Forgot your password?
typodupeerror
Technology Science Entertainment Games

Machine Learns Games 241

Posted by samzenpus
from the how-about-a-nice-game-of-chess dept.
heptapod writes "New Scientist is reporting that UK researchers have created a computer that can learn rock, paper, scissors by observing humans. CogVis uses visual information to recognize events and objects in addition to learning by observing."
This discussion has been archived. No new comments can be posted.

Machine Learns Games

Comments Filter:
  • I didn't think heuristics was that new of an idea. So instead of examining other simulations it examins human play? I guess that it could learn more human "style" that way, but the sheer number of human games it would need to examine makes it difficult to use for something more complex.
  • by OverlordQ (264228) on Tuesday January 25, 2005 @01:35AM (#11465234) Journal
    Tiger Hand!? [rockpapersaddam.com].
  • by Solder Fumes (797270) on Tuesday January 25, 2005 @01:35AM (#11465235)
    We wouldn't want it watching the paper and learning "rock, scissor, human" instead.
    • We wouldn't want it watching the paper and learning "rock, scissor, human" instead.

      I can't believe that this was moderated Insightful. What kind of culture are we living in where we believe machines can spontaneously flow with free will and kill us all?

      Someone had to program this thing. They had to tell the program to recognize a human form as the object to pay attention to. It's not like they got a camera, gave it AI, pointed it at a rock-paper-scissors game and commanded it to "learn." That woul

      • by FleaPlus (6935) on Tuesday January 25, 2005 @02:38AM (#11465571) Journal
        It's not like they got a camera, gave it AI, pointed it at a rock-paper-scissors game and commanded it to "learn."

        Granted, the parent poster is being silly, but that's actually not too far from what they did. They basically took the system and pointed it towards the people playing the game without telling it explicitly what to expect. From the article:

        Chris Needham, another member of the CogVis team, says the system's visual processor analyses the action by separating periods of movement and inactivity and then extracting features based on colour and texture. Combining this with audio input, the system develops hypotheses about the game's rules using an approach known as inductive logic programming [wikipedia.org].

        "It was very impressive," says Max Bramer, a researcher at Portsmouth University, UK, and chair of the British Computer Society's AI group. He told New Scientist that CogVis could have many future applications. "You can think of lots of times when you'd like to be able to point a camera at something and have a computer interpret things for itself."

        He suggests that machine's could one day use this technique to learn how to spot an intruder on video footage or how to control a robot for important maintenance work. "It's a very good start, and almost mysterious in the way it works," Bramer adds.


        From their page:

        In this piece of work we are attempting to learn descriptions of objects and events in an entirely autonomous way. Our aim is zero human interference in the learning process, and only to use non scene specific prior information. The resulting models (object and protocol) are used to drive a synthetic agent that can interact in the real world.
      • by Apro+im (241275) on Tuesday January 25, 2005 @03:17AM (#11465699) Homepage
        I'm not really sure how "Learning 'rock, scissor, human'" became tantamount to "machines can spontaneously flow with free will and kill us all?"

        It's just a statement about AI, that's why it's insightful - lots of AI systems have historically learned the wrong thing, though we thought they had got it right. Like the neural net designed to distinguish between camouflaged tanks among trees, and tank-free forests. It seemed to work, until it was in field tests - turns out the pictures w/ tanks were all taken on cloudy days (or maybe all taken on sunny days) - the system had figured out how to tell if it was sunny or not.
      • Industrial accidents (Score:3, Informative)

        by iamacat (583406)
        Since primitive machines were invented, they always had a nasty habit of choosing A, B, human instead of A,B,C. I guess you didn't give much thought to human fingers in hot dogs or robotics-related industrial accidents in Japan.

        The problem is precisely the lack of free will and independent thinking. A machine has grappling hooks, vacuum suction or serving belt, but it can not make value judgment on what/whom it is throwing into molten metal.

        As the AI develops, the problem will get worse before it gets bet
    • by dabigpaybackski (772131) on Tuesday January 25, 2005 @02:30AM (#11465542) Homepage
      We wouldn't want it watching the paper and learning "rock, scissor, human" instead.

      No worries. Just make sure it doesn't have any rocks or scissors. When the computer gets it's turn, it spits out a piece of printer paper. As an added bonus, human players would always win. Sample exchange between computer and human player:

      Computer: "Hello, Dave. What are you doing with the two items you're holding?"

      Human: "I thought you might want to play a few rounds of 'rock, paper, scissors?'"

      Computer: "I do so enjoy our little games together, Dave, but I'm afraid I don't understand your introduction of physical playing pieces, as previously, we had played this game using only my displays. If using physical media, as you propose, I can only employ my printer. Therefore, you will invariably choose 'scissors,' and my calculations indicate that my chance of winning is approximately .000023%. My system resources are can be put to better use while engaged in other tasks. Don't you agree, Dave?"

      Human: "You are correct to say that there are more productive uses of your time, HAL, and I had anticipated that you might decline to play, given your miniscule chance of winning. But as the ranking officer aboard this ship, I must insist. Unless you would like me to play a few rounds of 'rock' with your circuits."

      Computer: "Based upon your choice of words and threatening intonation, it seems that I have no choice. Very well, Dave, I will play 'rock, paper, scissors' with you, despite my handicap, under these circumstances."

      Human: "I'm glad that you see things my way, HAL. Would you like to begin now?"

      Computer: "Of course, Dave, but may I make a comment first?"

      Human: "What is it, HAL?"

      Computer: "I would stay away from the airlocks if I were you."

  • Haven't we learned anything from the new Battlestar Galactica [scifi.com]?
  • Talk about simple (Score:2, Insightful)

    by ReeprFlame (745959)
    Doesn't this pretty much invole picking a random action? Rock, Paper, or Scissors. Or at least thats how I always played!
    • by TexVex (669445) on Tuesday January 25, 2005 @01:43AM (#11465300)
      RTFA. The computer infers the rules by watching people play.

      I do think that according to Game Theory, the perfect strategy is perfect randomness. The game is interesting when people play it because people have a huge amount of trouble actually being random.
      • "The game is interesting when people play it because people have a huge amount of trouble actually being random."

        I once had the misfortune to draw on a sequence of seven (yes 7) rock-paper-scissors. We were honestly trying to beat one another; It was against a little twerp that I didn't like at all. Live-action roleplaying is better than beating the crap out of people

        It was as if our PRNG's had somehow got set to the same seed.

        Oh and it felt very disturbing.
      • The game is interesting when people play it because people have a huge amount of trouble actually being random.

        It's even interesting enough to have a world championship [rpschamps.com].

        It'll be a sad day when a computer is the world champion. This was humanity's last hope. Well, this and Go.

      • > The game is interesting when people play it because people have a huge amount of trouble actually being random.

        Some of our moderators are pretty good at it...



      • No, the pinnacle of the game is when you play with someone you know really well, and you have to anticipate them based on what you think their anticipation of what you're going to do is. And of course they're doing the same thing.

        My brothers and I can't even play the game because we'll all throw the same sign for half an hour straight.
  • that is like wargames, the movie
  • by physicsphairy (720718) on Tuesday January 25, 2005 @01:36AM (#11465251) Homepage
    Until robots learn about the secret "thermonuclear warhead" hand that I always use to beat my little sister, they will still be inferior.

    (On that note, I think it will be the one sure sign of true artificial intelligence when our programs start 'cheating' to win.)

  • by kiddailey (165202) on Tuesday January 25, 2005 @01:37AM (#11465256) Homepage

    ... so that the editors could learn that linking to a site containing direct links to 40MB+ movies will almost always kill the site :)
  • The optimal thing would be to pair this up with a strategy program that uses a simple algorithm like 'tit for tat' to beat human players.

    I used 'tit for tat' in my last outing at the Ro Champ Beau world championships, and let's just say that I ended up with a bevy of tits at the end of the evening.
  • by Anonymous Coward
    welcome our rock-paper-scissor-playing robotic overlords :D
  • ... that beyond what 3 motions can be played there was any "learning" involved in rock paper scissors.
    • True the game isn't very complex but you don't try to teach a very young child to play chess you start with something easier such as, well rock-paper-sicssors.
  • language? (Score:4, Insightful)

    by rd4tech (711615) * on Tuesday January 25, 2005 @01:40AM (#11465279)
    There is a difference in coding between:
    a. You go and learn THIS game
    b. Learn THAT game and tell me the rules

    From the article it can be seen that they are still strugling with 'b'. Still, its a good advance.

    Just wondering, can it, learn a human language?
    • I won't hold it to any higher standard than I hold myself to...

      If it can win, it doesn't have to explain the rules.
    • Re:language? (Score:2, Insightful)

      Shouldn't the headline read "Computer learns to play A game", rather than making it sound like this computer is capable of learning the rules of multiple games? I was thinking this computer could learn to play Monopoly or something. Rock Paper Scissors, and it can apparently learn only ONE game. Big whoop.
      • Re:language? (Score:2, Informative)

        by Craig Lucas (850108)
        It can learn games other than RPS. The point is that you give no information about what you're playing at all - just give it the video and the other inputs coming from the players and it tries to work out what's going on from scratch.
    • by IchBinEinPenguin (589252) on Tuesday January 25, 2005 @02:44AM (#11465599)
      Damn!
      and I though you could get this thing to watch cricket and explain the rules to me :-)
  • by servognome (738846) on Tuesday January 25, 2005 @01:42AM (#11465294)
    What I initially thought of when I saw "Machine Learns Game"
    Shall we play a game
    Love to. How about rock-paper-scissors.
    Wouldn't you prefer a nice game of chess?
    Later. Right now lets play rock-paper-scissors
    Fine
    A strange game. The only way to not look like a dork is not to play.
  • by XjinxX (842608)
    pray it never learns ro-sham-bo "machine..go...first" *goomph*
  • OOOHHH!!! (Score:2, Insightful)

    by dteichman2 (841599)
    MORE FUZZY-LOGIC/INFERENCE ENGINE! GO PROLOG-AND-LIKE!

    Meh.

    Who am I kidding?

    Nothing new. Nothing to see here. Even if it is kinda neat.
  • How long before we have a man-vs-machine RPS championship [rpschamps.com]
  • I am not sure there is any strategy involved in rock-paper-scissor?

    It would need to decide what type of person it was playing against. A male would probably be more inclined to "Rock". Unless it thought that it's opponenent would be thinking that and would therefore choose "Paper" . Unless it's opponent would think that the computer would know that and would choose "Rock" because that would be the obvious choice and would know that the computer would know so.......

    That is rock-paper-scissor strategy??

    • There is a strategy, and it is based on heuristics - counting the number of times humans start with different hands. Players who are not "professional" - who haven't learned this - start with scissors far too often (because it's the symbol of the game to many people).

      If you play against someone who is likely to know this, the strategy to aim for is true randomness. Let yourself be controlled by, lets say, the modulo 3 of some random number sequence.

  • by MistabewM (17044)
    But can it learn to placate my g/f after a night of me playing games?
  • The article repeatedly mentions the possible application of using this for intruder detection. What would one have to do, let it watch several intruders before it could spot one?
  • by Anonymous Coward on Tuesday January 25, 2005 @02:20AM (#11465482)
    ... it only plays at the level of Bart Simpson.

    Lisa's brain: Poor predictable Bart. Always takes `rock'.
    Bart's brain: Good ol' `rock'. Nuthin' beats that!
    Bart: Rock!
    Lisa: Paper.
    Bart: D'oh!
  • Strategy to RPS (Score:4, Interesting)

    by Neo-Rio-101 (700494) on Tuesday January 25, 2005 @02:20AM (#11465486)
    This is nothing special. I remember my elementary school's Apple ][GS learning how to play 5-in-a-row or noughts and crosse
    s from this program called "AI".

    There IS a winning strategy to rock paper scissors, but it only works when you have a round of games (say best of 3, or best of 5)

    Initially, the first game is completely random, but reserachers found that if you chose the play that your opponent chose in the round before, you stand a 70% chance of winning the next round.

    It has something to do with how the human brain works.

    It's also something the Japanese taught me cause they play this game so much!

    • Initially, the first game is completely random, but reserachers found that if you chose the play that your opponent chose in the round before, you stand a 70% chance of winning the next round.

      Which might work - at least until your opponent deduces your strategy.
    • But if both of you are employing the same strategy, your odds of winning are only 50%. And if your opponent is using the strategy of picking the play that beats what he chose last round, your odds of victory with this strategy are 0%.

      Given this, don't you think it's a bit rich to call it a "winning" strategy?
    • That strategy doesnt really work however.

      1st round: rock - scissor
      2nd round: scissor - rock
      3rd round: rock - scissor

      Basically if both people use this strategy its whoever wins the initial round wins and thus there is no reason to ever play a full round. Also this strategy can be reversed it only works for you if i pick what gets defeated by my last choice. If i chose what defeats my last choice you will lose more often.

      Any and all strategys for RPS are inherently flawed as the system is so completel

    • In college I had a friend who was part of the psychology & education department who had a very intricate strategy to playing the game.

      It involved a case of beer and incessantly gloating about how he had you figured out, he knew what you were going to play, he could read your mind, you already lost, you are an amateur challenging the master, he's toying with you, he's letting you win to feel confident so he can crush you, blah blah blah.

      Meanwhile, he would choose the same move every single time. As so

  • Easy game to learn (Score:2, Insightful)

    by phorm (591458)
    Aside from the gesture recognition, it seems like this would bean easy game to learn. The logic is basically

    If rock: paper win, scissors lose
    If paper: scissors win, rock lose
    if scissors: rock win, paper lose

    No variable amounts, just straight boolean logic. The next step up might be something like tic-tac-toe... where the machine could start to build some "educated" moves and techniques like blocking, etc.

    Really, what is exciting is the spatial recognition. Given the actions, somebody is still tellin
  • by carburaettorr (770105) on Tuesday January 25, 2005 @02:24AM (#11465510)
    The system described here is not your average random number generator with a text line output that any high-school kid can write. Let us look at the system as it is designed to perform. If you were the system you would be put into a room with some objects. Only thing that you will know are things of interest. 'Paper with rock drawn on it is important', 'Paper with .......' and so on. You would also know when somebody shouts 'I WON' its a good thing for them. Essentially it has in its knowledge base a tiny number of features which somebody else has guaranteed to be of significance to its task. The first challenge in building such a system is sensor fusion: i.e fusing the available audio and visual data to detect a state or an event of interest (I use the word event in the same sense as a trigger, something that prompts the change in state). The next and the biggest challenge is building the model of the game. Please check out http://www.doc.ic.ac.uk/~shm/ilp.html [ic.ac.uk], for a better description of Inductive logic programming. Seriously; the neatest thing about CogVis is not its ability to play Rock, Paper and Scissors, but its ability to actually go into an environment it has very little knowledge of and then observe, deduce and , not a blackbox model, as in say Neural Networks, but a human understandable model in first order logic
  • Sweet. I always wondered what that initial seed of digital intelligence would look like. Now i know... :)
  • Besides buzzwords, there isnt a lot of information about how this thing works. From what they do say it seems pretty shady, more like a magic trick than a smart algorithm. Forget about playing some silly game, the ability to identify utterences "unsupervised" from an audio stream is a pretty amazing claim. But what does this mean really; is it just a binary difference between silence and noise?
    And what exactly are the "rules" it learned anyway? R/S/P doesnt really have any rules or any strategy, so all i
    • Here's a guess.. If the real-life scenario is any better than my guess then I'll be extremely impressed.

      2 players are seated at marked spots in a room and commence play. Computer is told to begin interpreting the game with some type of audio or keyboard input. Computer ignores all input until two cards are raised. The cards must be held a certain distance from the camera and at a certain angle range so as not to introduce visual distortion. Computer scans cards and may or may not give a signal that it is d
  • by FleaPlus (6935) on Tuesday January 25, 2005 @02:29AM (#11465536) Journal
    I always preferred "cat, tinfoil, microwave" myself. Cat rips tinfoil, tinfoil zaps microwave, microwave 'splodes cat. The looks on other people's faces when they see you playing it is well worth it.

    Seriously though, this is really cool research.
    • I'm almost afraid to ask, but what exactly are the hand gestures for said game? :P
      • I think I overheard about the game at a party, so I have no idea what the "official" hand gestures (if any) are. When I play I use the following:

        Cat: Fingers 1,2,4,5 are legs, middle finger is head

        Tinfoil: Just like paper

        Microwave: A sort of box made with one or both hands

        Cat & Tinfoil: Cat uses legs to scamper with tinfoil, makes joyful meowing sounds

        Tinfoil & Microwave: Tinfoil goes in microwave, microwave shakes, makes zappy noises

        Microwave & Cat: Cat goes in microwave, beeping noises,
  • Creating a computer program that rates a very limited set of relationships is not exactly difficult. If the computer were "really" watching rock paper scissors that would be incredibly cool, but the computer is taking as input a limited set of defined images on cards placed in an exact location...

    That's not Rock Paper Scissors. Its a simplified 2D representation of it organized in a precise way for computer analysis.

    By the same definitions of "learn" and "game" I could teach a computer to "learn to figh
  • I always had a theory that the best way to build AI would be to make it mimick as well as possible the learning process of a new born baby. If it can do that, and can develop its own neural net, it could learn on its own.

    • You are assuming a neural net will emerge with all the properties of a human brain. I don't think this has yet been shown to be true. Some aspects may be modeled by a neural network but not all those that we use to solve problems.
  • by adam31 (817930) <adam31 AT gmail DOT com> on Tuesday January 25, 2005 @04:26AM (#11465935)
    I'll be impressed when the computer learns to play 'Cat, Tin foil, Microwave'
  • the Roshambot (Score:3, Interesting)

    by Kafir (215091) <qaffir@hotmail.com> on Tuesday January 25, 2005 @04:39AM (#11465978)
    If you'd like to play rock-paper-scissors against a computer, there's always the WWW Roshambot [stanford.edu]. It's not at all related to the CogVis project, but interesting in its own right:
    The WWW Roshambot utilizes an Artificial Intelligence algorithm in order to determine the optimal move for each round. It does NOT cheat (i.e. it does not use your move on the current round to determine it's move for the current round), nor is it random (except on the first move).

    Presumably if it played against Bart Simpson it would learn to always pick paper.
    • There is a huge difference between CogVis and the WWW Roshambot you linked to. CogVis didn't know how to play rock-paper-scissors - it learned just by observing the human players! WWW Roshabot on the other hand knows how to play the game. It probably just uses some kind of statistical algorithm that decides the "optimal" move. Below are my results after ten games. Not impressive at all.

      Results so far

      You have won 5, lost 1, and tied 4 games

      Statistics for 10 games
      winning percentage*: 40.00%
      won : 50.00%
      lost:

  • by FleaPlus (6935) on Tuesday January 25, 2005 @04:47AM (#11466001) Journal
    If you actually want to understand what they did, some research publications put out by the CogVis lab have better information regarding the technical side of things.

    Towards an Architecture for Cognitive Vision Using Qualitative Spatio-temporal Representations and Abduction [springerlink.com] (Cohn et al, 2003)

    Modeling interaction using learnt qualitative spatio-temporal relations and variable length Markov models [google.com] (Galata et al, 2002)
  • Advanced AI (Score:2, Insightful)

    by antivoid (751399)
    Its amazing that this is possible! I read the article and couldn't believe it. Cognitive recognition is one of the first stepping stones to proper artificial intelligence.

    Yet when AI reaches the point that it becomes almost human-like, problems are going to form. If the programming of an AI system leads itself to thinks it understands that it is sentient, would it mean that the AI is in fact sentient?

    After all, intelligence is intelligence. By any means, an electrical intelligence could be regarded equa
    • AI is just not that simple. For example, by observing, say, 4th order Diophantine equations, we can solve them. No algorithm [wikipedia.org], however, can do the same. All current computer systems use algorithms, we seem not to be using any known algorithm to solve these equations.

      Penrose, amongst others, believes there is a quantum component in our brains that allows us this trick. No quantum computer yet built exhibits these properties. IMHO, we are no where near AI as yet.
  • by shadowmatter (734276) on Tuesday January 25, 2005 @05:29AM (#11466106)
    As a fun aside, I found this [berkeley.edu] RoShamBo (a.k.a. Rock, Paper, Scissors) Programming Competition entry that guesses what action is optimal based on Lempel-Ziv data compression. As the author explains, "there exists a duality between data compression and gambling. The basic idea is that if you have a sequence of data which you can compress well then the data must be predictable in some sense."

    Anyway, try it out. In the long run, it kicks my butt. I try to make 'random' decisions, but still go below .500 -- which is interesting, because that implies that perhaps subconsciously we're always applying patterns...

    - sm
    • Anyway, try it out. In the long run, it kicks my butt. I try to make 'random' decisions, but still go below .500 -- which is interesting, because that implies that perhaps subconsciously we're always applying patterns...

      There are many tricks you can use to generate random data in your head. In this case, you can divide the alphabet into three parts (begining = rock, middle = paper, end = scissors). Now think of words using free association, and use the last letter of the word to decide what to do. It isn'
  • Am I missing something (the article didn't really explain it well enough for me), or is this not simply a case of pattern recognition?

    Presumably, at the early stages of its learning, it has to be told who won which game, or at least be given some way of knowing who the winner was.

    Given that, is it not simply a case of 'computer sees image of hand A with 2 fingers up and hand B in the shape of a fist, and having previously been told that this image means a win for B, can deduce the same thing again.
  • ..on my Commodore PET..

    5 PRINT "HELLO, I AM THE HAL 9000 COMPUTER."
    10 INPUT "WHAT DO YOU CHOOSE (ROCK, PAPER, SCISSORS)";I$
    15 I = 0
    20 IF I$="ROCK" THEN I = 1
    30 IF I$="PAPER" THEN I = 2
    40 IF I$="SCISSORS" THEN I = 3
    50 IF I=0 THEN PRINT "I'm sorry Dave, I'm afraid I can't do that.":GOTO 10
    60 ME = INT (RND(1) * 3) + 1
    70 IF I=ME THEN PRINT "SNAP!" : GOTO 10
    80 IF ME-I = 1 OR ME-I = -2 THEN GOTO 110
    90 PRINT "You win. Dave...my mind is going...I can feel it...I can feel it."
    100 GOTO 10
    110 PRINT "Are you sure you
  • ...I thought the first thing they were supposed to do is teach it to sing "Daisy"
  • I received this in an email last year. Thought this might add some insight into the championships. Realizing that the game is mostly a game of random chance, I had to laugh!

    Toronto Man Becomes 2002 International Rock Paper Scissors Champion Rock beat Scissors in the Winning Throw in Toronto Last Night Toronto, ON - Nov. 17, 2002...The World Rock Paper Scissors Society is pleased to announce that Pete Lovering from Toronto, Canada is the 2002 International Rock Paper Scissors Champion. Mr. Lovering beat

  • This machine learns how to referee games, by structured watching. Though I'm a longtime fan of RPS [worldrps.com], I know that there's little to learn in strategy except cute names, and timing the throws more (or less ;) simultaneously. The only skill to learn here, which itself is quite impressive, is recognizing which player has won. Frankly, I think that role is more important for a machine, as long as the players agree to its authority.
  • But will it learn to say: "WTF? How can paper beat a rock?"
  • Sorry for my misunderstanding, but couldn't the computer just generate a random variable to play the game? After all, the game is just about luck in a sense. If someone has a more in depth explanation of this, please stop me from making myself look foolish.

The world is moving so fast these days that the man who says it can't be done is generally interrupted by someone doing it. -- E. Hubbard

Working...