15 of 60

1 RL Course by David Silver - Lecture 8: Integrating Learning and Planning
RL Course by David Silver - Lecture 8: Integrating Learning and Planning
Google DeepMind
2 RL Course by David Silver - Lecture 1: Introduction to Reinforcement Learning
RL Course by David Silver - Lecture 1: Introduction to Reinforcement Learning
Google DeepMind
3 RL Course by David Silver - Lecture 2: Markov Decision Process
RL Course by David Silver - Lecture 2: Markov Decision Process
Google DeepMind
4 RL Course by David Silver - Lecture 5: Model Free Control
RL Course by David Silver - Lecture 5: Model Free Control
Google DeepMind
5 RL Course by David Silver - Lecture 6: Value Function Approximation
RL Course by David Silver - Lecture 6: Value Function Approximation
Google DeepMind
6 RL Course by David Silver - Lecture 4: Model-Free Prediction
RL Course by David Silver - Lecture 4: Model-Free Prediction
Google DeepMind
7 RL Course by David Silver - Lecture 3: Planning by Dynamic Programming
RL Course by David Silver - Lecture 3: Planning by Dynamic Programming
Google DeepMind
8 RL Course by David Silver - Lecture 10: Classic Games
RL Course by David Silver - Lecture 10: Classic Games
Google DeepMind
9 RL Course by David Silver - Lecture 7: Policy Gradient Methods
RL Course by David Silver - Lecture 7: Policy Gradient Methods
Google DeepMind
10 Google DeepMind: Ground-breaking AlphaGo masters the game of Go
Google DeepMind: Ground-breaking AlphaGo masters the game of Go
Google DeepMind
11 Match 1 - Google DeepMind Challenge Match: Lee Sedol vs AlphaGo
Match 1 - Google DeepMind Challenge Match: Lee Sedol vs AlphaGo
Google DeepMind
12 Match 2 - Google DeepMind Challenge Match: Lee Sedol vs AlphaGo
Match 2 - Google DeepMind Challenge Match: Lee Sedol vs AlphaGo
Google DeepMind
13 Match 1 15 min Summary - Google DeepMind Challenge Match
Match 1 15 min Summary - Google DeepMind Challenge Match
Google DeepMind
14 Match 3 - Google DeepMind Challenge Match: Lee Sedol vs AlphaGo
Match 3 - Google DeepMind Challenge Match: Lee Sedol vs AlphaGo
Google DeepMind
Match 2 15 Minute Summary - Google DeepMind Challenge Match 2016
Match 2 15 Minute Summary - Google DeepMind Challenge Match 2016
Google DeepMind
16 Match 3 15 Minute Summary - Google DeepMind Challenge Match 2016
Match 3 15 Minute Summary - Google DeepMind Challenge Match 2016
Google DeepMind
17 Match 4 - Google DeepMind Challenge Match: Lee Sedol vs AlphaGo
Match 4 - Google DeepMind Challenge Match: Lee Sedol vs AlphaGo
Google DeepMind
18 Match 4 15 Minute Summary - Google DeepMind Challenge Match 2016
Match 4 15 Minute Summary - Google DeepMind Challenge Match 2016
Google DeepMind
19 Match 5 - Google DeepMind Challenge Match: Lee Sedol vs AlphaGo
Match 5 - Google DeepMind Challenge Match: Lee Sedol vs AlphaGo
Google DeepMind
20 Match 5 15 Minute Summary - Google DeepMind Challenge Match 2016
Match 5 15 Minute Summary - Google DeepMind Challenge Match 2016
Google DeepMind
21 DQN SPACE INVADERS
DQN SPACE INVADERS
Google DeepMind
22 DQN Breakout
DQN Breakout
Google DeepMind
23 Asynchronous Methods for Deep Reinforcement Learning: Labyrinth
Asynchronous Methods for Deep Reinforcement Learning: Labyrinth
Google DeepMind
24 Asynchronous Methods for Deep Reinforcement Learning: MuJoCo
Asynchronous Methods for Deep Reinforcement Learning: MuJoCo
Google DeepMind
25 Asynchronous Methods for Deep Reinforcement Learning: TORCS
Asynchronous Methods for Deep Reinforcement Learning: TORCS
Google DeepMind
26 Differentiable neural computer family tree inference task
Differentiable neural computer family tree inference task
Google DeepMind
27 StarCraft II DeepMind feature layer API
StarCraft II DeepMind feature layer API
Google DeepMind
28 DeepMind Health – Partnership with the Royal Free London NHS Foundation Trust
DeepMind Health – Partnership with the Royal Free London NHS Foundation Trust
Google DeepMind
29 DeepMind Health – Michael Wise – a patient's journey
DeepMind Health – Michael Wise – a patient's journey
Google DeepMind
30 Streams – a platform for a digital NHS
Streams – a platform for a digital NHS
Google DeepMind
31 DeepMind Lab - Nav Maze Level 1
DeepMind Lab - Nav Maze Level 1
Google DeepMind
32 DeepMind Lab - Stairway to Melon Level
DeepMind Lab - Stairway to Melon Level
Google DeepMind
33 DeepMind Lab - Laser Tag Space Bounce Level (Hard)
DeepMind Lab - Laser Tag Space Bounce Level (Hard)
Google DeepMind
34 Exploring the mysteries of Go with AlphaGo and China's top players
Exploring the mysteries of Go with AlphaGo and China's top players
Google DeepMind
35 Demis Hassabis on AlphaGo: its legacy and the 'Future of Go Summit' in Wuzhen, China
Demis Hassabis on AlphaGo: its legacy and the 'Future of Go Summit' in Wuzhen, China
Google DeepMind
36 The Future of Go Summit: AlphaGo & Ke Jie match 1 moves analysis
The Future of Go Summit: AlphaGo & Ke Jie match 1 moves analysis
Google DeepMind
37 The Future of Go Summit: AlphaGo & Ke Jie match 2 moves analysis
The Future of Go Summit: AlphaGo & Ke Jie match 2 moves analysis
Google DeepMind
38 The Future of Go Summit: Pair Go moves analysis
The Future of Go Summit: Pair Go moves analysis
Google DeepMind
39 The Future of Go Summit: AlphaGo & Ke Jie match 3 moves analysis
The Future of Go Summit: AlphaGo & Ke Jie match 3 moves analysis
Google DeepMind
40 Emergence of Locomotion Behaviours in Rich Environments
Emergence of Locomotion Behaviours in Rich Environments
Google DeepMind
41 StarCraft II 'mini games' for AI research
StarCraft II 'mini games' for AI research
Google DeepMind
42 Trained and untrained agents play StarCraft II full 1vs1 game
Trained and untrained agents play StarCraft II full 1vs1 game
Google DeepMind
43 DeepMind open source PySC2 toolset for Starcraft II
DeepMind open source PySC2 toolset for Starcraft II
Google DeepMind
44 ICML 2017: Test of Time Award (Sylvain Gelly & David Silver)
ICML 2017: Test of Time Award (Sylvain Gelly & David Silver)
Google DeepMind
45 Ke Jie and DeepMind's Go Ambassador Fan Hui review the 3rd AlphaGo vs Ke Jie game
Ke Jie and DeepMind's Go Ambassador Fan Hui review the 3rd AlphaGo vs Ke Jie game
Google DeepMind
46 Ke Jie and DeepMind's Go Ambassador Fan Hui review the 1st AlphaGo vs Ke Jie game
Ke Jie and DeepMind's Go Ambassador Fan Hui review the 1st AlphaGo vs Ke Jie game
Google DeepMind
47 Ke Jie and DeepMind's Go Ambassador Fan Hui review the 2nd AlphaGo vs Ke Jie game
Ke Jie and DeepMind's Go Ambassador Fan Hui review the 2nd AlphaGo vs Ke Jie game
Google DeepMind
48 AlphaGo Zero: Discovering new knowledge
AlphaGo Zero: Discovering new knowledge
Google DeepMind
49 AlphaGo Zero: Starting from scratch
AlphaGo Zero: Starting from scratch
Google DeepMind
50 Defining principles for tech companies in the NHS: DeepMind Health's Collaborative Listening Summit
Defining principles for tech companies in the NHS: DeepMind Health's Collaborative Listening Summit
Google DeepMind
51 A systems neuroscience approach to building AGI - Demis Hassabis, Singularity Summit 2010
A systems neuroscience approach to building AGI - Demis Hassabis, Singularity Summit 2010
Google DeepMind
52 Retour de Rémi Munos en France et ouverture de DeepMind Paris
Retour de Rémi Munos en France et ouverture de DeepMind Paris
Google DeepMind
53 Grid cells - Caswell Barry, UCL
Grid cells - Caswell Barry, UCL
Google DeepMind
54 DeepMind Health Research and Moorfields Eye Hospital NHS Foundation Trust: What our research shows
DeepMind Health Research and Moorfields Eye Hospital NHS Foundation Trust: What our research shows
Google DeepMind
55 DeepMind Health Research and Moorfields Eye Hospital NHS Foundation Trust: A Patient's Story
DeepMind Health Research and Moorfields Eye Hospital NHS Foundation Trust: A Patient's Story
Google DeepMind
56 Deep Learning 3: Neural Networks Foundations
Deep Learning 3: Neural Networks Foundations
Google DeepMind
57 Deep Learning 5: Optimization for Machine Learning
Deep Learning 5: Optimization for Machine Learning
Google DeepMind
58 Deep Learning 8: Unsupervised learning and generative models
Deep Learning 8: Unsupervised learning and generative models
Google DeepMind
59 Reinforcement Learning 1: Introduction to Reinforcement Learning
Reinforcement Learning 1: Introduction to Reinforcement Learning
Google DeepMind
60 Deep Learning 2: Introduction to TensorFlow
Deep Learning 2: Introduction to TensorFlow
Google DeepMind

The video provides a 15-minute summary of the historic second match between AlphaGo and Lee Sedol, discussing the game's progression and key moves. Viewers can learn about the application of AI in games and the current state of AI research. The video is a great resource for those interested in AI and game theory.

Key Takeaways
  1. Watch the video to understand the game's progression
  2. Analyze the key moves made by AlphaGo and Lee Sedol
  3. Research the development of AlphaGo and its applications
  4. Explore the current state of AI research in game theory
💡 The match between AlphaGo and Lee Sedol demonstrates the capabilities of AI in complex games, highlighting the potential for AI to surpass human intelligence in specific domains.

Related Reads

📰
Pakistan’s Digital Summer: Two Inaugurations, One Real Question
Pakistan launches its first AI-ready data centre and Google's local office, sparking questions about the country's digital future
Medium · AI
📰
The study NVIDIA doesn’t want you to see.
NVIDIA's dominance in AI hardware is being questioned in a mysterious study, sparking curiosity about the industry's massive spending
Medium · AI
📰
What is a Free Relationship with Technology?
Learn how to cultivate a free relationship with technology by incorporating philosophy and art in the age of AI
Medium · AI
📰
The House Doesn’t Publish Its Tells via @sejournal, @pedrodias
AI labs are using pre-2022 books to avoid AI-generated content and developing text watermarking, indicating content-at-scale strategies may not be effective
Search Engine Journal
Up next
Since February, Reddit has lost almost half of its keyword rankings.
Edward Sturm
Watch →