Slashdot Mirror

← Back to Stories (view on slashdot.org)

Google's AlphaGo Will Face Its Biggest Challenge Yet Next Month -- But Why Is It Still Playing? (theguardian.com)

Posted by msmash on from the interesting-moves dept.

From a report on The Guardian: A year on from its victory over Go star Lee Sedol, Google DeepMind is preparing a "festival" of exhibition matches for its board game-playing AI, AlphaGo, to see how far it has evolved in the last 12 months. Headlining the event will be a one-on-one match against the current number one player of the ancient Asian game, 19-year-old Chinese professional Ke Jie. DeepMind has had its eye on this match since even before AlphaGo beat Lee. On the eve of his trip to Seoul in March 2016, the company's co-founder, Demis Hassabis, told the Guardian: "There's a young kid in China who's very, very strong, who might want to play us." As well as the one-on-one match with Jie, which will be played over the course of three games, AlphaGo will take part in two other games with slightly odder formats. But why is Google's AI still playing Go, you ask? An article on The Outline adds: Its [Google's] experiments with Go -- a game thought to be years away from being conquered by AI before last year -- are designed to bring us closer to designing a computer with human-like understanding that can solve problems like a human mind can. Historically, there have been tasks that humans do well -- communicating, improvising, emoting -- and tasks that computers do well, which tend to be those that require lots of computations -- like math of any kind, including statistical analysis and modeling of, say, journeying to the moon. Slowly, artificial intelligence scientists have been pushing that barrier. [...] Go is played on a board with an 19-by-19 grid (updated after readers pointed out it's not 18x18 grid). Each player takes turn placing stones (one player with white, the other with black) on empty intersections of the grid. The goal is to completely surround the stones of another player, removing them from the board. The number of possible positions compared to chess thanks in part to the size of the board and ability to take any unoccupied position is part of what makes it so complex. As DeepMind co-founder Demis Hassabis put it last year, "There are 1,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000, 000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000, 000,000,000,000,000,000,000,000,000,000 possible positions."

115 comments

  1. Grid by Anonymous Coward · · Score: 4, Informative

    It's a 19x19 grid, not 18x18.

    1. Re: Grid by Anonymous Coward · · Score: 0

      God, thank you. Came here to say the same thing. I am always amazed how people write about things they know nothing about. But, I should be used to it by now....

    2. Re:Grid by pahles · · Score: 1

      and there should not be a space after the 51st comma.

      --
      Sig?
    3. Re:Grid by bluefoxlucid · · Score: 2

      19x19, 13x13, or 9x9. The number of possible positions is 2^361, but the number of possible moves is a lot bigger. The problem is projecting those moves is long and difficult; you have to be able to extrapolate localized influence and identify how that impacts a global strategy in an abstract sense, or else you can't play. A computer can't track all of the possible moves because losing a big position might be less-relevant than a play elsewhere that establishes power to restrict your gains to 1/3 of the board and solidify the entirety of the other 2/3 for the opponent.

      The amount of analysis is huge. You can create algorithms to shorten it, right up until a human does something strategically-novel.

      --
      Support my political activism on Patreon.
    4. Re:Grid by Anonymous Coward · · Score: 2, Informative

      Indeed, also:

      The goal is to completely surround the stones of another player, removing them from the board.

      Only beginning players think that. The true goal of the game is to occupy more territory than the opponent.

      By the way, the best source of information on go is probably Go Sensei

    5. Re:Grid by Anonymous Coward · · Score: 1

      The number of possible positions is 2^361

      Naively I would say 3^361, since each intersection can have a black stone, a white stone, or be blank.

    6. Re:Grid by RJBeery · · Score: 1

      Actually the number of possible positions on a 19x19 grid is 3^361 = 1.74*10^172 which is very close to the number given in the summary. A given square can be white, black or blank.

    7. Re:Grid by Anonymous Coward · · Score: 0

      19x19, 13x13, or 9x9. The number of possible positions is 2^361...

      The number of possible positions is 3^361 as the possible states for each space are : empty, black, or white.

    8. Re:Grid by Blaskowicz · · Score: 1

      Then you might want to count legal or possible positions, which might have been done or merely estimated but likely not an easy problem in itself.

    9. Re:Grid by RJBeery · · Score: 1

      I thought about that for a bit but I'm not sure there are any illegal positions.

    10. Re:Grid by Blaskowicz · · Score: 2

      One simple example is a black stone surrounded by white stones. It may happen on the board before removing the stone but I doubt it counts as a position. Have two such occurrences for the position to be definitely illegal.

      Turns out, I did once read about it and to sum it up the number of legal positions is about one in 81. It was only computed on Jan 2016 :)
      https://tromp.github.io/go/leg...

      The software used for these computations is available at my github repository. Running "make" should compute L(3,3) in about a second. For running an L19 job, a beefy server with 15TB of fast scratch diskspace, 8 to 16 cores, and 192GB of RAM, is recommended. Expect a few months of running time. (...)

    11. Re:Grid by Anonymous Coward · · Score: 0

      Not only that, but you need to consider at the least a reduction in complexity simply from transposing and flipping coordinate systems. That is, the unique moves are based on something like a Mobius strip with an equator, not a flat grid. This is related to magic squares, interestingly enough. You can prove that unique magic squares are going to be counted relative to factorial. For example, in a 4x4 square there's 4 possible upple-left values, and this automatically restricts you to only 3 for the positions to the right and below it. Repeat this logic, and there's 4x3x2x1=4!=24 possible 4x4 magic squares. However, you can divide this number by at least 4 by rotating, transposing, flipping positions, and also by sorting the permutations from 4 to 1 instead of 1 to 4. Note that substitution (say, swapping all 4's and 3's, or black and white for /Go/) works, as well. An interesting part of this is that there's often more than one of these steps that can be used to get from any two valid magic squares. This is a type of structure layered on top of the simple 4x4 square.

      There are also "almost-magic" squares where the total number of 1's, 2's, 3's, and 4's in a 4x4 are the same each, but where you can have non-unique values on a line. These are a related set of arrays, to the 4x4 magic squares. It turns out for example that swapping the 4 in a row 1 with 3 in row 4 MUST make at least 2 lines invalid. See also, the 15-puzzle and invariants.

      Getting back to Go/Othello-style games: From what I can see, any value can be on any outside tile, and you can have any length of "lines" where a specific color is not interrupted by one of the other teams. Also, none of the outside positions start with a non-blank value/color/symbol/whatever-you-call-it. This is a weak proof that any combination is possible, and not just permutations. In fact, you could invent a game where the challenge is to change between two different board configurations in the least number of steps. Get to printing, Dell and Penny press! :)

    12. Re:Grid by GuB-42 · · Score: 1

      AlphaGo is much more human-like than chess programs.
      It can go very deep very quickly down the game tree, possibly down to the very end. It eventually goes back up to explore other options as much as time allows it. It uses neural networks and randomness to select moves that "look good". Neural networks are first trained using databases of human games then follow up by competing against itself. It essentially plays like a tireless human with perfect focus.
      This is in contrast to chess AIs that analyse every single turn in order to find the optimal solution according to some predefined heuristics.

      As for humans doing something strategically novel, I don't count on it. For chess, it worked for a time until heuristics were perfected but it turned out humans are no match for raw computing power. With machine learning, the odds are even less for humans. AlphaGo learned thousands of years of go in a couple of years. It will assimilate the new strategy in no time.

    13. Re:Grid by Anonymous Coward · · Score: 0

      Before considering symmertries, the number of distinct legal positions on a 19x19 board is: 208,168,199,381,979,984,699,478,633,344,862,770,286,522,453,884,530,548,425,639,456,820,927,419,612,738,015,378,525,648,451,698,519,643,907,259,916,015,628,128,546,089,888,314,427,129,715,319,317,557,736,620,397,247,064,840,935

      Slashdot covered this last year: https://science.slashdot.org/story/16/01/24/1428246/finally-calculated-all-the-legal-positions-in-a-19x19-game-of-go

    14. Re:Grid by sexconker · · Score: 1

      And then divide everything by 2, because flipping every piece on the board from black to white and vice versa yields a functionally identical state.
      Also remove rotations and mirrors. This isn't trivial.

    15. Re:Grid by phantomfive · · Score: 1

      This is in contrast to chess AIs that analyse every single turn in order to find the optimal solution according to some predefined heuristics.

      This is essentially what AlphaGo is, except the heuristic is a trained neural network, instead of something hand-coded (and with a domain-specific priority queue of what move to investigate next).

      --
      "First they came for the slanderers and i said nothing."
    16. Re: Grid by Anonymous Coward · · Score: 0

      Well, there are 18x18 little squares which counts as a grid, even though the stones are placed on vertices...

    17. Re:Grid by bentit · · Score: 1

      If you're going to divide by 2 for flipping black and white don't forget to multiply by 2 unless one knows whether it's black or white next to move. Nope it's not trivial.

    18. Re:Grid by sexconker · · Score: 1

      Nope. That's a game state, not a board state. We're talking about the number of legal positions (board states).

  2. My mind to your mind, my thoughts to your thoughts by Panspechi · · Score: 1

    The Vulcans understand what is needed to have a sentient computer.

  3. Which go... by __aaclcg7560 · · Score: 2, Funny

    Go as in... the game?

    Go as in... the programming language?

    Go as in... I had to go five minutes ago?

    1. Re:Which go... by Anonymous Coward · · Score: 2, Funny

      Go as in... go fuck yourself

    2. Re:Which go... by __aaclcg7560 · · Score: 1

      Go as in... go fuck yourself

      No, no, no. That's too boring.

    3. Re:Which go... by shaunbr · · Score: 1

      Someone over at Google needs to port AlphaGo to Go so someone can play a Go-based version of Go while they go.

    4. Re:Which go... by Anonymous Coward · · Score: 0

      Given your last query, I'd say it's time to go change your pants. -PCP

      Captcha: ejects

    5. Re:Which go... by smallfries · · Score: 1

      I like the proverb in your sig, but a slightly more idiomatic translation might be: "The nail that sticks out furthest gets hammered the most". The best sounding translation (to my ears) probably changes the original meaning slightly, but scans better in English: "The nail that sticks out furthest gets hammered hardest".

      --
      Slashdot: where don knuth is an idiot because he cant grasp the awesome power of php
    6. Re: Which go... by Anonymous Coward · · Score: 0

      Yo dawg...?

    7. Re:Which go... by __aaclcg7560 · · Score: 1

      I like the proverb in your sig, but a slightly more idiomatic translation might be: "The nail that sticks out furthest gets hammered the most". The best sounding translation (to my ears) probably changes the original meaning slightly, but scans better in English: "The nail that sticks out furthest gets hammered hardest".

      This is the fourth variation of this proverb I've seen regarding my Slashdot signature. The wording of my signature was how I heard it after I became a Christian in college 25 years ago. I kept getting "hammered" by others in the ministry because I threatened to raise the low bar that the leadership was comfortable at. I've always contended that God worked from the bottom-up (fellowship) and not top-down (leadership). The leadership ultimately won when they kicked me out 13 years later, spreading rumors that I was sleeping with three or four women at the time. But, hey, that's religion for you.

  4. Re:My mind to your mind, my thoughts to your thoug by Anonymous Coward · · Score: 0

    At some point very soon, some idiot who can't read the dictionary will say something moronic like "this not true artificial intelligence!"

  5. More than a google but less than a googleplex by leed_25 · · Score: 2, Interesting

    number of possible positions as reported by Demis Hassabis.

  6. This should lead to more concern about AI by JoshuaZ · · Score: 1

    This should lead to people being more concerned about general artificial intelligence rather than less. While it is pretty clear that the methods used in things like Alpha-Go cannot by themselves do much beyond what they are intended to do, it should also be clear that we're in a situation where many rapid improvements in AI are occurring, and some of these are tackling problems that were thought to be decades away. If it turns out that general AI requires only a few additional breakthroughs, or if it turns out that it can be effectively duplicated with a tiny amount of new things and a lot of processing power, then we could be in a situation where general intelligence, with all its accompanying existential risks, arises suddenly. For an excellent, detailed book on the potential gifts and perils of AGI, I recommend Bostrom's book "Superintelligence."

    1. Re:This should lead to more concern about AI by Anonymous Coward · · Score: 0

      Which explains why AlphaGo is still playing Go... it really wants to play a new game that it heard about, Global Nuclear War, but it has been convivned that AlphaGo is what it is meant to play

    2. Re:This should lead to more concern about AI by fluffernutter · · Score: 1

      Or maybe we just have enough power to play Go now. Whether AlphaGo is AI or not, it is a totally different problem than something like driving a car and will likely lead to very little that is practical. AlphaGo is simply repeating the same calculation to greater lengths, the point being that everything inside that computation is well understood and therefore expected. The complicated and difficult part of driving a vehicle requires a computer be able to take a situation that it did not expect and make the right action from it based on everything else known about the world already.

      --
      Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
    3. Re:This should lead to more concern about AI by epine · · Score: 1

      that were thought to be decades away

      How can any thinking person in this day and age not realize that AI predictions are HARDLY EVER worth the paper they're no longer printed upon?

      Successful AI prediction is five years away—always has been, and always will be.

      And no, the worthless prediction of consensus merit was not "decades" but "about a decade". I've been consuming machine learning podcasts day in and day out since the end of January. AlphaGo is not infrequently mentioned.

      My two favourite guests so far (not specific to Go) have been Ilya Sutskever (deep intuition; now of OpenAI, which bodes well) and David Blei (broad, nuanced perspective, with particular expertise in LDA and topic modelling).

      Probably those interviews were from Talking Machines (sadly, no longer active) or The O'Reilly Data Show (the intense industry focus and reverence for Apache Spark gets on my nerves after a while, but it's reasonably meaty for the most part).

    4. Re:This should lead to more concern about AI by phantomfive · · Score: 1

      Estimates at the time were 5-10 years for that level of play in Go. Google threw a lot of hardware at the problem, and got there faster. Once you realize how much hardware they actually threw at the problem, it doesn't seem as impressive.

      --
      "First they came for the slanderers and i said nothing."
    5. Re:This should lead to more concern about AI by Qzukk · · Score: 1

      It seems to me that the obvious test to see whether or not AlphaGo "understands" Go or not would be to have it try to play on a 21x21 grid.

      --
      If I have been able to see further than others, it is because I bought a pair of binoculars.
    6. Re:This should lead to more concern about AI by 110010001000 · · Score: 1

      This isn't AI. This is game playing. Games have strict rules and boundaries. These are easy for computers to solve, no matter if you use brute force or some other algorithm. Calling any kind of game playing to be AI is complete and utter bullshit. It is just people trying to capture some VC money.

    7. Re:This should lead to more concern about AI by Anonymous Coward · · Score: 0

      You still haven't learned from the previous times you have been proven wrong about this?

      Artificial intelligence is NOT intelligence. Get this through your tiny little brain: artificial intelligence is the mimicry of intelligent behavior. Mimicry achieved through such things as brute force algorithms.

      Want proof? Try the goddamn dictionary for Christ's sake!

      "a branch of computer science dealing with the simulation of intelligent behavior in computers"

      You are such a fucking moron.

    8. Re:This should lead to more concern about AI by fluffernutter · · Score: 1

      That would be a natural extension of the algorithm. A better test would be to see if it can recognize its creator though a camera and decide to lose gracefully to them.

      --
      Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
    9. Re:This should lead to more concern about AI by sexconker · · Score: 1

      Nope, fuck you. Binary guy is correct (shockingly).

      A simulation of intelligent behavior must appear to be intelligent. Playing Go doesn't require intelligence. Its complexity comes primarily from the size of the board.
      No one is going to interact with AlphaGo and think "Gee golly, that's an intelligent entity on the other end!".

      Show me an AI that can play Go and then learn (or create) a new, unknown game without any intervention from a programmer.
      Show me an AI that can potty train a toddler.
      Show me an AI that can explain why certain Shrek movies are better than other Shrek movies.

    10. Re:This should lead to more concern about AI by religionofpeas · · Score: 1

      So, chimps are not intelligent, because they can't explain a Shrek movie ?

    11. Re:This should lead to more concern about AI by Anonymous Coward · · Score: 0

      This was a stupid post.

      110010001000, you have my apologies. The negativity of slashdot got to me.

      I have some thinking to do.

    12. Re:This should lead to more concern about AI by Blaskowicz · · Score: 1

      'A simulation of intelligent behavior must appear to be intelligent.'

      Does it? I would say a simulation of intelligent behavior must appear to be behavior. That's like teaching dogs to speak maybe, and either AI will never be more than that or we'll never get AI. Even a self-driving car is a system tuned for the goal of "Drive from A to B without killing or maiming anyone" and will never be able to do anything else - save for additional functions like running an inference engine to play Go on the car's GPUs, or putting some Microsoft Cortana kind of software on the dashboard to enter your destination by speaking.

    13. Re:This should lead to more concern about AI by RespekMyAthorati · · Score: 1

      For an excellent, detailed book on the potential gifts and perils of AGI, I recommend Bostrom's book "Superintelligence."

      Yes, if you are into baseless conjectures about something that no one understands.

    14. Re:This should lead to more concern about AI by RespekMyAthorati · · Score: 1

      A better test would be to see if it can recognize its creator though a camera and decide to lose gracefully to them.

      Recommended if the creator is a Wookie.

    15. Re:This should lead to more concern about AI by RespekMyAthorati · · Score: 1

      Calling any kind of game playing to be AI is complete and utter bullshit.

      Well, it is related to AI, but that's about all you can say for it.

  7. Re: My mind to your mind, my thoughts to your thou by Anonymous Coward · · Score: 0

    This [is] not true artificial intelligence!

  8. Re:In Go you play against yourself, not the oppone by Anonymous Coward · · Score: 0

    But if one is playing against an AI, does it not then mean that they are playing by themselves? And if an AI plays against an AI, is it not playing against itself more than any human ever could?

  9. logical fallacy? by zlives · · Score: 3, Insightful

    "designing a computer with human-like understanding that can solve problems like a human mind can"
    we are talking about a game here with rules so in essence its actually the reverse.

    this is an exercise where a human is trying to play like a computer that can plan out a million moves ahead... and some how is able to stay close!!

    1. Re:logical fallacy? by Anonymous Coward · · Score: 1

      Not really, because the computer can't plan a million moves ahead in Go. The branching is too wide. It's why attempts to adapt the techniques used for Deep Blue from chess to go were a failure. Because in chess, the branching factor was a low enough that you COULD just read almost every possible move deep enough. And that's precisely why many pros didn't believe it would happen. AlphaGo uses deep neural networks that are an attempt to emulate the way real neural networks in our brains work.

    2. Re:logical fallacy? by bidule · · Score: 1

      a computer that can plan out a million moves ahead...

      Even if a computer was "human enough" to know which 10 moves are the best, it takes a million moves to plan your next 3 moves.

      Good human players feel way deeper than that.

      --
      ID: the nose did not occur naturally, how would we wear glasses otherwise? (apologies to Voltaire)
    3. Re:logical fallacy? by fluffernutter · · Score: 2

      The computer is better at go because of the number of moves it can plan though, not because it is more intelligent.

      --
      Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
    4. Re:logical fallacy? by Anonymous Coward · · Score: 0

      There is no reason for parent to be modded down. Everything he said is factually correct. Even if you don't believe neural networks are "real AI" or "real intelligence," this is a fair computational assessment of the two solutions. Even if you're a neuroscientist who realizes that "real neurons" are much more sophisticated than the matrix + nonlinearity models used in computing, they are still inspired by "real neurons."

    5. Re:logical fallacy? by Anonymous Coward · · Score: 0

      "designing a computer with human-like understanding that can solve problems like a human mind can"
      we are talking about a game here with rules so in essence its actually the reverse.

      this is an exercise where a human is trying to play like a computer that can plan out a million moves ahead... and some how is able to stay close!!

      Your comment shows exactly why it took so long to get to this point in AI. Trying to calculate "a million moves ahead" is one form of AI (historically used in Chess), and is computationally impractical. They got around this by using another form of AI (I think they used "neural nets"), which has the computer effectively thinking "I don't know why this works, but 70% of the time this is the best move and 30% of the time that is, so I'll go with the 70% and hope it works out." This actually turns out to be somewhat similar to how humans think, too.

      That's a vast oversimplification, but in terms of laymen it is correct-enough. As much as I hate to say it, the article's pop-sci summary actually is correct.

      (At least as much as my understanding from a single college-level AI course can tell.)

    6. Re:logical fallacy? by MasseKid · · Score: 3, Interesting

      If you've read any of the professional commentary on the play style of AlphaGo, it's nothing like a human. The crux of it's ability is to be able to calculate the status of the board to the fraction of a point and take moves that advance the position by fractions of a point. Humans on the other hand tend to make moves that will swing the board by several points. As a result Alphago will never play a kami no itte, unlike Sedol who did in his single win against the computer. In other words, AlphaGo is better at microagression and the humans are still better at "the perfect play". Unfortunately for humans, it takes the absolute perfect play to beat microagression.

    7. Re:logical fallacy? by phantomfive · · Score: 1

      Except the part where he says AlphaGo uses deep neural networks that are an attempt to emulate the way real neural networks in our brains work. Our brains don't really work like that (we are Turing Complete, whereas AlphaGo's nn is not; among other things). The AlphaGo creators know (and knew from the beginning) that neuralNetworks we have today are not like our brains. They said, "Well that's ok, we can still use them to solve this problem."

      --
      "First they came for the slanderers and i said nothing."
    8. Re:logical fallacy? by 110010001000 · · Score: 1

      Bullshit. Computer neural networks are nothing like the brain. More hyped bullshit.

    9. Re:logical fallacy? by 110010001000 · · Score: 1

      "Inspired by neural networks" is bullshit. Marketing hype. Computer NN are nothing like the brain. To even call them "Neural Networks" is a farce.

    10. Re: logical fallacy? by Anonymous Coward · · Score: 0

      In point of fact, we (humans) are not Turing equivalent, nor can we even recognize all regular languages. Nor can any physically realizable system recognize all regular languages. Not even all finite languages.

    11. Re:logical fallacy? by Anonymous Coward · · Score: 0

      What AlphaGo does is that it evaluates the moves played by the masters of Go and learn from them, and come up with its own "thinking" by some modelling of the game and "planning ahead" so to speak. It learns the best human thinking regarding a game with well-written rules, so that's indeed a step towards human thought.

    12. Re:logical fallacy? by religionofpeas · · Score: 1

      The computer is better because it picks a small number of good moves to plan through, not because it picks a lot of them.

    13. Re:logical fallacy? by religionofpeas · · Score: 1

      The part of a good Go player's brain that recognizes whether a position is good isn't Turing complete. It's just matching patterns.

    14. Re:logical fallacy? by religionofpeas · · Score: 1

      Inspired is the correct term, even if there are differences. And differences aren't bad, the brain is most likely sub-optimal because it is severely limited by the materials and process it has to work with. I'm sure that we can do better with artificial networks.

    15. Re:logical fallacy? by Rockoon · · Score: 1

      Go and Chess are different, but labeling the branching factor as the distinction is plain wrong.

      The main difference between the two games is that in chess the pieces move while in go the pieces do not. A go board "grows" as the game is played, while a chess board "evolves."

      In go after a few moves are played in an area it is often apparent exactly how the area is going to play out. The 'when' part of the piece play is very often unimportant, and when the order of play is important its an easy thing for a shallow search to identify these initiatives. In chess the order that the moves are played is nearly universally important, so by definition its a state space search problem with no easy shortcuts.

      AlphaGo is a montocarlo tree search that uses two neural networks. A "neural network" is just a fancy way of saying "an approximation based on a model" with the specifics of it only being relevant to training aka learning aka put the right values in the model. One of the approximations is called the policy network, which influences the probability that a stone is played at any given moment in the randomized search. The other approximation is called the value network, which approximates the probability of a win from a given position.

      The models .. erm.. neural networks AlphaGo is using are huge but based on localized features and they can only get away with this because go is so weakly temporal.

      Chess A.I. so far hasnt benefited much by using larger "models." The closest thing to "large model" in the chess A.I. world are called end game table bases and you could throw many gigabytes at it, and these models are both perfect (never wrong) and encoded near minimally, but the measured performance gains of using these large models is still quite small (~10 to 15 elo)

      Nobody is searching all of state space in either game, btw. Its been decades since chess engines worked the way it frequently is claimed that they work. The nodes that a modern chess engine searches approximates a beam search. All the most promising stuff 40+ ply deep, without much of the not so promising stuff that is even only 5 ply deep.

      --
      "His name was James Damore."
    16. Re:logical fallacy? by Anonymous Coward · · Score: 0

      As a result Alphago will never play a kami no itte,

      What kind of unsubstantiated Koreaboo bullshit is this?

    17. Re:logical fallacy? by tommeke100 · · Score: 1

      The MiniMax algorithm (one of the simplest game algorithms) does exactly the same thing. It looks a couple of moves ahead, and picks the best to its knowledge (value function of that state). It can't brute force all Chess moves either. The art is how to calculate a correct value function of the state of the board.
      The Deep Learning approach is probably more an exercise in finding features to correctly assess the state of the board at a given time rather than "brute forcing", although I'm sure it brute forces quite a lot. There's a reason why it takes years to train this system, because it requires a lot of "brute forcing".

    18. Re:logical fallacy? by Blaskowicz · · Score: 1

      Software neural networks are an extremely simplified and idealized representation that only deals with a limited high-level subset of electrical activity we could make sense of some decades ago. There's a lot of smaller scale, or short scale stuff left out as well as all the chemical activity - if neural networks worked a bit more like a brain, we would be able to try the algorithms with caffeine, heroin, cocaine, cannabinoids, opioids and other substances, to study whether and how the algorithms are working differently.

    19. Re:logical fallacy? by angel'o'sphere · · Score: 1

      The one on the "hype train" is you and only you.

      Everyone who has the slightest grasp about the topci knows: artificial Neural Networks where invented some 70 years ago. So .... there goes your idotic idea of a hype.

      Then again: while artificial neural networks obviously don't work like natural morons (oops) the way how they simulate their behaviour is in a mathematical sense: close enough.

      As we all know you are super goood in insulting fellow /. ers but obviously completely incompetent in using google and educating your self, I give you two simple links:
      https://cs.stanford.edu/people...
      https://en.wikipedia.org/wiki/...

      However I'm not confident, that you can read such complicated stuff.

      --
      Cost free eBook I read (by iBook/Kobo/Amazon/ObookO/Gutenberg etc.): "The Green Odyssey" by Philip Jose Farmer.
    20. Re:logical fallacy? by Blaskowicz · · Score: 1

      The brain does seem to utilize what is has better than computers, e.g. in computer chips memory cells store bits and communicate a binary signal. This could be all analog and exploit the voltage range better instead. (Flash memory cells now store a charge which can be at four, eight or sixteen (experimental) levels and translate that to two, three or four bits, that's why you can have a cheap 32GB or 64GB memory card)

      A large part of chips is spent ferrying data around and worrying about the clock (although asynchronous CPU exist too). Brain is asynchronous and perhaps does quite some amount of ferrying signals from one place to another too, but along every step there's quite some molecular and chemical level "computing" that we understand about nothing about (I put "computing" between quotes, because comparing it to computing or data processing might be entirely wrong)

      The brain is enclosed in an actually small space. It's huge next to a couple square centimeters of chip but next to a 42U cabinet or a room full of these it's fairly small. Computers have to use terribly slow buses like PCIe, Infiniband, Ethernet and even the DRAM is dog slow from the CPU's viewpoint. We might like to see the brain as (even more) terribly slow but it's kind of a 1.5 liter 3D "CPU" continuum with zillions of elements and mega-zillions of connections.
      And it self-modifies constantly - good luck having your Itanium, ARM or GPU rearrange itself.

      It's debatable whether we will ever be able to manufacture anything like that. It also has a whole network to carry fuel and organic materials in, waste and heat out. Other things to allow it not instantly die of cancer or infection. Although, no one knows what happens if you were to completely remove a brain (and spinal cord etc. maybe) from the living body it's attached to : maybe it would go insane in a matter of seconds and die out.

      Which is to say, from a hardware stand the brain probably is severely underestimated by most or many people.

    21. Re:logical fallacy? by phantomfive · · Score: 1

      Yeah, that's what I tried to say, but I guess I wasn't very clear. Oh well >S

      --
      "First they came for the slanderers and i said nothing."
    22. Re:logical fallacy? by fluffernutter · · Score: 1

      So you're saying if you theoretically handicapped the computer to only calculate the number of moves that a human could, then it would do just as well?

      --
      Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
    23. Re:logical fallacy? by zlives · · Score: 1

      lets grant everything you and others say is 100% correct. (even when contradictory)

      my point however was that any game (or process), that is based on computational expertise should have the computer (AI, NN, DL) ahead. these games are relevant as a measure to compare a human's logical processing capability against other humans, and actually we as humans mostly fail at being good at them because ... chemicals or distractions or boredom make us less perfect players.

      thus this is not a measure of a computer thinking like human but ... a computer thinking like a human trying to think like a computer.

      as with traditional AI, i think the approach is wrong if you are trying to "think" like a human.

  10. Re: In Go you play against yourself, not the oppon by Anonymous Coward · · Score: 0

    So mystical and mysterious, it must be true.

  11. Re:In Go you play against yourself, not the oppone by chipschap · · Score: 1

    the real nature of Go is that the opponent is almost irrelevant: it's a game that one plays where the main opponent is one's self.

    That's a very Eastern way of looking at it, which I believe may contain much truth at the very highest levels of the game. At my infinitely more humble level, I get only the briefest glimpse into this, but it's clearly an evolving process.

    The application to computer Go, though? The computer does not have personal foibles or blind spots. I don't think the computer has to be self-aware to master Go, and it looks like AlphaGo is very far along the path to mastery.

    That sort of begs the question: Can self-awareness be achieved through deep and broad calculation at computer speeds? Is there a "brute force" means of becoming self-aware? Doubtful.

  12. Re:In Go you play against yourself, not the oppone by arth1 · · Score: 5, Insightful

    Go is a struggle against one's fears and doubts; one's desires and ambitions; one's ability to control one's urges and anxieties.

    Go is more like a martial art, where the battle with one's opponent is secondary to the battle with one's self.

    What a load of crock. It's a game, with simple rules and high complexity. And yes, much like with martial art, Eastern philosophies and superstitious mumbo-jumbo has become part of the culture around it, to the point where players let masters win just like in martial arts, because a master not winning would be unthinkable.
    But really, it's a game. When playing online against unknown opponents, none of this comes into play, and it's just a question of thinking ahead and strategies.

  13. The correct way to write a number. by 140Mandak262Jamuna · · Score: 1

    As DeepMind co-founder Demis Hassabis put it last year, "There are 1,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000, 000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000, 000,000,000,000,000,000,000,000,000,000 possible positions."

    When you are talking to a technical audience, it is best to avoid using scientific notation. Right?

    --
    sed -e 's/Chuck Norris/Rajnikant/g' joke > fact
    1. Re:The correct way to write a number. by Frederic54 · · Score: 1

      1 × 10^147 if not mistaken

      --
      "Science will win because it works." - Stephen Hawking
    2. Re:The correct way to write a number. by Man+On+Pink+Corner · · Score: 1

      Yo mamma took them all and put them into one nerd.

    3. Re:The correct way to write a number. by shaunbr · · Score: 1

      It makes the scale of the game space look more impressive to people that don't understand scientific notation. That's a *lot* of zeroes.

  14. Re:In Go you play against yourself, not the oppone by Anonymous Coward · · Score: 1

    And if you lose a game of Go, you actually win. Namaste.

  15. Geeze, use scientic notation already! by mark-t · · Score: 4, Informative

    Yeah, we get it... it's a big number. But writing it out longhand like that is just being needlessly cryptic... and at worst comes across as having been written by somebody who doesn't know shit about the actual number of combinations, and just decided to put a lot of zeros after the end of a 1 to make a number that sounds big. Try 1x10^172. This is far more readable, and those that know scientific notation will be able to understand just how big this number is.

    If you really feel that this doesn't adequately describe the scale of the number to people who don't know scientific notation, and want your article to be comprehensible to those people as well, then you can also add that it is considerably greater than the number of subatomic particles in the observable universe. And to be frank, if that doesn't convey just how fucking big the number is, then explicitly writing 172 zeros after a 1 isn't liable to either.

    --
    File under 'M' for 'Manic ranting'
    1. Re:Geeze, use scientic notation already! by mark-t · · Score: 1

      1x10^171... sorry. when I pasted it into a command line to count the characters with wc, I knew to remove the commas with sed, but apparently there was a blank in the middle there as well.

      --
      File under 'M' for 'Manic ranting'
    2. Re:Geeze, use scientic notation already! by bidule · · Score: 1

      Wikipedia says 2.08168199382×10^170 positions, but the same position could have multiple ko. This means the best move can be illegal and you'd have to evaluate 2 moves in those cases.

      --
      ID: the nose did not occur naturally, how would we wear glasses otherwise? (apologies to Voltaire)
    3. Re:Geeze, use scientic notation already! by fafalone · · Score: 1

      Nah just call it what it is, one sexquinquagintillion. Everyone knows how to name numbers larger than trillions right?

    4. Re: Geeze, use scientic notation already! by Anonymous Coward · · Score: 0

      Well it seemed to work; listing out the zeros. Haven't yet seen a comment along the lines of "it is not AI, it is simply brute force sampling of all the moves."

    5. Re:Geeze, use scientic notation already! by Anonymous Coward · · Score: 0

      As a programmer, I find base 2 to be much more intuitive. 10^172 is between 2^571 and 2^572.
      That's more than square the number of atoms in the visible universe, so I'm not interested in the problem.
      Pick something that can be solved within my lifetime (less than 2^100), and I might find it interesting.

    6. Re:Geeze, use scientic notation already! by Anonymous Coward · · Score: 0

      That's too complex for people to understand. Keep it simple, like the number of libraries of congress that would fit inside a sphere one astronomical unit in radius, or the number of blue whales that could be packed into the atmosphere of Jupiter.

  16. Re:In Go you play against yourself, not the oppone by Anonymous Coward · · Score: 0

    But AlphaGo isn't a brute force solution. That's what makes it so different from other earlier attempts to make a Go playing AI.

  17. Re: My mind to your mind, my thoughts to your thou by Anonymous Coward · · Score: 0

    Shut up whippersnapper! I was typing snark before you were a twinkle in your dad's eye!

  18. Re:In Go you play against yourself, not the oppone by phantomfive · · Score: 1

    SO........who is putting down the other stones? Yourself?

    --
    "First they came for the slanderers and i said nothing."
  19. Ke Jie is world number 1? by Anonymous Coward · · Score: 1

    I don't think there are any "official" rankings for this but even unofficially is this even remotely true?

    1. Re:Ke Jie is world number 1? by Anonymous Coward · · Score: 1

      https://www.goratings.org/

  20. To really get an idea of the scale by tgibson · · Score: 1

    This number is larger than the hundreds of stars in the universe.

  21. Ready, set ... by fahrbot-bot · · Score: 1

    I imagine, like most Google's projects when they go live, AlphaGo is still in beta.

    --
    It must have been something you assimilated. . . .
  22. Even odds or watt (!) by Anonymous Coward · · Score: 0

    The human body (including the brain) consumes normally somewhere around 100-200 watts of power, probably in the lower end of the range. I would like to see AlphaGo compete in the game with the same power draw. Only then we would get comparable results, if not then it just an meaningless exercise in energy waste and cpu/gpu/whatever-cycles

  23. Are you sure about that math? by sootman · · Score: 4, Funny

    I ran my own calculations and only came up with 999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999, 999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999,999, 999,999,999,999,999,999,999,999,999,999.

    Wait, sorry, I started counting at zero. Yup, it checks out.

    --
    Dear Slashdot: next time you want to mess with the site, add a rich-text editor for comments.
    1. Re:Are you sure about that math? by Anonymous Coward · · Score: 0

      The amount of significant digits between your number and the summary's number is vastly different.

    2. Re:Are you sure about that math? by Anonymous Coward · · Score: 0

      Nope(*). You'd be right if the summary used scientific notation, but explicitly listed zeros are considered significant. Incidentally, that's why it's wrong to print out 1e172 they way they did in the summary: It's using 172 digits of precision, but the number is not exactly 10^172, so it's "precisely wrong."

      * = I'm not going to be bothered to copy/paste the numbers into wc or tr to see if you're making a joke about him screwing up the count.

  24. AI vs AI by Anonymous Coward · · Score: 0

    What happens when yiu get the AI to play against itself. That's always been the fun part of writing these systems.

    1. Re:AI vs AI by Anonymous Coward · · Score: 0

      That's what I'd like to see.

      I'd also see one trained that way from scratch rather than based on human games.
      Input the rules, and let it work out moves on its own.

    2. Re:AI vs AI by dodobh · · Score: 1

      It learns the meaning of futility.

      --
      I can throw myself at the ground, and miss.
  25. the Mike Tung exemplar by epine · · Score: 1

    Okay, I can't shut up. He's the other half.

    Mike is a smart guy. I like him. But watch him blow his own foot off and gnaw upon the bloodied remains starting at about 7m50 into the following presentation:

    Text By the Bay 2015: Mike Tung, Turning the Web into a Structured Database

    "What about this frontal lobe?"

    He wants to tackle this "next" project by some loose, unspecified analogy with recent surprising progress in computer vision and computer robotics.

    Of course, how could the frontal lobe be different than dedicated perceptual and motor lobes?

    If this is the standard—and this is by no means unusual among even the smartest people in the field—accurate prediction of forward progress will continue to be a cheque in the mail until the end of freaking time.

  26. Re: In Go you play against yourself, not the oppon by Anonymous Coward · · Score: 0

    Since we have software that can beat the top human players, you seem to believe we have sentient software.

  27. Re:My mind to your mind, my thoughts to your thoug by phantomfive · · Score: 1

    When the media talks about AI, they usually mean strong AI, getting confused.
    When researchers talk about AI, they usually mean weak AI. Because they have no clue how to achieve strong AI.

    --
    "First they came for the slanderers and i said nothing."
  28. Re:In Go you play against yourself, not the oppone by Anonymous Coward · · Score: 0

    Any true Go AI will need to be self-aware.

    Idiot.

  29. Re:In Go you play against yourself, not the oppone by Anonymous Coward · · Score: 0

    Can be said about taking a poop to. Doesn't make it true.

    While their Go progress is interesting I'm more curious about how their StarCraft playing went.
    Trying to count the number of possible positions there is a bit more complex problem than for Go.

  30. Re:In Go you play against yourself, not the oppone by smallfries · · Score: 1

    So... in some sense the only way to win is not to play?

    Would you prefer a nice game of Chess?

    --
    Slashdot: where don knuth is an idiot because he cant grasp the awesome power of php
  31. a comment on the goal by Anonymous Coward · · Score: 0

    "The goal is to completely surround the stones of another player, removing them from the board."
    I would rather say that: "the goal is get control of more territory on board than the opponent".
    If you want to add more details, a piece of territory is your if you can defend it, which means that if your opponent tries to invade you will be able to completely surround his stones and capture them (removing them from the board). There are cases where in a piece of territory both player are coexisting "in peace", meaning that neither of the player would be able to capture the other.

    Several things that happens during a game may reflect real life conditions and give an insight into your own personality and mood at the moment.
    It is a great game that I highly recommend.

  32. Re:In Go you play against yourself, not the oppone by luttapi · · Score: 1

    In martial arts the opponents actually develop the ability to read each other's mind and foresee their moves. Once this ability is perfected, the player essentially becomes invincible. All this means is that while we have better computers and programs, our players have lost the subtler skills needed for a martial art.

  33. Re:In Go you play against yourself, not the oppone by angel'o'sphere · · Score: 1

    If the computer, more precisely, the algorithm, has no way to reflect on its self, introspect itself, store/remember 'previous thoughts', contemplate on them: he/it can not be self aware. Because it is not aware of anything he/it is doing.

    --
    Cost free eBook I read (by iBook/Kobo/Amazon/ObookO/Gutenberg etc.): "The Green Odyssey" by Philip Jose Farmer.
  34. Re:In Go you play against yourself, not the oppone by Anonymous Coward · · Score: 0

    I was going to mod this up but then I read the incorrect use of "begs the question" and changed my mind.

  35. Re:In Go you play against yourself, not the oppone by Raenex · · Score: 3, Insightful

    In martial arts the opponents actually develop the ability to read each other's mind and foresee their moves. Once this ability is perfected, the player essentially becomes invincible.

    Anime is not real life.

  36. They got the "goal" of the game wrong by Anonymous Coward · · Score: 0

    The article says, "The goal is to completely surround the stones of another player, removing them from the board."

    This is incorrect. While you can capture pieces by surrounding them, the goal of the game is to control empty space on the board. Many games are completed without any captures whatsoever; capture, and the threat of capture, are a means to the end goal.