Forgot your password?
typodupeerror
Robotics Entertainment Games Hardware

When Robots Play Games 184

Posted by timothy
from the everyone-loses dept.
Roland Piquepaille writes "If the theory of evolution has worked well for us -- even if this is arguable these days -- why not apply it to mobile robots?, asks Technology Research News. Several U.S. researchers just did that and trained neural networks to play the Capture the flag game. Once the neural networks were good enough at the game, they transferred them to the robots' onboard computers. These teams of mobile robots, named EvBots (for Evolution Robots), were then also able to play the game successfully. This method could be used to build environment-aware autonomous robots able to clear a minefield or find heat sources in a collapsed building within 3 to 6 years. But the researchers want to build controllers for robots that adapt to completely unknown environments. And this will not happen before 10 or maybe 50 years. You'll find more details and references in this overview, including a picture of EvBots trying to find their way during a game." Read on for a similar robot competition held this weekend in France.

saunabad writes "The annual Eurobot autonomous robot contest for amateurs is held this weekend on La Férte-Bernard, France. This year's theme is 'coconut rugby,' and the robots are collecting small stress balls from the field and carrying them to the opponent's end, or shooting them in the rugby goal, while avoiding the randomly placed obstacles at the same time. Each team has a one main robot and an optional small assisting robot."

This discussion has been archived. No new comments can be posted.

When Robots Play Games

Comments Filter:
  • Aimbots... (Score:3, Funny)

    by criordan (733016) on Sunday May 23, 2004 @03:26PM (#9231980) Homepage Journal
    Aimbots have been around in CS for years. Is this really news?
  • Maybe.... (Score:2, Offtopic)

    by TastyWords (640141)
    we could see some of this in the next "Austin Powers" movie and they can involve the FemBots?
  • by Anonymous Coward on Sunday May 23, 2004 @03:33PM (#9232027)
    You know all they're going to do is run into each other and explode.

  • First-aid (Score:5, Funny)

    by Scrameustache (459504) on Sunday May 23, 2004 @03:34PM (#9232034) Homepage Journal
    find heat sources in a collapsed building within 3 to 6 years.

    Yeah, I think the body will be cold by then...
  • Real Dolls [realdolls.com] and QRIO [sony.net] for me to have any vested interest :-)


    Kinda like AI.. only replace Jude Law [imdb.com] and give me Rebecca [imdb.com]

    *sigh*... how great the world would be.
  • by brxndxn (461473) on Sunday May 23, 2004 @03:38PM (#9232051)
    until they master jumping around while strafing and shooting..

    I am not impressed until I see one jump+crouch and scream 'I pwn j00!'

  • Arguable? (Score:3, Insightful)

    by Squidbait (716932) on Sunday May 23, 2004 @03:40PM (#9232065)
    If the theory of evolution has worked well for us -- even if this is arguable these days

    Do I detect the scent of an evolution denier? And it is interesting that you implicitly question the validity of a theory even as you cite an example of its successful application.
    • Re:Arguable? (Score:2, Insightful)

      by Anonymous Coward
      It doesn't help evolution's status as a scientific theory that every time someone mentions its problems and/or shortcomings, they are subjected to intense social pressure to stop talking about it and conform.

      The current group of theories that make up evolutionary theory as a whole (from paleontology, biology, molecular biology, etc.) DO have some problems serious to warrant real discussion and investigation. But instead of recognizing this and just considering that there are parts of evolution that we don'
      • Re:Arguable? (Score:2, Insightful)

        by Squidbait (716932)
        Fair enough. I would never be one to squash legitimate debate about any issue, and I would encourage reasonable criticism of any scientific theory. But given the battle, particularly in the US, between believers in evolution and religious folk of many kinds (who are in the majority), nine times out of ten when someone makes off the cuff remarks questioning the validity of evolution, they are not coming from a scientific standpoint. If the orginal post was meant as a joke, then I'll shut up, although agai
        • And what of those religious folks who have genuine scientific objections, and who also would not accept darwinism even if they lost their faith? Are they to be ignored too? Or is it completely unreasonable for one to outline the fundamental problems with the darwinist model?
          • Re:Arguable? (Score:3, Interesting)

            by Quelain (256623)
            "And what of those religious folks who have genuine scientific objections, and who also would not accept darwinism even if they lost their faith? Are they to be ignored too?"

            Anyone who has credible evidence will not be ignored, no matter which cult they do or don't belong to. If they don't have credible evidence to back up their claims, they will be ignored (or laughed at). It's as simple as that.

            "Or is it completely unreasonable for one to outline the fundamental problems with the darwinist model?"

            P
            • Anyone who has credible evidence will not be ignored, no matter which cult they do or don't belong to. If they don't have credible evidence to back up their claims, they will be ignored (or laughed at). It's as simple as that.

              My experience is quite different. People will ignore me, and mock me - and then after I persist and point out their childishness and present my arguments, they apologise and get into more serious debate. The mockery comes first, and after that is weathered out, then comes the liste

              • So you have credible evidence then? Why aren't you rushing off to your nearest university biology department?

                Your statement:
                "Or is it completely unreasonable for one to outline the fundamental problems with the darwinist model?"

                I want to hear your claims because you are implying that there actually are fundamental problems.

                Should I get modded down for pointing out the many serious problems with Linux?

                Have you stopped beating your wife?

                Come on, yes or no answers please!

                You could have outlined one of
                • So you have credible evidence then? Why aren't you rushing off to your nearest university biology department?

                  For the same reason the atheist does not run into the church to speak with the priest with his evidence. He doesn't expect they will listen, because there's more than just evidence involved in the person's conversion.

                  I want to hear your claims because you are implying that there actually are fundamental problems.

                  You want to hear my claims? It's nice of you to be interested, but it's still noth

                  • He doesn't expect they will listen, because there's more than just evidence involved in the person's conversion.

                    When it comes to science though, evidence is the *only* thing which matters. If you have some valid evidence and don't think they will listen, I'd be more than happy to hand it in for you, a Nobel prize would be way cool.

                    Would you say that you have more than "just evidence" involved in your 'conversion' to creationism?

                    "Darwin on Trial" by Phillip E. Johnson.

                    Johnson is a lawyer, and argues
                    • Would you say that you have more than "just evidence" involved in your 'conversion' to creationism?

                      I am against Darwinism (see below) whether or not I ascribe to creationism.

                      Johnson is a lawyer, and argues like one. Science is not a courtroom, and even if it was, he still has no "Exhibit A" to submit to the jury, nor even an expert witness.

                      Please take the time to read this critique of "Darwin on Trial".

                      It may not surprise you to hear that I've already seen that critique. Have you read Johnson's bo

                    • I am against Darwinism (see below) whether or not I ascribe to creationism.

                      So you are using the term 'darwinist' to describe someone who feels that the theory of evolution provides an adequate explanation for the history of life on earth, right?

                      It may not surprise you to hear that I've already seen that critique.

                      Have you read it though? There are many others if you don't like that one.

                      Have you read Johnson's book? The point of his book is that darwinism, by and large, simply isn't about science. The
                    • So you are using the term 'darwinist' to describe someone who feels that the theory of evolution provides an adequate explanation for the history of life on earth, right?

                      I am using the term 'darwinism' to be a catch all phrase for, among other things:
                      1. All life shares a common ancestor
                      2. Slow, gradual change rather than punctuated equilibrium (therefore a rejection of Gould's 'Hopeful Monster')
                      3. Naturalism (as opposed to scientific theories that are not naturalistic)

                      Possibly other things I'm for

                    • This may be THE FUNNIEST thing I have ever read on Slashdot.

                      Congratulations on the 1 dozen + post troll, may we all live long enough to again view such majesty.
                    • Nice list, but completely misses the point. Naturalism:
                      "(philosophy) the doctrine that the world can be understood in scientific terms without recourse to spiritual or supernatural explanations"

                      So, even for someone who does not hold to the naturalistic philosophy would still agree that all those theories hold scientific explanations without recourse to the supernatural. The difference is that he admits:
                      a. That not everything can be understood through science (indeed, some questions can only be answered
                    • I am using the term 'darwinism' to be a catch all phrase for, among other things:
                      1. All life shares a common ancestor
                      2. Slow, gradual change rather than punctuated equilibrium (therefore a rejection of Gould's 'Hopeful Monster')
                      3. Naturalism (as opposed to scientific theories that are not naturalistic)

                      Possibly other things I'm forgetting.

                      So, you can see fundamental flaws in all that then? If you have devised this concept, it's your problem then isn't it?

                      I'm currently writing a response to this artic

                    • So, you can see fundamental flaws in all that then? If you have devised this concept, it's your problem then isn't it?

                      I don't understand...those things I listed are precisely what those who previously were called "evolutionists" taught. Besides, I haven't made up the term darwinism [m-w.com]. You may also find the definition of neo-darwinism informative.

                      Yes, I will, but I'd much rather hear what argument from the book you find most convincing.

                      That assumes I found one particular argument more convincing than a

                    • Oh, side note - my email may be having troubles, so can you post a quick note here please to let me know when you've sent it? That way I'll know if it's worked or not.
                    • Imagine a scenario of one who claims to be able to tell the location of hidden objects through psychic powers. It can be tested in this manner: Hide an object from the one claiming psychic powers. Remove all possibilities of them having known where you hid it.
                      If they are able to tell you where the object is hidden, then a supernatural answer may be required after all natural explanations are elimintated.

                      The supernatural and spiritual are not beyond science's potential to test.

                      None of these are scient

              • Anyone who has credible evidence will not be ignored, no matter which cult they do or don't belong to. If they don't have credible evidence to back up their claims, they will be ignored (or laughed at). It's as simple as that.

                My experience is quite different. People will ignore me, and mock me

                Tyreth, you do not have credible evidence.
                That is why you are ignored and laughed at.

                I tend to restrict my debates on slashdot

                lol! You troll every single thread that mentions the word "evolution"!

                Though you may

                • lol! You troll every single thread that mentions the word "evolution"!

                  Yes I do. But you have difficulty distinguishing between "reminding people of the debate through introduction" and "presenting a full length argument against evolution and defending it from all attacks". The latter is a nightmare - ever tried to answer 5 posts for every single post you make? So I do the former. Which is what I originally said. Which gives you no reason to laugh.

                  You claimed genuine scientific objections, then you

      • If you want, you can still use Geocentric models of the solar system to calculate astronomical events. Heliocentric models are far simpler, but there are no priveldged frames of reference in Physics, so you can do it.
      • "It doesn't help evolution's status as a scientific theory that every time someone mentions its problems and/or shortcomings, they are subjected to intense social pressure to stop talking about it and conform."

        Bullshit. Science doesn't care a bit about the ramblings of a bunch of whacko creationists, or how others like myself respond to it, it's not a bloody democracy you know.

        If there is any pressure on you, it's basically a case of "stop talking shit". You can believe whatever weird shit you like, but
      • the "problems with evolutionary theory" consist of debates as to the details of how evolution works. However, few in the scientific community have any serious doubts that life evolves over time, or that the species we see on earth today evolved from other species, all the way back to single-celled organisms in the distant past. Those who do doubt these facts are about on par with those who think the philogiston theory of combustion has merit.
    • Re:Arguable? (Score:3, Insightful)

      by Jerf (17166)
      You can equally sensibly interpret the submitter's statement as an observation that we are not evolved for the civilization we live in, because in terms of "number of generations" since civilization (especially 20th and 21st century civilization), it hasn't been anywhere near enough, plus there has been a "lack" of selection pressures in many cases.

      (It's not quite that bad because our civilization has evolved to match us, but it's still not perfect; for instance, you can blame "lack of willpower" for the c
    • He could just be saying that perhaps evolution could have done a better job for us than it actually has. We've been at it for 3 billion years, and not got a lot to show for it really.
    • No, dammit. (Score:3, Insightful)

      by JessLeah (625838) *
      That isn't the scent of an evolution denier. That is the scent of someone making a goofy little crack about how stupid people are nowadays, despite the effects of billions of years of evolution. There is credible evidence that we really are getting dumber [fourmilab.ch].
    • Aww jeez, who modded that flamebait as insightfull? Sigh...

      If the theory of evolution has worked well for us -- even if this is arguable these days

      Do I detect the scent of an evolution denier?

      No.
      What he meant was: Evolution has worked well for us, though seeing all the idiots running around its arguable that evolving AIs will grow in intelligence.

      It was a sort of joke, wich you did not get, and attacked, and now you've woken up the creationist trolls.

    • even if it is arguable these daysDo I detect the scent of an evolution denier?

      I took the arguable "these days" part to be questioning if evolution is currently helping, or indeed affecting in any way, humanity any longer. It seems all of the anti-survival behavior we humans engage in does not affect our ability to reproduce.

  • by keller (267973) on Sunday May 23, 2004 @03:41PM (#9232074)
    Nope, when they master capture the flag, how soon before these Evil Robots are ready to take over the world...
  • Bzflag (Score:3, Informative)

    by super-momo (691644) on Sunday May 23, 2004 @03:45PM (#9232096)
    Wikipedia mentioned Quake and UT. Bzflag [bzflag.org] is also a great CTF game, and a classic.
  • by Entropy Unleashed (682552) on Sunday May 23, 2004 @03:46PM (#9232103)
    Robots are cool and all, but why bother building and programming robots to find mines when we already have [theage.com.au] biological robots that can do the same thing while running off of water and a little bit of food. It seems a bit like a wonderful solution to a problem that doesn't exist - evolution has been doing pretty darn well at doing this sort of thing so far, so I'm not really sure why would need robots after all this time.
  • It would be cool if we could evolve robots so they can make an accurate choices based on facts, like a human being would, without being biased.

    Some examples of the tasks a robot could do are judge criminal cases, mark exam papers, and moderate slashdot posts.

    However although the robot will probably make the right choice more times than a human we still wouldn't trust these important decisions to them.
  • by MajikMan (453619) on Sunday May 23, 2004 @03:47PM (#9232113) Homepage
    ...no one asks it to play global thermonuclear war.

    I vote we drop capture the flag, and just start up the tic tac toe game right now.
  • by crem_d_genes (726860) * on Sunday May 23, 2004 @03:48PM (#9232124)
    "After several hundred generations, the neural networks had evolved well enough to play the game competently and were transferred into real robots for testing in a real environment. "The trained neural networks were copied directly onto the real robots' onboard computers," said Nelson. "

    As someone who spent a considerable amount of my childhood less interested in 'organized' sports and instead playing this game, it seems the whole point of playing Capture the Flag was to develop strategies in how to win. We had a set of rules that evolved over the years, depending on how many kids were playing, what time sunset (or the first person called back to their house would be), etc. We even had evolving words that were based on nonsense - or the inability of one of the younger kids to say a word (for instance - in some "Steal the Flag" games - the term "electricity" is used to talk about a strategy that involved making a line of kids that attacked from one end - they all held hands in the stragegy so that if anyone was captured they would automatically be "freed" by the "electricity" back to their own side. We deemed this a violation of the intent of the game, so we had a *no electricity* rule some little kids couldn't pronounce right - so it became "no a-la-ca-triss" - or something like that).

    The game wasn't about *object avoidance*, it was about kicking ass through completely ad hoc strategies that had to be original because the teams always traded players rapidly, so you didn't want to make a rule or come up with something that would come back to bite you.

    In this way - the random nature of our game was more like evolution than the winning was (it shuffled the components and allowed for *mutations*). The fact that the model showed no improvements with greater numbers of computers is not in line with what actually happens. The best games were the huge ones.

    This simulation was probably a lot of fun to watch once the program was transferred to the robots though...
    • First though you had to learn how to walk, how to chase things, etc. You already had a huge advantage on the field. These systems need to learn the basics of individual operations before you can expect them to start using group operations.

      Evolution didn't start out with multicellular organisms. It started out with single cell systems that had to compete on its own. THEN you started seeing multicellular organisms evolve.

      The robots are just starting out. Give them some time before you apply the whole "
      • I agree with you about the time needed.
        I wonder though - is this really an evolution model - or simply the normal iterations that one would go through testing anything from scratch? This seems more like *intelligent design* - not really random - but specifically selected to *win* - Whereas evolution occurs in response to a changing environment - not to preset rules.
        • But isn't the environment just a set of rules? True they may change after a time, but that's just a new game.

          With the new rules you will get a new "race" of robots that learn faster or are prediposed to the new game. They'll adapt to the new changes and either the old robots will find a niche to survive or they will go extinct.
  • Aww man, (Score:2, Flamebait)

    why do the French have to make everything, including robots, seem lame? I can remember a time, like 5 minutes before I read the article, when I thought robots were badass. Now the whole favourable perception has been ruined.
  • It's not as though this approach hasn't been thought of before. The problem is the limitations of contemporary AI tools, combined with limitations on our hardware. Also, as tasks become more and more complex, it becomes much more difficult to "evolve" systems that behave exactly as you want them to. There are a number of stories of neural nets being trained to recognize some feature from a set of training inputs, and instead keying in on some completely different and irrelevant detail.
    • Also, as tasks become more and more complex, it becomes much more difficult to "evolve" systems that behave exactly as you want them to. There are a number of stories of neural nets being trained to recognize some feature from a set of training inputs, and instead keying in on some completely different and irrelevant detail.

      Hmmm. That describes my boss pretty well. I think i'll check for a port on the back of his neck.
    • The complexity of the problem that you can train a neural network to do has a lot to do with the complexity of the network; a single neuron is basically only good for either-or classification of linearly-separable input categories. A larger group can perform much more complex functions; my current project (IAAUSRNNPT*), combined with a nifty look-up table, is getting close to being able to read any kanji character in the Japanese alphabet (IAAAAL**). To get remotely close to the complexity of the human brai
  • If they can evoluate, why not try to show them how to find the best solution on a given computer program?
    I'd like to see how a robot could work on his own code too, to try to always be faster.

    Given the fact these robots (programs after all) can evoluate/learn and re-use this evolution, they should be able to learn until their hardware limis them.

    As I see it, its all about a really basic but really well done base code, who will start the comparison, memory and self-modification of the comparison code that
    • That's almost exactly what neural networks do. They are essentially re-programming themselves based on the results of their actions, without actually writing lines of code.
    • If they can evoluate, why not try to show them how to find the best solution on a given computer program?

      It's already been done. [genetic-programming.org] Genetic algorithms is the evolution of a particular solution to achieve the best results, such as finding the best solution to the travelling salesman problem. Travelling to 100 cities in the best possible order to minimize the distance travelled, this would take 100! (that's factorial, not just an exclamation) calculations to search through the entire solution space. Using
    • While we're at it, we'd better evolve programs that can explain the operation of the code they write. The "best" solution might indeed be fast/efficient/whatever, but it might also be completely incomprehensible to humans - especially if the problem to be solved is even slightly complex.

      I recall seeing some comments on this topic from researchers using genetic algorithms to evolve circuits. The evolved circuits worked really well, but nobody could get a grip on how they actually worked (at least, that wa
  • The problem with robots which are evolved to handle a task, is that you have no idea of what the programming is like in their heads. You have a working robot, but you don't know how to program one yourself. All you can do is train it and analyze the "brain patterns" afterwards. These things are hard to decipher.

    The robots-as-dogs method will probably win because it gets results quicker than the programmed-thought-by-human method. Could be dangerous though, because you don't know what makes it tick.

    Asimov'
  • 10 to 50 years? (Score:4, Insightful)

    by drinkypoo (153816) <martin.espinoza@gmail.com> on Sunday May 23, 2004 @04:16PM (#9232267) Homepage Journal

    But the researchers want to build controllers for robots that adapt to completely unknown environments. And this will not happen before 10 or maybe 50 years.

    I disagree that it will necessarily take even 10 years and it will certainly take less than 50. Pathfinding and object search algorithms are strong even today. With a combination of radar, sonar, lidar, and optical recognition, I think we should be able to create robots which traverse formerly-unknown terrain in ten years or less.

    I'm not trying to trivialize the difficulty of the problem, all the stuff we take for granted as we navigate a room is really quite a lot to deal with and it is only through practice that we are so successful, but an awful lot of effort is going into these problems (I know "more than ever before" is cliche and obvious but nonetheless...) and it is a top priority for so many very smart people that I cannot see it taking even a decade for useful robots with these capabilities to be in use.

    Of course, it depends on what you want them to do when they get there...

    • Re:10 to 50 years? (Score:4, Insightful)

      by hypnotik (11190) on Sunday May 23, 2004 @06:26PM (#9233028) Homepage
      10-50 years is probably a realistic estimate. Spend a bit of time in the AI/AL world and you'll get a picture of how much we still need to learn. Evolving things in a computer simulation is fine, but once you step out into the real world, you see a whole new set of problems. In fact, evolving anything is hard. Your simulation has to be perfect otherwise you end up with a solution which has evolved to take advantage of flaws in the simulation and not perform the task.

      Back in the early days of Genetic Algorithms, there were experiments which tried to evolve robots in simulation to go to the end of a corridor and turn in a specified direction. However, once the robots were evolved and "built" in the real world, they often failed. The reasons for the failure were numerous, from not having the same dimensions for the corridor to different motor sensitivities in the robot itself.

      They've gotten around this somewhat by feeding randomness into the simulation (see Nick Jacobi's Minimal Simulations). However, for any complex real world type problems, there just remains too many variables to vary and evolution doesn't work as efficiently.
  • by X86Daddy (446356) on Sunday May 23, 2004 @04:19PM (#9232284) Journal
    Friday's A Softer World [asofterworld.com] strip was about this very topic!

    Read the rest from their homepage [asofterworld.com].
  • My bowling instructor at State (yes, I took bowling, stop laughing) once said that we could easily build robots that could bowl. But then, what would the point be? They'd bowl a perfect game every time.

    And now I have the answer to that question. We need robots that can bowl through minefields! It's all clear to me now.

  • by Tablizer (95088) on Sunday May 23, 2004 @04:21PM (#9232295) Homepage Journal
    The introduction makes it sound like training neural networks is evolution. Neural networks and genetic algorithms are two very different technologies, although they can be combined.
  • Basic Fallacy (Score:3, Insightful)

    by AlecC (512609) <aleccawley@gmail.com> on Sunday May 23, 2004 @08:19PM (#9233740)
    There is a basic fallacy in this sort of research - that evolution will necesarily develop some kind of intelligence to solve problems. Evolution will do "what it takes" to solve a problem - and no more. If you attempt to use evolutionary techniques to, for example, solve mazes, you will end up with a system very good for solving mazes - and nothing else.

    This happened in computing in the 70s. Intel found it convenient to solve the problem of calculator design by buoilding the 4040 - the first microprocessor, But this was in no way *necessary* - Intel could have continued down the old line of discrete logic.

    Evolution is a powerful tool - but not a panacea

"For the man who has everything... Penicillin." -- F. Borquin

Working...