Google's AlphaGo Beats Lee Se-dol In the First Match (theverge.com)

New submitter Fref writes with news from The Verge that "A huge milestone has just been reached in the field of artificial intelligence: AlphaGo, the program developed by Google's DeepMind unit, has defeated legendary Go player Lee Se-dol in the first of five historic matches being held in Seoul, South Korea. Lee resigned after about three and a half hours, with 28 minutes and 28 seconds remaining on his clock. "
Lee will face off against AlphaGo again tomorrow and on Saturday, Sunday, and Tuesday. Also at the New York Times. Science magazine says the loss may be less significant than it seems at first.

Comments

Lee underestimated the computer

[rmdingler]

This is a great accomplishment for A.I., but it's likely he will rebound from this opening round loss.

Re:Lee underestimated the computer

[infolation]

And it will be no match for him at kickboxing.

Re:Lee underestimated the computer

[AmiMoJo]

It might he harder for Lee to beat a computer because he says that he relies heavily on reading his opponent. Unlike poker you can't just calculate odds on everything, and unlike chess there are too many permutations to plan right to the end of a game.

It really depends if he can find a way to figure the computer out without the usual cues he gets from human players.

Re:Lee underestimated the computer

[BigFire]

AlphaGo did beat the European champion 5 out of 5. So Lee will have to step up his game a bit.

Re:Lee underestimated the computer

[FatdogHaiku]

Poker is called a "game of imperfect information" and just calculating the odds will not make you a winner. Like Go, interaction with the players is important in that you may be getting false or misleading information. This 2015 heads up match was an interesting contest:
https://en.wikipedia.org/wiki/Claudico

Re:Lee underestimated the computer

[BitwiseX]

And it will be no match for him at kickboxing.

I've read that somewhere before..

Re:Lee underestimated the computer

[arth1]

AlphaGo did beat the European champion 5 out of 5. So Lee will have to step up his game a bit.

The European master was 2 Dan. Lee is 8 Dan (or 9, but the last one is honorary, so it doesn't count for comparing strengths). In the world of Go, that difference is rather staggering. It's like the difference between a chess ELO rating of 2100 and 2600. Someone consistently beating the lower ranked player may not have a chance against the higher rated player.

If Lee were to play the European master, he'd be expected to trounce him too. Quite thoroughly.

What counts in AlphaGo's favor here is that it has had over a year to improve. That it did win the first game says a lot more about its strength than any play against much lower ranked players.

Re:Lee underestimated the computer

[Anonymous Coward]

In fact let me spell it out for people who've never paid any real attention to Poker.

At any moment, a single human player of (say) Texas Hold 'Em can see some of the state of the game, but not all of it, and, except for the final round of betting (the "river" round) there is still a random element which in most cases can be decisive.

It might seem as though a perfect player would calculate the odds that they've got the best hand, and bet accordingly. But actually that's awful because now the other players can determine from how you bet exactly what cards you've got. Instead then, a good player must "balance" their behaviour so that whatever they do their opponent doesn't learn anything valuable without paying for it. Some high end professional players have balanced play where they'll occasionally bet very strong with total air (ie they know they don't have the best hand) so that even when you suspect they have great cards you can't be sure. Seeing just one hand of this looks completely insane - but they make money every year, because they don't play just one hand, they play thousands of hands and over time this unpredictability makes them hard to beat.

Re:Lee underestimated the computer

[TylerJWhit]

Unfortunately, the reality is, an efficient AI can and will learn their opponents behavior, and couple that with statistical probabilities and the opponent is essentially screwed.

Re:Lee underestimated the computer

[Anonymous Coward]

They're professional ranks, not amateur ranks, so it's more like ELO 2700 vs 2900 or so. Still a significant difference, but not quite as staggering.

Re:Lee underestimated the computer

[Anonymous Coward]

It's not just that it won a game. It won the game AS WHITE! I haven't been seeing this pointed out elsewhere...

The other human professionals who read out this game to score it said that the win was beyond the Komi (the 7.5 points granted to white due its not going first). IOW, AlphaGo beat TWO levels of disadvantage; playing as white and achieving more than the balance granted by komi.

I think Lee Sedol will have a much more difficult time in the next game as AlphaGo will have black (first move). And while komi is supposed to make up for the first move advantage, it doesn't take away the feeling from the _human_ playing white that they play the entire game attempting to come from behind.

Re:Lee underestimated the computer

[Bender0x7D1]

I've read that somewhere before..

It was shown in the documentary Kung Fury. The best robots were unable to master the intricate fighting moves.

Re:Lee underestimated the computer

[arth1]

I think Lee Sedol will have a much more difficult time in the next game as AlphaGo will have black (first move). And while komi is supposed to make up for the first move advantage, it doesn't take away the feeling from the _human_ playing white that they play the entire game attempting to come from behind.

True, but Lee is also known for playing an exceptionally strong white (much like Viktor Kortchnoi excelled at black in chess).

The biggest damage might be psychological here - he went to the game thinking he would win the fight 5-0 or 4-1, but now have to rethink that - while 4-1 is still possible, the odds are not on his side.

Re:Lee underestimated the computer

[slashping]

You don't need intricate fighting moves when your fists are made from steel.

Re:Lee underestimated the computer

[Schmorgluck]

Poker is a game of incomplete information. Go is a game of complete information. Making an A.I. good at playing go is interesting because of the game's complexity and vast universe of permutation, but making an A.I. good at playing poker is probably more interesting in terms of decision-assistance (for example). Making an A.I. good at both would be quite the breakthrough.

Re:Lee underestimated the computer

[Anonymous Coward]

The reason it's not being pointed out elsewhere is because you seem to have been misinformed.

Not sure which human professionals you listened to at which point in the game, but at the point when Lee gave up, he would have lost by about 3 points, taking into account the Komi.

In a game where games are often and won/lost by 0.5 points, 3 points is of course huge, and thus Lee gave up.

I watched the whole game live, and I have reasons to be confident that Lee will win the remaining 4 games. AlphaGo made some ridiculous moves that a pro gamer would *never* make. You maybe thinking "well, and Lee STILL lost?". But it was very apparent from Lee's game play that he was *testing* AlphaGo throughout the game. When the computer made a series of dumb moves at mid-late stage of the game, Lee got a little too comfortable and perhaps a bit cheeky, going after AlphaGo in the lower right hand corner, when he should have simply grabbed more *houses* for himself there. That's ultimately why Lee lost (at least IMHO). If Lee played the whole game *normally*, I expect he would have won comfortably.

Re:Lee underestimated the computer

[K. S. Kyosuke]

I watched the whole game live, and I have reasons to be confident that Lee will win the remaining 4 games.

So, how confident are you now?

Re:Lee underestimated the computer

[Anonymous Coward]

Nice prediction, bozo!

Re:Lee underestimated the computer

[Anonymous Coward]

Tell that to the pros who make a living playing online.

Someone has to say it

[Anonymous Coward]

I for one welcome our new computer overlords

Big Whoop

[Anonymous Coward]

My AI has never lost a game of naughts and crosses. It's not hard to play a game.

Wake me up when this AI does something interesting...

Re:Big Whoop

[Lisandro]

It's not hard to play a game.

Well, that's up for debate. Go is arguably the hardest game to play (and master) there is.

Re:Big Whoop

[Anonymous Coward]

Hi, it's called "tic-tac-toe".

Thanks,
America

Re:Big Whoop

[Anonymous Coward]

"Go is arguably the hardest game to play (and master) there is."

Hardly. Try Diplomacy some time. Complex negotiation and justifying back-stabbing,

If the computer disdains or is incompetent at unstructured negotiation with other players,
let's see how long it will last with the players ganged up against it.

Re:Big Whoop

[Anonymous Coward]

No, Go is still more difficult.

In diplomacy there are fewer moves to plan for, and the irrational behavior of the actors can also be planned for. In any single situation there may only be 1-10 outcomes and these can be weighted and a most likely outcome chosen. Over the course of a year, if wishing to implement a 12month strategy, there may only be 1-100 such situations, so even assuming worst case scenarios, the decision tree really isn't that large.

The difficulty, of course, is assigning the weightings, but humans become pretty good at assessing other people's behavior, and any capable group (I think a team here is important) of political specialists could handle this quite well.

Anyway, all I wanted to write is that political manoeuvrings aren't actually that complex, certainly not compared to a provably complex game like Go.

Re:Big Whoop

[GTRacer]

Clearly *you* my friend have never played Bamboozled! https://www.youtube.com/watch?...

Re:Big Whoop

[Tom239]

Well, that's up for debate. Go is arguably the hardest game to play (and master) there is.

Hex (a.k.a. Con-tac-tix or Nash) is a very subtle and interesting game. Programs still can't beat the best human Hex players. DeepMind's CEO was quoted in the NYTimes today saying, "Really, the only game left after chess is Go". I wish reporters knew to ask him, "What about Hex?"

Re:Big Whoop

[Anonymous Coward]

I respectfully disagree.

Diplomacy with it's self-references and multi-body perturbations is closer to a set of non-linear partial differential equations whose solution is extremely difficult and idiosyncratic and chaotic.

Ever had a disagreement with someone who is basing their behavior upon yours, who in turn is basing their behavior upon theirs? Now add as many as five more players, all intertwining their interactions with yours over time. We are not talking here about enumerating numerical solutions to nice set of equations, either. We are talking about inputs including revenge, boredom, capriciousness, contrariness.

Basically Chess and Go are nerfed versions of real world problems that humans have learned to deal with pretty well.
Even humans have significant problems with these cicular and self-referential domains, see R.D. Laing's book Knots, for example.

Re:Big Whoop

[dtmos]

Go is arguably the hardest game to play (and master) there is.

No. Try human copulation.

Re:Big Whoop

[Anonymous Coward]

I don't know about that, wikipedia, the relevant source of everything in the world is quite categorical:

The game can never end in a tie, a fact proved by John Nash: the only way a player can prevent an opponent from forming a connecting path is to form their own path. In other words, Hex is a "determined" game.

When the sides of the grid are equal, the game favors the first player. A standard non-constructive strategy-stealing argument proves that the first player has a winning strategy as follows:

Since Hex is a finite, perfect information game that cannot end in a tie, either the first or second player must possess a winning strategy. Note that an extra move for either player in any position can only improve that player's position. Therefore, if the second player has a winning strategy, the first player could "steal" it by making an irrelevant move, and then follow the second player's strategy. If the strategy ever called for moving on the square already chosen, the first player can then make another arbitrary move. This ensures a first player win.

I found other sources much less categorical (the like of: "first player has demonstrably an advantage, but the winning strategy cannot be computed"), so feel free to fix the article (with sources).

Re:Big Whoop

[Lisandro]

Wha'? That's as simple as a baby toy shapes. Which only makes it more fun.

Re:Big Whoop

[Anonymous Coward]

All that and you don't know that it's means it is?

Re:Big Whoop

[Anonymous Coward]

BURN

Re:Big Whoop

[amorsen]

I found other sources much less categorical (the like of: "first player has demonstrably an advantage, but the winning strategy cannot be computed"), so feel free to fix the article (with sources).

Your quote from Wikipedia and the one in quotes are in complete agreement. It would be difficult to fix it.

Re: Big Whoop

[Anonymous Coward]

There's nothing to fix. You're just an idiot.

Re:Big Whoop

[Anonymous Coward]

Go is a limited, finite game, ideal for a computer to shine. I am still waiting for a computer who can recognize my bags at a conveyor belt at least as efficiently as me, with whom I can have a conversation just as I do with people, that gives useful hints to queries (unlike the idiotic feedback from the Siri, Google Now and friends) etc. It has long been the case that computers are far better than we are at very specific, narrow disciplines; take them out of there, and they are pathetic. The only thing surprising about the Go event is that it did not happen like ten, or even twenty, years ago. You may be impressed, but I find this most underwhelming.

Re:Big Whoop

[Lisandro]

I am still waiting for a computer who can recognize my bags at a conveyor belt at least as efficiently as me

I've worked with industrial vision devices in the past and trust me, you could set up a machine to recognize luggage as efficiently as a human being today if you wanted to. In fact, it will do better.

The only thing surprising about the Go event is that it did not happen like ten, or even twenty, years ago. You may be impressed, but I find this most underwhelming.

That's likely because you don't understand what it involves. Go is unlike chess in the sense that just throwing raw computing power at the problem won't help you at all; for a "small" 13x13 there are over 10^300 valid game trees to compute, and the number gets exponentially worse once the board increases in size. For reference, the estimated number of atoms in the universe is 10^130.

Google's AlphaGo engine is an actual machine-learning AI which had to be trained plays the game much like a regular person would - Myungwan Kim actually remarked that it feels like playing against a human being. Having a competitive Go engine today is a major milestone, make no mistake about it.

Re:Big Whoop

[Anonymous Coward]

I wish reporters knew to ask him, "What about Hex?"

You might as well wish for a pony, too. The media doesn't care about doing any investigation. Look at all their recent Narratives: programming can be learned in an hour of code; it's sexually frustrated misogynerds who can't get laid that are preventing women from learning programming; Sanders has already lost the primary (sadly, though, Sanders probably won't win the primary in the end); Trump can't ever win the primary; Clinton was a civil rights activist and is a better choice than Sanders for people concerned about civil rights; there is a programmer shortage; feminism supports trans women; etc.

Re:Big Whoop

[JMZero]

In game design theory, the kinds of politics you describe is normally treated as a form of luck - and as such it makes determining "who is better at this game" a meaningless question. If 3 random people are playing Risk against the best Risk player in the world (the person who understands the game the best), there's a rational argument that their best strategy is to co-operate and eliminate him first (no matter what he does or says or how he behaves). This sort of interaction effectively decouples skill from game success.

This is also why modern game design has generally abandoned games with lots of politics - they effectively all become the same game, and that game is really uninteresting after a while.

Re:Big Whoop

[JMZero]

I see no reason to believe this game wouldn't fall to serious effort (ie. computers would surpass human players if a good team made a large effort as has been made here with Go). Rather, I think there's many reasons to believe it would be much easier to reach that point. The game state is much clearer and much more amenable to search than Go is, and the game doesn't have the kinds of complicating factors that would make me think of it as a "hard game" for computers to play (eg. hidden information, simultaneous action selection, broad ranges of choices, deeply "non-local" interactions). The game has interesting properties, but none of those seem relevant to the task at hand, other than that they suggest the game probably isn't a tough one for computers to play well (eg. the fact that 9x9 boards have an explicit winning strategy).

Re:Big Whoop

[Tom239]

Yes, a strategy for perfect Hex play is difficult to find (known to be a PSPACE-complete problem). Hex (on large-sized boards like 19x19) is like Go in that we approach it with heuristics because finding a perfect solution is intractable. In practice, both Go and Hex have rules to put both players on closer to an even footing (komi in Go, the "pie rule" (a.k.a. swap) in Hex).

Re:Big Whoop

[Joce640k]

It's not hard to play a game.

Well, that's up for debate. Go is arguably the hardest game to play (and master) there is.

It still follows very fixed rules.

Lee's defeat at Go doesn't demonstrate machine "intelligence" any more than Kasparov's defeat at chess did. It just shows better algorithms and advances in computer processing power.

Re:Big Whoop

[Joce640k]

The only thing surprising about the Go event is that it did not happen like ten, or even twenty, years ago. You may be impressed, but I find this most underwhelming.

That's likely because you don't understand what it involves. Go is unlike chess in the sense that just throwing raw computing power at the problem won't help you at all; for a "small" 13x13 there are over 10^300 valid game trees to compute,

....except that no program ever computes the full game tree. It would be impossible to do in many games, eg. Chess.

The trick is to prune the tree. This requires skill by human programmers.

(What this news is really about is that some humans have managed to produce a workable pruning algorithm for Go, it really has very little to do with AI).

Re:Big Whoop

[Lisandro]

It requires more than skill. Pruning such a massive game tree is no minor feat - in fact, we don't know how to do it even today. All Go engines are based on some form of adaptive AI.

Again, chess is waaaaay easier in comparison. Pretty much all chess engines work the same way: they start with precomputed moves from an opening book and then move to what's an essentially brute force approach where the engine tries positions, assigns them scores and then picks the highest score available. How these positions are scored / discarded is what separates them, but the base procedure is unchanged. This is also what leads to what chess players call "computer moves" - most chess engines will favor unassuming, conservative moves yielding small positional advantages instead of, well, more "human", intelligent ones. Picking up pivotal moments from classic games (move 17 on Fischer-Byrne, for example) and feeding them to top-rated chess engines is an enlightening exercise.

This is all but impossible to perform in Go with even modest board sizes. The game complexity, given its simple rules, is just staggering.

Re:Big Whoop

[Lisandro]

Go is a far better demonstration of "intelligence" than chess in the sense that you require some form of actual AI to be competitive in it. The opening book+brute force combo used by modern chess engines is useless here.

AlphaGo relies heavily on machine learning neural network to play.

Re:Big Whoop

[Joce640k]

No argument that chess is simpler but at the end of the day the process is the same. A bunch of humans did trial and error with lots of heuristics and ran machines against each other all day long to find out which ones worked best. When it gets complex you can vary the scores for each heuristic randomly and let the machines fight it out while you sleep. End of story.

The big advantage of machines compared to humans is that they're methodical and don't overlook stuff. They don't get tired, they don't have bad days.

Re:Big Whoop

[Joce640k]

Neural networks are just a fancy form of heuristic.

At the end of the day the underlying main loop in the program will be very similar to chess (or reversi, or tic-tac-toe,,,,).

a) It generates moves
b) It gives the new board position a score
c) It plays the move with the highest score

Part (b) is the tricky bit but at the end of the day it's just a case adding up values output by a set of heuristics applied to the board. The machine isn't showing any "intelligence" at all. All the intelligence comes from the people who decided on what heuristics to use.

Re:Big Whoop

[RockDoctor]

....except that no program ever computes the full game tree. It would be impossible to do in many games, eg. Chess.

True, but not particularly helpful.

The trick is to prune the tree. This requires skill by human programmers.

And a necessary prerequisite before pruning the game tree is an efficient algorithm for comparing the likely score for one board position compared to another well before you get to the end of the game. While the AlphaGo team have made progress in this aspect - as well as in other aspects of the game, this remains one of the more difficult aspects of computer Go.

RIP

[Anonymous Coward]

Rest in Peace Slashdot....

I have followed you (and had even registered nickname with five digits) for years, only to drop out few years ago once you started these stupid website redesigns for survival.

There are no more commenters, as I remember when often it was three digits if not even four, and now it is often just two digits if even barely that.

5-10 years ago the news like this would have triggered 1000-1500 comments, but now just few dozens.

Re:RIP

[fisted]

5-10 years ago the news like this would have triggered 1000-1500 comments, but now just few dozens.

Within the first 25 minutes after the submission, sure.

Goodbye AC, we won't miss you, nor your alleged 5-digit uid account.

Re:RIP

[Ogive17]

I've been reading this site for over a decade. Very few topics (relative to total stories) ever eclipsed 1000 comments.

Re: RIP

[Anonymous Coward]

Typical 1995 slashdot posting:

50 substantive comments
450 goatse links

Nostalgia is fun! (From a 4-digit uid who is too lazy to log in from his phone).

Re: RIP

[Anguirel]

You forgot the 80 attempts at "First Post"

Re: RIP

[bugnuts]

5-10 years ago it also would've been more significant.

It wasn't that long ago where chess was able to beat a master.

PS: 5 digit uid.

Can we move the goalpost?

[Anonymous Coward]

I want to say "Nice but computers will never beat us at X" but I don't know what X is. Help?

Re:Can we move the goalpost?

[Jason Levine]

I want to say "Nice but computers will never beat us at X" but I don't know what X is. Help?

Tic Tac Toe. A computer can never beat me at Tic Tac Toe. Of course, given a good enough computer, I'll never beat it at Tic Tac Toe either.

Re:Can we move the goalpost?

[Anonymous Coward]

Eventually of course computers will best us at everything.

After all, we are just computers made out of meat.
It would be pretty arrogant to presume that 80 kilograms of meat is the ultimate intelligence in the universe.

Re:Can we move the goalpost?

[spacepimp]

A computer will never beat us at offline interpretation of literature. Well at least not for a long time.

Re:Can we move the goalpost?

[edremy]

Umm, be careful about that. Topic modeling is a thing, and even if it and similar techniques are only useful right now for digesting large amounts of text I suspect we'll see interesting improvements in the next decade or two.

Re:Can we move the goalpost?

[spacepimp]

Sure. The nuance of literature is not overly binary. Most AI attempts at parsing phrases are little more than workarounds. Actual Intelligence in computing is at its earliest stages. Do I predict that one day a computer will be able to give compelling interpretations of a text ... YES... However not in isolation from internet reviews or a data pool of prior interpretations to parse with a markov engine...

Re:Can we move the goalpost?

[Anonymous Coward]

The only winning move is not to play

Re:Can we move the goalpost?

[Anonymous Coward]

Learning. Current computer systems require much more learning to achieve the same levels as we do. My 4 year old nephew sees a picture of an elephant and is good at identifying elephants from then on. Googles image recognition took 10s of thousands of images of elephants to be able to reliably recognize them.

One may want to say that this will improve, but we run into a catch 22 which is already being witnessed. Computers are considered superior because they make far fewer mistakes. But to start to operate at the level of a human, you need to have processing power comparable to what a human brain can manage. It's seeming this is impossible to do with a single processor, and will need to use massive clusters of computers. And then we run into the problem seen with modern super computers. One CPU is unlikely to make mistakes, but in massive clusters, mistakes become inevitable (see probability of unlikely events in large populations). And now you need error handling to ensure that part of your system isn't making mistakes. But these systems can't be fool proof. So you start having computer system which have the same problems humans have. They're no longer perfect, and performance takes a hit as it's having to check itself as well as the interconnects for the CPUs are incredibly slow due to limits of physics. It was an area I started researching when I was searching for a thesis topic, it's fascinating and becoming a bigger and bigger problem as systems become more complex.

It's almost ironic really, as computers become more complex and approach the complexities of the human brain and able to do the same things, all the problems of the human brain start to manifest in the computer systems.

Re:Can we move the goalpost?

[Visarga]

Machine learning techniques work well with incomplete and noisy data, so they would work well even when there are occasional failures in some nodes.

Re:Can we move the goalpost?

[ledow]

"The Game".

See? They've already lost.

Re: Can we move the goalpost?

[Anonymous Coward]

They already have. God's kind of an idiot.

Re:Can we move the goalpost?

[Ginguin]

I hate you so much right now. I was doing so well, too! With this audience I am sure quite a few people swore out loud (as I did). For those that didn't get the reference... welcome to the big leagues: The Game

Mildly interesting

[Anonymous Coward]

But not nearly a milestone such as the first Chess grandmaster win or the Jeopardy win.

I would like to see how well the computer does at Diplomacy with its complex negotiations.

No doubt the AI singularity will come, but we aren't even close yet.

Re:Mildly interesting

[Anonymous Coward]

...and then WHAM, it hits you over the head and you look up too late to see anything but dust.

I strongly suspect we will have the first higher animal-level general AI before the end of the year, and the Singularity not long after that (probably before the end of 2017). People will complain when it does because they won't see it like it is supposed to be the Rapture or something, then a few months later, they'll be falling in love with their phones and buying realistic robot bodies for them.

a life0cidal bi-deitical zionist nazi?

[Anonymous Coward]

it gets even creepier? cease fire.. in the moms we trust

like chess,

[vasilevich]

What happened in Chess will happen in Go... it is just a matter of time...

Re:like chess,

[dargaud]

Are there any classic games left where humans have a marked advantage over computers ?

Re:like chess,

[vasilevich]

Poker, perhaps, but certainly not for long.

Re:like chess,

[mattdunelm]

Tag?

Re:like chess,

[Anonymous Coward]

Predator drones always beat humans at tag...

Re:like chess,

[Anonymous Coward]

The board game Diplomacy.

Re:like chess,

[Tom239]

Are there any classic games left where humans have a marked advantage over computers ?

Hex. It has neither a centuries-long tradition nor a large player base, but many of us who have learned the game consider it classic in the sense of having great depth and beauty. There is active work on Hex programs and they are still far behind the best human players.

Re:like chess,

[Anonymous Coward]

Ping-Pong. Although, computers may win at ping.

Re:like chess,

[Megane]

They could expand upon the principles of winning at Rock-Scissors-Paper to win at Tag. And as another poster has pointed out, they're already pretty good when drones are involved. Or laser turrets.

Re:like chess,

[Anonymous Coward]

Hex is difficult largely for the same reason that Go is difficult. It's an additive game.

Re:like chess,

[amRadioHed]

They can already win at Pong too.

Re:like chess,

[braindrainbahrain]

Some games were designed to deny the computer's advantage over humans, these would be Arimaa and Octi. I understand the former has indeed been defeated by the machines by now. The latter, Octi, probably survives because it is such an obscure game.

Re:like chess,

[I'm not god any more]

Are there any classic games left where humans have a marked advantage over computers ?

obligatory: Game AIs

Re:like chess,

[Anonymous Coward]

Poker?! I hardly know'er!

Significance

[wonkey_monkey]

Science magazine says the loss may be less significant than it seems at first.

Err, no, not really. It's still has about the same significance as it first seemed to me.

How is this AI?

[Anonymous Coward]

It's just a bunch of rules. If it's really AI, does it get annoyed? Flustered? Does it daydream? Does it feel happy when it wins?

Re:How is this AI?

[BigFire]

The fact that there's so little rule makes the game harder for computer. And the possibility of play is many order of magnitude harder than chess.

Interesting game....

[tekrat]

The only winning move is not to play.

The system is learning

[Anonymous Coward]

Whats significant here isn't that it beat him in the first round, or that it may win, but that if it wins, it will be a remarkable achievement because the method is not the same one that was used to beat kasparov or used by deep blue to beat players in jeopardy. It is in a more general purpose algorithm that is being used. This system is actually learning.

Google's Deep Blue moment ?

[Anonymous Coward]

Surely, there was no human help behind this one ?

So-called 'AI'

[kheldan]

We keep hearing about 'AI this' and 'AI that', but by my standards there is no such thing. I can't sit down with a computer, have a conversation, and for one second feel like I'm talking to the intellectual equivalent (or better) of a human being, therefore there's been no such thing an 'artificial intelligence' as of yet. All we've got are so-called 'expert systems', which at best mimick a human being's ability to think -- but only on specific subjects. Even so-called 'machine learning' is a far cry from actual sentience. Honestly, media people, can we get that straightened out?

Re:So-called 'AI'

[Dutch Gun]

"Expert systems", as you indicated, is a far better term for what we have these days, but I think it just doesn't have that same media-friendly click-generating panache. Hell, I'm just thankful whenever the media calls it "AI" instead of the ridiculous over-broad term "robots" when they're just talking about computer algorithms.

Re:So-called 'AI'

[Anonymous Coward]

We keep hearing about 'AI this' and 'AI that', but by my standards there is no such thing. I can't sit down with a computer, have a conversation, and for one second feel like I'm talking to the intellectual equivalent (or better) of a human being, therefore there's been no such thing an 'artificial intelligence' as of yet. All we've got are so-called 'expert systems', which at best mimick a human being's ability to think -- but only on specific subjects. Even so-called 'machine learning' is a far cry from actual sentience. Honestly, media people, can we get that straightened out?

No. If you think you need to be able to have a conversation with something to be able to call it intelligent then clearly dogs aren't intelligent. You need to start somewhere.

Re:So-called 'AI'

[kheldan]

I don't consider 'intelligent' and 'sentient' to be interchangeable terms, and since you bring it up I'm not so sure that the term 'artificial intelligence' is all that appropriate, either. 'Artificial Sentience', perhaps?

I don't expect to have a conversation with a dog. I expect the dog to listen, and me to have to make guesses based on the dog's behavior as to whether or not it's listening to me/comprehending me. But since you bring it up: We so far haven't even managed to produce a machine that can 100% emulate a dog's cognitive function, either. I think they barely have been able to emulate a housefly brain, haven't they? And that at great expense. Nope, sorry, I'm sticking to what I said before: I, personally, won't consider anything to be 'sentient' until it can have a casual conversation with me, and leave me wondering whether or not it was a person or a machine. It needs to have a sense of humor all it's own (not just regurgitating someone else's jokes). In fact I think the best test of whether a machine is truly sentient on a par with human beings, is if it can create art, and I don't mean random digital art, I mean formulate an intent to evoke an emotional response with it's artwork, then manage to pull it off with some degree of success. I think that more or less encapsulates what it means to be human, and when the day comes that we can create a machine that does that, then I'll admit we have Artificial Sentience. Until then it's just 'expert systems' and clever simulations.

Re:So-called 'AI'

[Lotana]

You are referring to something which is called "Strong AI". This article talks about "Weak AI". The difference is that the weak AI is focused on mastering a small, specialized set of problems (Playing Go, chatter-bots, etc) while strong AI takes on the big consciousness part of the problem. This distinction was created by John Searle who came up with a thought experiment that counters Turing Test: Chinese Room

Weak AI has had much more research focus and success. One big part of the reason is that it is so much easier to test: Just keep the AI playing against players and keep score. Secondly it gives a clearly defined scope to the problem: Go AI does not need to be able to carry out a deep, meaningful conversation while playing. Logic here is that AI that masters a subset of general problems in a believable way is still useful and worthwhile. Medical expert system is still helpful to the users despite not being able to play chess.

With Strong AI where do you even start? This rapidly runs into the philosophy of the mind questions. How do you test whether something is conscious? How do you express such concepts as code in hardware? Trouble is that we can't even exactly model how the brain functions!

IMO, the only practical solution at the moment is to dodge answering the question and just create an artificial simulation of the human brain down to the base level and see what happens. Unfortunately even now we don't have the hardware powerful enough for such a simulation.

Actual game, anywhere ?

[Laxator2]

Is there a link to the actual game played ? The pgn file, for example ? Would be curious to see the actual moves.

Re:Actual game, anywhere ?

[Anonymous Coward]

go uses sgf

https://gogameguru.com/i/2016/03/Lee-Sedol-vs-AlphaGo-20160309.sgf

Re:Actual game, anywhere ?

[Anonymous Coward]

Since I doubt that most people unfamiliar with go have a way to view an sgf file here is link to the gogameguru article about the game. At the bottom there is a javascript applet you can use to play through the game.

https://gogameguru.com/alphago-defeats-lee-sedol-game-1/

Re:TRUMP RULES! CRUZ DROOLS!

[Anonymous Coward]

I flew to NY and drove through Queens... they didn't speak any different English than anywhere else I've been.

What hardware do they use?

[jernejk]

I can't find anywhere what hardware is used in the game, and also what hardware was used for training.

I assume it's a cluster of computers talking to each other, writing down trees of possible moves (to RAM, but still writing down), after they have played more games than a single human could in their lifetime against one human brain who is not allowed to talk to other players and has only seen a small fraction of games the AI version has.

Yes, computers with enough resources to compute numerous trees will win. But human brain is still more creative. And incredible, considering how much it can do on a much smaller sample of data and ridiculously small short term memory (sth like 7-10 items only!)

Re:What hardware do they use?

[chuckugly]

To make it fair we could limit the AI to the physical size and energy constraints of a human brain ;)

Re:What hardware do they use?

[Anonymous Coward]

I can't remember the numbers exactly, but I read somewhere that it's a cloud cluster of about 1K CPUs and 10K GPUs.

Eh

[Anonymous Coward]

Wake me up when AI can consistently beat my 5 year old at Candy Land. It can't even gain an advantage in simple child games.

Deepmind Video

[Roceh]

Here's a talk by deepmind about this AI https://www.youtube.com/watch?...

Alphabet

[sexconker]

Alphabet's AlphaGo.

Re:Multi-billion dollar corporation defeats one gu

[Anonymous Coward]

Same thing with Deep Blue, but theses days a cheap desktop computer can beat Deep Blue.

Lee spent way less energy than the data center

[SDPost]

Lee spent a lot less energy than the data center that powered AlphaGo. Even if lee burned 3000 calories in 3.5 hours, that would only be around 3.4 kilo watt hour. AlphaGo's energy usage would be in the tens (or hundreds?) of "mega" watt hour.

Re:Lee spent way less energy than the data center

[Anonymous Coward]

So What?
If we built an AI that is twice as intelligent as the most intelligent human that ever exist but consume 50 times more energy, do you think it is unfair? will the AI care?
can you make a human twice as intelligent as a genius?, no you cannot, bottom line is that what matters is the intelligent level, I'm sure the AI will find a way to design a more efficient machine shall it wish

Paging swillden

[Anonymous Coward]

Where are swillden or his brother dwillden? For goodness sakes, there's a Google related story that requires someone to wave the company flag. Get out here boys!

For the next game, Lee is white (disadvantaged)

[Anonymous Coward]

It's not just that AlphaGo won the first game. AlphaGo won the first game AS WHITE! As I haven't been seeing this pointed out elsewhere...

The other human professionals who read out this game to score it said that the win was beyond the Komi (the 7.5 points granted to white due its not going first). IOW, AlphaGo beat TWO levels of disadvantage; playing as white (player 2) and achieving more than the balance granted by komi.

I think Lee Sedol will have a much more difficult time in the next game as AlphaGo will have black (first move). And while komi is supposed to compensate for the first move advantage, it doesn't take away the feeling from the _human_ playing white that they play the entire game attempting to come from behind.

The acs are the most interesting

[Anonymous Coward]

Why post anonymously?

Not yet solved

[Mostly a lurker]

Contact bridge is one game where (in spite of some serious efforts like Ginsberg's Intelligent Bridge Player) AIs can still not play near the level of top human players. The game combines imperfect information with being a partnership game. Perhaps an even greater challenge, you are prohibited by the rules from using optimum bidding systems and card signalling methods as these are too difficult for the average player to defend against.