Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AlphaGo beats Lee Sedol again in match 2 of 5

gogameguru.com

Post reply on HN
Print view View original

521–530 of 576 posts

Re: AlphaGo beats Lee Sedol again in match 2 of 5

#521

Post by zeven7 »

z0r wrote: post #513
zeven7 wrote: post #507

Earlier quoted context omitted.

It didn't look like a Takemiya-style move to me. Takemiya tends to play for a huge moyo in the center. AlphaGo had no such moyo. It wasn't only a strange move; it was also a strange time to play it, and it definitely went against conventional wisdom.

The result of the shoulder hit coordinated with black's bottom formation, and the extension on the 4th line that threatened to cut white's stones off was flexible and could have easily formed an impressive moyo on the bottom. It did not play out that way, but I think that black's strategy was as cosmic as anything Takemiya might have played. His games did not always end with a giant moyo, he was also very flexible. I…

Thanks for linking to the examples! That is interesting indeed.

Re: AlphaGo beats Lee Sedol again in match 2 of 5

#522

Post by dragontamer »

rwill128 wrote: post #515
dragontamer wrote: post #477

Earlier quoted context omitted.

On the contrary. I think that Chess machines play perfectly for the next 8 moves, but don't necessarily sense the importance of a Knight Outpost (which may have relevance 20 moves ahead. A proper Knight Outpost will remain a fork threat for the rest of the game). It is far easier for a Human to beat a Chess Machine at positional play (ex: a backwards pawn shape will probably be a problem at endgame, 30+ moves from no…

This was true 10-15 years ago. It is no longer true. Chess engines have positional evaluation algorithms that have been trained using many millions of games, and the weighting parameters for different kinds of positional features have been adjusted accordingly. Do some reading on Stockfish for example if you doubt the veracity of my statement.

Yes, I do realize that.

But its just as you say: its weighting parameters and heuristics. When Stockfish recognizes a backwards pawn, it deducts a point value. When Stockfish recognizes "pawn on 6th row", it adds a point value to that pawn.

But that's a heuristic. A trained heuristic using games, but still comes down to what I understand to be a +/- point value (like... +35 centipawns).

In contrast, a chess engine truly knows that if you do X move, it will force a Rook / Minor piece exchange in 8 moves.

When you play positionally vs Stockfish, you're arguing with a heuristic (a heuristic which has been refined over many cycles of machine learning, but a heuristic nonetheless that comes down to "+/- centipawns") . When you play tactically vs Stockfish, it is evaluating positions more than a dozen moves ahead of what is humanly possible.

When you play against Stockfish in endgame tablebase mode, it plays utterly, and provably, perfectly.

Take a pick of what game you want to play against it. IMO, I'd bet on its positional "weakness" (yes, it is still very strong at positional play, but it is the most "heuristical" part of the engine)

Re: AlphaGo beats Lee Sedol again in match 2 of 5

#523

Post by atom-morgan »

semi-extrinsic wrote: post #478
atom-morgan wrote: post #398

Earlier quoted context omitted.

As a competitive speedcuber (Rubik's Cubes) this makes sense. If I watch a fellow cuber solve a cube, I understand their process even if it's a different method than the one I'd use. But a robot solving it? To my brain it looks like random turns until...oh shit it's finished.

Have you ever managed to learn the human Thistlethwaite algo? It basically lets you solve the cube like a robot would. I'm pretty rusty at cubing nw, but I always wanted to learn it.

I have not. It's just not something I'm very interested in.

Re: AlphaGo beats Lee Sedol again in match 2 of 5

#524

Post by Florin_Andrei »

Cookingboy wrote: post #494
fma wrote: post #333

Earlier quoted context omitted.

I'm not a Go player but play other competitive sports. Humans have a herd mentality...as Op mentioned there's certain styles of playing...which has their own strengths and weaknesses. Sometimes people will not examine other styles that may have better strengths and just focus on the exist one. Then comes along someone who 'thinks outside the box' with a new style and revolutionize the playing field. Think Bruce Lee a…

Lee isn't even the best human player at the moment, he has a 2-8 loss record against Ke Jie, who's actually ranked number 1 at the moment. Ke Jie is an arrogant 18 year old and he's been saying on social network in the past couple days how he will defeat AlphaGo.

Exponential progress is going to bear down on Ke Jie like a ton of bricks soon.

Re: AlphaGo beats Lee Sedol again in match 2 of 5

#525

Post by andreyk »

Jach wrote: post #467
astrofinch wrote: post #245

So given that this victory seems to be happening a decade or so before experts predicted, how likely are we to see similar acceleration in reaching other AI milestones? (Especially given that AlphaGo is using the same algorithm that won the Atari games, so it has the potential to be very general in its application)

You probably saw this, but linking anyway: https://news.ycombinator.com/item?id=10983539 The general point is that it's more evidence improvement can come in discontinuous leaps, it doesn't have to be some smooth (even if accelerating) incremental process. So timeline predictions should probably be wide, with closer-to-present lower bounds (especially when successful generalizable techniques become public). I don't t…

Hugely agree with your response. Just to add slightly to that, I think the fact that Facebook had a quite similar Go AI in the works (just without self play reinforcement learning) in the works is an indicator of how clear this research direction was. The complexity of Google's solution (2 neural nets and a fast evaluation function, plus other small details?) really indicates to me a lot of manual engineering and iteration went into this. So it is not really an indicator of cool new theoretical breakthroughs, but an indication that applied engineering to make use of known techniques can achieve great things.

Re: AlphaGo beats Lee Sedol again in match 2 of 5

#526

Post by andreyk »

Very impressive. Since there is a ton of hype about this and many media stories (at least NYTimes, with no citation at all) saying that this came 'a decade early', I think its worth looking over Yann LeCun retrospective on research in this area (https://www.facebook.com/yann.lecun/posts/10153340479982143). Clearly he was saying all this to preface the results of Facebook research in comparison to Google's, but I still think it is a very good overview of the history and shows the ideas did not come about suddenly. Quoting a few key things since the whole things is very long:

"The idea of using ConvNet for Go playing goes back a long time. Back in 1994, Nicol Schraudolph and his collaborators published a paper at NIPS that combined ConvNets with reinforcement learning to play Go. But the techniques weren't as well understood as they are now, and the computers of the time limited the size and complexity of the ConvNet that could be trained. More recently Chris Maddison, a PhD student at the University of Toronto, published a paper with researchers at Google and DeepMind at ICLR 2015 showing that a large ConvNet trained with a database of recorded games could do a pretty good job at predicting moves. The work published at ICML from Amos Storkey's group at University of Edinburgh also shows similar results. Many researchers started to believe that perhaps deep learning and ConvNets could really make an impact on computer Go.

...

Clearly, the quality of the tactics could be improved by combining a ConvNet with the kind of tree search methods that had made the success of the best current Go bots. Over the last 5 years, computer Go made a lot of progress through Monte Carlo Tree Search. MCTS is a kind of “randomized” version of the tree search methods that are used in computer chess programs. MCTS was first proposed by a team of French researchers from INRIA. It was soon picked up by many of the best computer Go teams and quickly became the standard method around which the top Go bots were built. But building an MCTS-based Go bots requires quite a bit of input from expert Go players. That's where deep learning comes in.

...

A good next step is to combine ConvNets and MCTS with reinforcement learning, as pioneered by Nicol Schraudolph's work. The advantage of using reinforcement learning is that the machine can train itself by playing many games against copies of itself. This idea goes back to Gerry Tesauro's “NeuroGammon,” a computer backgammon player that combined neural nets and reinforcement learning that beat the backgammon world champion in the early 1990s. We know that several teams across the world are actively working on such systems. Ours is still in development.

...

This is an exciting time to be working on AI."

Re: AlphaGo beats Lee Sedol again in match 2 of 5

#527

Post by ant6n »

colanderman wrote: post #360
philh wrote: post #306

Earlier quoted context omitted.

> It never actually would make the winning move – however obvious – until it had no other options! I'm confused. Why would 'make the winning move' not be the way to maximise probability of winning?

The AI is based on the minimax algorithm [1]. Because of the way Minimax works, the only way for a possible next move to be designated a "win" is if it is a guaranteed win. (The tree is (effectively) fully explored, and the opponent is given the benefit of the doubt in the face of incomplete information.) So, if there are multiple such winning moves, and care is not taken to distinguish the "magnitude" of the win, th…

But once you have guaranteed winning moves, why not pick the shortest one available (in terms of turns)?

Re: AlphaGo beats Lee Sedol again in match 2 of 5

#528

Post by kowdermeister »

astazangasta wrote: post #377
kowdermeister wrote: post #370

Earlier quoted context omitted.

> We made this, and it is not an intelligence, it is a series of mathematical optimization functions You can say the same about your mind too which is a bunch of optimization nodes. If something is intelligent, does it matter if it's evolved in nature or created by a species who is evolved in nature? > In the future I hope that this research will move more towards models that provide interpretable results I think it'…

It's a bit precious I think to say that a human is a "bunch of optimization nodes". I can write code to create a CNN, and I can draw a graph of how it operates on a piece of paper. We can't even decode a few million rat neurons the same way. A CNN is a pure mathematical function - if you want, you could write it down that way. Given a set of inputs, it will always produce the same output. We don't call a linear regre…

My exaggeration was intentional to point out that if you scale up NN based systems, we are not that different :) I do appreciate it, but let's not forget that we have finite nodes, so at one point a machine can surpass us with "just mathematical functions".

> A CNN is a pure mathematical function That's their basic property, but who are we to say that our cell based neural network is superior? Cells are just compositions of atoms and they are defined by quantum mechanics, which is... "just" math and information.

I also think that Go might be a great communication tool between AI and humans. If you look at the commentary from this angle if's fun to think about like this.

Re: AlphaGo beats Lee Sedol again in match 2 of 5

#529

Post by raattgift »

gohrt wrote: post #488
eva1984 wrote: post #179

Earlier quoted context omitted.

Yeah, can confirm. I am not a Go player, and didn't know much about it. But top Go players are really well respected as some of the highest talents in the society, almost like being feared. It is smart people's game, eventually. And the fact, that, prior to the deep learning revolution, Go is the only board game that human cannot be beaten, add even more myth and charm to the game and players alike. Now, it comes to…

How well would the computer fare if it didn't have access to a library of human-played games, and only got self-study?

How well would the computer fare with a slightly different game -- something like go, but with differing rules? Would a smart human learn faster (in real time, or alternatively with comparable energy use) than an artificial reinforced deep learning system?

And who can make the most interesting new go-like game?

Perhaps this could be tested with chess or checkers, even.

Re: AlphaGo beats Lee Sedol again in match 2 of 5

#530

Post by jjaredsimpson »

gohrt wrote: post #488
eva1984 wrote: post #179

Earlier quoted context omitted.

Yeah, can confirm. I am not a Go player, and didn't know much about it. But top Go players are really well respected as some of the highest talents in the society, almost like being feared. It is smart people's game, eventually. And the fact, that, prior to the deep learning revolution, Go is the only board game that human cannot be beaten, add even more myth and charm to the game and players alike. Now, it comes to…

How well would the computer fare if it didn't have access to a library of human-played games, and only got self-study?

Same argument applies to humans
Post reply on HN