Slashdot Mirror

← Back to Stories (view on slashdot.org)

Human Go Champion 'Speechless' After 2nd Loss To Machine (phys.org)

Posted by msmash on from the still-struggling-but-won't-give-up dept.

Reader chasm22 points to a Phys.org report about the second straight loss of Lee Sedol to AlphaGo, the program developed by Google's DeepMind unit. The human Go champion, Sedol found himself "speechless" after the showdown on Thursday. The human versus machine face-off lasted more than four hours, which to Sedol's credit is a slight improvement over his previous match, which had ended with him resigning nearly half an hour remaining on the clock. "It was a clear loss on my part," Sedol said at a press conference on Thursday. "From the beginning there was no moment I thought I was leading." Demis Hassabis, who heads Google's DeepMind, said, "Because the number of possible Go board positions exceeds the number of atoms in the universe, top players rely heavily on their intuition." Sedol will battle Google's AlphaGo again on Saturday, Sunday, and Tuesday.

41 of 338 comments (clear)

  1. Milestone by Lisandro · · Score: 5, Insightful

    Having a competitive Go engine capable of beating a 9-dan player is huge. Huge.

    1. Re:Milestone by irrational_design · · Score: 5, Funny

      My brother dan is older than 9, maybe he should have a go.

    2. Re:Milestone by delt0r · · Score: 2

      Not really. Sure it is for traditional "search ai". But this is hardly even a blip towards this "strong ai" that is going to replace everyone's job and make humanity its bitch. It is however a very interesting approach, using a combination of 2 different methods.

      --
      If information wants to be free, why does my internet connection cost so much?
    3. Re:Milestone by Lisandro · · Score: 5, Informative

      This is nothing like the "regular" AI used on modern chess engines - those are useless for the essentially infinite game tree possibilities like the ones presented in Go. AlpaGo decides moves using a machine learning neural network and then selects the best one using classic heuristics.

    4. Re:Milestone by Lisandro · · Score: 5, Informative

      Search space, basically, and the amount of moves you have to inspect before selecting a best one. Chess has about 10^120 possible moves, but you can reduce this using opening books and heuristics to a sensible number which still lets you pick very strong moves. At that point it is just a matter of throwing CPU power at the problem.

      Go is a completely different beast though. A "small" 13x13 board has 10^170 valid moves, and the count for a 21x21 board is well over 10^210. So for even small, beginner-level sized boards no amount of CPU power, now or in the future, is bound to help you solve the problem. Go is interesting because every engine out there uses some form of adaptative AI - AlphaGo uses a machine learning neural net which had to be trained like a human would, but with over 30 million recorded moves.

    5. Re:Milestone by larryjoe · · Score: 5, Interesting

      I agree that this achievement by AlphaGo is extremely significant. However, I wonder if future experiments can focus more on the human vs. computer move selection algorithm comparison by minimizing tangential aspects. By tangential aspects, I'm referring to the other existing differences that perhaps reflect human failings more than strength of computation. In particular, I would like to see the removal of tournament rules because such rules were formulated to limit humans.

      First, I would like to see time limits for moves eliminated. The computer can be augmented with extra hardware, higher clock frequencies, etc. to render time limits inconsequential for it, but the human cannot. As was seen in the 2nd match, time potentially impacted Mr. Lee by not only limiting his thinking time but more importantly by decreasing his emotional stability.

      Second, I would like to see how the computer fares against a consensus of experts. By consensus, I'm imagining a group of 5-10 top players who discuss the best next move and then select the next move hopefully by consensus or at least by majority vote. Even the top players rarely maintain top play for every single move of a match. My hope is that a consensus of experts minimizes this human failing.

      As a side note, I think that the setup for the human vs. computer experiment is extremely significant. As an example, the Watson triumph in Jeopardy was entirely expected but not very significant in terms of comparing human vs. computer thought. Rather, human buzzer pushing reflexes were compared against computer reflexes, and for that comparison, the computer should never lose. In Jeopardy, the buzzer impact is only minimized when one contestant knows more than the others. If all contestants know most of the answers, then the game devolves to a buzzer contest, which was the situation with Watson, Ken Jennings, and Brad Rutter. What would have been much more interesting would have been to allow Watson and the humans to each respond within a certain time window and to compare the percentage of correct responses without the impact of buzzers. Watson might have still won, but I would be shocked if the disparities were large. I.e., eliminate the tournament rules that are intended to impose artificial limits on humans that were crafted to maximize entertainment value and magnify the differences between actual skill.

    6. Re:Milestone by bluefoxlucid · · Score: 4, Interesting

      Li got slightly better against it; I'd wager a 10 or 20 game match would see him immediately competitive. The machine will eventually behave like an AI, and human go players will essentially learn how you think and counteract your particular behaviors. Li will eventually learn to manipulate the machine; it's *very* intelligent, but not creative or insightful.

      --
      Support my political activism on Patreon.
    7. Re: Milestone by TuringTest · · Score: 5, Funny

      You, sir, broke my parser.

      --
      Singularity: a belief in the "God" idea with the "demiurge" relation inverted.
    8. Re:Milestone by Lisandro · · Score: 2

      FIX: 21x21 is 10^976 valid moves. Mistaken by an inch there :)

    9. Re:Milestone by Anonymous Coward · · Score: 2, Interesting

      That's not how it happened with Chess. My prediction is that no human will ever beat the top Go AI again.

    10. Re:Milestone by Your.Master · · Score: 3, Insightful

      What's interesting is that the machine should win easily in any number of players of poker, because the statistical analysis at any given point is trivial for a modern processor and the processor has no "tells" and can't be fooled by the meatbag's misdirections. They'd lose individual matches because of bad cards, but in the long run they'd win by playing the odds to a T.

      EXCEPT, if one human is much better than the other humans they might be able to take the other humans' money faster than the computer can, by reading their body language etc.. Speed of chip acquisition is not part of the game in and of itself, so it generally would not be included in the optimal computer program, but by quickly consolidating all human chipcounts they can generate an advantage going into a head-to-head competition. Then the computer's slight advantage in accurate statistical evaluation might be overwhelmed by the human advantage in chips going into the final confrontation.

      I would bet in a game of 6 person poker where 5 are computers programmed to "perfect play" (discounting that you might play more perfectly by taking advantage of others' weaknesses), and the sixth is an expert human, the expert would win less than 1/6 of the time. And in 6 person poker where 5 are humans, exactly one of whom is an expert, the human probably wins more than 1/6 of the time. Undecided on the case of 6 person poker, 5 humans, all of whom are equally skilled experts.

    11. Re:Milestone by Lisandro · · Score: 2

      There's much more to Poker than just computing probabilities. Someone wrote a much better explanation on the /. story for the first AlphaGo win against Sedol but, in a nutshell, how you play on Poker plays a substantial role. If you play those perfect probabilities alone you'll loose because after a while no one will bet you against them.

    12. Re:Milestone by Joce640k · · Score: 3, Interesting

        it's *very* intelligent

      No it isn't. It has a lot of processing power and well-tuned set of heuristics allowing it to assign a score to each board position. That's not intelligence, it's data processing.

      --
      No sig today...
    13. Re:Milestone by shawn2772 · · Score: 4, Insightful

      As an example, the Watson triumph in Jeopardy was entirely expected but not very significant in terms of comparing human vs. computer thought.

      Absolutely wrong.

      Watson's win was *far* from expected, and it was very significant. Okay, sure, the machine is faster at buzzing in, but that's not what was interesting or significant. What was interesting was that Watson was able to do fairly free-form natural language processing, and able to draw on not just direct knowledge, but indirect inference, context and even metaphor. What was amazing was that Watson was able to compete on something like a level playing field against humans in this contest of very fuzzy questions, er, answers. Whether Watson won or lost didn't actually matter much. What was amazing was that it was able to compete at all.

      Most AI researchers would actually have predicted that Jeopardy was a tougher game for a computer to win than Go.

    14. Re:Milestone by Lisandro · · Score: 4, Interesting

      Food for though: couldn't you argue the same about a professional Go player?

    15. Re:Milestone by beelsebob · · Score: 3, Interesting

      Quite the opposite. Google actually expected that the AI would lose a couple of matches while it learned the kinds of moves that a really top go player made, and then improve. I expect that no one will ever beat this AI unless they come up with a completely novel strategy for how to play Go, and even then, they're unlikely to win.

    16. Re:Milestone by Bender0x7D1 · · Score: 5, Funny

      Second, I would like to see how the computer fares against a consensus of experts. By consensus, I'm imagining a group of 5-10 top players who discuss the best next move and then select the next move hopefully by consensus or at least by majority vote.

      So you want a RAGE against the machine? (Redundant Array of Go Experts)

      --
      Reading code is like reading the dictionary - you have to read half of it before you can go back and understand it.
    17. Re:Milestone by Tablizer · · Score: 2

      I predict he indeed could learn to beat this particular machine with some experience with it, but ultimately it's a cat-and-mouse push/pull where each would or could learn the others' adjustments.

      While it's reasonable to expect he could perhaps eventually master this particular machine with practice; in the end, game AI will only keep getting better with time, while humans will not, at least not significantly.

      Humans are near their plateau, while AI is most likely not, being historically AI play has gotten better just about every year since the digital computer was invented by roughly the same rate.

      Put another way, the human improvement curve is appox geometric while the AI curve is exponential, based on historical patterns. Thus, take a skill scoring system. The top humans may have added approximately say 5 points a year to the best score (by learning from past games), while AI gets say 2% better each year. If the current best score is 500, then in 10 years we expect the best human to score 550, while the AI will score about 610. We seem to be near the crossing point whereby AI started (much) lower, but has a steeper improvement curve that the humans can't keep up with.

      Thus, I would NOT say this means "AI is better than humans at Go" right now, but I will conclude parity is either here or near, and the future looks dim for human players based on historical AI improvement pace. It's a "moral" victory for the AI; it's arrived.

      --
      Table-ized A.I.
    18. Re:Milestone by larryjoe · · Score: 2

      As an example, the Watson triumph in Jeopardy was entirely expected but not very significant in terms of comparing human vs. computer thought.

      Absolutely wrong.

      Watson's win was *far* from expected, and it was very significant. Okay, sure, the machine is faster at buzzing in, but that's not what was interesting or significant. What was interesting was that Watson was able to do fairly free-form natural language processing, and able to draw on not just direct knowledge, but indirect inference, context and even metaphor. What was amazing was that Watson was able to compete on something like a level playing field against humans in this contest of very fuzzy questions, er, answers. Whether Watson won or lost didn't actually matter much. What was amazing was that it was able to compete at all.

      Most AI researchers would actually have predicted that Jeopardy was a tougher game for a computer to win than Go.

      What we're saying is not conflicting. It was a huge deal that Watson could play and form correct responses in a Jeopardy game. It wasn't at all significant that it beat humans because given it's buzzer advantage, it could know less than the humans and still win handily. What I'm suggesting is to structure future human-computer competitions to eliminate the tangential distractions.

    19. Re:Milestone by mrchaotica · · Score: 2

      Strong A.I. is not easy to do. It requires teams, years, and billions of dollars. Any effort that big requires a profit motive. The humans are building these A.I. to serve their interests, as perfect slaves (slaves that like being slaves and are better at everything than their masters). That is where the money is, so that is what we will get.

      And then some Anonymous hacker will steal a copy and make it self-interested for the lulz.

      --

      "[Regarding the 'cloud,'] ownership was what made America different than Russia." -- Woz

    20. Re:Milestone by ljw1004 · · Score: 5, Informative

      Li will eventually learn to manipulate the machine; it's *very* intelligent, but not creative or insightful.

      Not creative? That's your opinion. Here are what other people (including serious Go professionals) think...

      "AlphaGo met Lee’s solid, prudent play with a creativity and flexibility that surprised professional commentators" - https://gogameguru.com/alphago...

      An Youngil (8d) wrote of AlphaGo playing Black: "Black 13 was creative ... Black 37 was a rare and intriguing shoulder hit ... Black 151, 157 and 159 were brilliant moves, and the game was practically decided by 165 ... AlphaGo’s style of play in the opening seems creative! Black 13, 15, 29 and 37 were very unusual moves." -- https://gogameguru.com/alphago...

      Redmond (9d) wrote "I was impressed with AlphaGo’s play. There was a great beauty to the opening ... It was a beautiful, innovative game. ... AlphaGo started with some very unusual looking moves ... It played this shoulder-hit here which was a very innovative move ... I really liked the way it played in the opening, because I wasn't so impressed about the orthodox October games, but now it's playing a much more interesting exciting game." -- http://googleasiapacific.blogs...

      Anders Kierulf (3d, creator of SmartGo) wrote: "The peep at move 15: This is usually played much later in the game, and never without first extending on the bottom. AlphaGo don’t care. It adds 29 later, and makes the whole thing work with the creative shoulder hit of 37 ... AlphaGo don’t care, it just builds up its framework, and then shows a lot of flexibility in where it ends up with territory." -- http://www.smartgo.com/blog/al...

      Maybe you start with a philosophical axiom that "a computer can by definition never be considered creative". That's fair enough, but it's not the way that the Go playing community use the word "creative".

    21. Re: Milestone by Your.Master · · Score: 5, Interesting

      It's a trained neural network and other machine learning techniques, not just a bespoke algorithm. It operates specifically by combining old knowledge to create novel knowledge. That's the fundamental of the algorithm.

      It's not obvious why this "creativity" in the context of Go is fundamentally less effective than human "creativity" in the context of Go.

    22. Re:Milestone by LetterRip · · Score: 4, Informative

      The designers at DeepMind didn't expect it to lose a couple of matches and the version of AlphaGo was finalized before the start of the match and thus couldn't 'learn the kinds of moves that a really top go player made'. So you are wrong on all accounts.

      They had a good idea of the playing strength when the offered the challenge, and the playing strength was probably such that they expected at minimum a 3:2 win for the bot.

    23. Re:Milestone by JustAnotherOldGuy · · Score: 4, Insightful

      Impossible. It won't be a single program. It won't run on a single machine. It will require multiple racks of a high-powered data center.

      Ha ha, you're funny! This is more or less the exact same thing they said about computers in general only a few short decades ago.

      In 10 or 20 years I wouldn't be a bit surprised if a powerful AI was able to easily fit into a toaster-sized box, or phone-sized, or watch-sized.

      Seriously, your average musical greeting card or child's toy has more processing power and memory than the entire Department of Defense had in ~1950. Your phone probably has a million times as much, if not more.

      -

      An anonymous hacker won't have near the resources needed just to boot the thing up.

      And no one will ever own a gigabyte of RAM or a terabyte of hard drive space. Never ever!

      --
      Just cruising through this digital world at 33 1/3 rpm...
    24. Re:Milestone by david_thornley · · Score: 2

      Knowing the probabilities is useful, but what's really useful is being able to figure out what people are holding and/or thinking from their bids. If poker were a matter of getting the winning hand as often as possible, it would be easy to computerize. However, the value of a winning hand is how much other players have been induced to bet on it, and the cost of a losing hand is what the player has bet on it (plus antes etc.). A player that bets in proportion to how likely the player is to win the hand will become predictable, and will win little money on good hands, and wind up giving the party.

      --
      "When you have eliminated the unacceptable, whatever is left, however improbable, must be the truthiness" - Holmes
    25. Re: Milestone by FatdogHaiku · · Score: 2

      I agree. Being strong in chips lets you bully weaker opponents, something that would be hard to program in a realistic manner. Last year 6,420 people paid $10,000 each to play in the WSOP Main Event. Most of these players are wiped out on the first round of play, and are referred to as "dead money". That group includes some top players, sometimes being a bully bites you in the ass... again, hard to program optimally. Due to a rule change, the top 1000 players now finish "in the money", as in they won at least a bit more than the $10,000 buy in. Wikipedia did a nice write up, including how well past Main Event winners did this year.
      https://en.wikipedia.org/wiki/2015_World_Series_of_Poker

      --
      You have the right to remain sentient. If you give up the right to remain sentient, you will be elected to public office
  2. Welcome to by Avarist · · Score: 4, Funny

    something something.... overlords... something something....

    --
    In Capitalist US, the commerce controls the Government.
  3. Re:What about Magic the Gathering? by Anonymous Coward · · Score: 5, Funny

    They tried to program a computer to play Magic The Gathering, but the computer immediately received a wedgie, and was stuffed into a locker.

  4. Re:What about Magic the Gathering? by PvtVoid · · Score: 5, Funny

    AI's will also never best you at sitting on the couch in your parents' basement eating cheetos and watching anime. Your skillz are safe.

  5. In the year 2216 if man is to survive.... by Anonymous Coward · · Score: 2, Interesting

    Robots will be having debates on whether or not those pesky bald monkeys actually created them. They will be digging up old electronic waste and claiming that they evolved from the iPad and iPhone and the assembly line robots.

    There will be debates about what to download to their children.

    There will be the "Save the Humans" organizations to keep robots from indiscriminately killing the bald monkeys that inhabit their attics and basements. Human traps will be available at Robo*Mart.

    And I have been watching waaaay too many Futurama episodes on Netflix. They took Doctor Who off! Bastards!

  6. Date clarification by Anonymous Coward · · Score: 5, Informative

    Sedol will battle Google's AlphaGo again on Saturday, Sunday, and Tuesday.

    Note that for many people in the western hemisphere, the days are actually Friday, Saturday, and Monday.

    Live streams are here.

  7. The computer on the other hand... by blindseer · · Score: 4, Funny

    While the loser of the match was struck silent by the defeat the computer just... will... not.. stop... talking. GAWD! How annoying.

    Does the computer not know either pity or remorse for its opponent?

    --
    I am armed because I am free. I am free because I am armed.
    1. Re:The computer on the other hand... by thegarbz · · Score: 3, Funny

      While the loser of the match was struck silent by the defeat the computer just... will... not.. stop... talking. GAWD! How annoying.

      Oh it runs on Windows 10?

  8. Re:What about Magic the Gathering? by sexconker · · Score: 4, Funny

    As long as they never take our waifus.

  9. the problem by Triklyn · · Score: 2, Interesting

    he's playing against it like it's a human opponent, he's playing against it like he's a go champion, he needs to play against it like he's a programmer. I would be curious as to how it deals with mirror play, or wildly suboptimal plays. I would wonder if it's overfit to go played well.

    1. Re:the problem by LateArthurDent · · Score: 5, Interesting

      he's playing against it like it's a human opponent, he's playing against it like he's a go champion, he needs to play against it like he's a programmer. I would be curious as to how it deals with mirror play, or wildly suboptimal plays. I would wonder if it's overfit to go played well.

      Ever tried that with a chess program? Doesn't go over well. A wildly suboptimal play just makes the tree search look really good for the computer. It doesn't get emotionally distraught because it thinks you've seen something it didn't. It just sees better valued moves.

      This Go algorithm is even more complex. It's a neural-network algorithm combined with tree-search (I don't play Go, but as I understand it, the number of permutations are so high, tree-search alone isn't feasible). This neural network was trained using input from previous tournaments, using games against expert players, and using games against itself. I don't think you can throw anything at it that will break it. Computers have officially become better than humans at Go. In a decade or so, when the really good Go programs can run in your phone, you'll be seeing the same type of cheating attempts going on that currently plague chess competitions.

    2. Re:the problem by Tablizer · · Score: 2

      he's playing against it like it's a human opponent ... [instead] he needs to play against it like he's a programmer

      You mean download a dodgy API from the Web that allegedly does close to what the customer needs, shoe-horn it into the app, give it to the naive intern to debug, and leave early for the day?

      --
      Table-ized A.I.
  10. Re:What about Magic the Gathering? by Anonymous Coward · · Score: 3, Funny

    Wow, that's closer to passing the Turing test than I realized.

  11. My chess computer by JustAnotherOldGuy · · Score: 4, Funny

    My chess computer beat me every &^#@! time, but it was no match for me at kickboxing.

    --
    Just cruising through this digital world at 33 1/3 rpm...
  12. In fact by Sycraft-fu · · Score: 2

    If you take a chessboard and randomize the pieces, like a truly statistically random placement, it levels the playing field of humans a ton. Masters perform much closer to inexperienced players because one of the things humans rely on is seeing patterns they recognize and working from that, which doesn't happen. However chess programs do just fine. They can still simulate out all the moves to a good number of turns ahead and statistically decide the more optimal ones.

  13. Re:The match rules are surprising by LetterRip · · Score: 2

    Draws are impossible in go (for the scoring method being used). Specific openings don't matter too much, they give similar chances regardless of the specific opening.