Slashdot Mirror
Robots, Robots, Robots
destructor writes: "It looks as though robots can answer the questions of life and capitalism through robotic soccer simulations. I found this article over at megarad.com that tells us about Dr. Balch's experiments with soccer robots [NYT, free reg, blah blah]. For now, it is purely a computer simulation, but I guess it will be turned into a physical environment rather soon."
Additionally Shabazz writes: "The SF Weekly has a story about a band called 'Captured! by Robots' that started by Jay Vance (who some may recognize from Skankin' Pickle) and several robots that he created. The band is a bit out there, but something that any true geek can appreciate. Maybe this is the start of something great!" Additionally Phred noted that the Oregon Robotics Tournament being held this Saturday (December 1).
That's pretty cool . There are lots of sort of soccer championships for robots, one of them with aibo, the famous robot dog. It's especialy interresting in the field of the computer vision interaction.
From the BBC's web site from August this year:
Robot world cup kicks off
Now, next time those professional athletes go on strike we can just replace them with robots!
"Everything you know is wrong. (And stupid.)"
Moderation Totals: Wrong=2, Stupid=3, Total=5.
How can robotic action teach us about human relations? Humans are so much more complicated than capacitors, circuits and processing units that there is very little that one could find that would allow us to understand human behavior.
If you look hard enough at anything, you will find what you are looking for.
Remember a scooner is a sailboat and there is no Easter Bunny.
I personally think that they should have robots competing to see who can complete a jigsaw puzzle first. It is completely unbiased and provides a test of both vision and dexterity, and both to fairly high degrees......but then what do I know?
"I'm tired of all this 'Aren't humanity great' bullshit. We're a virus with shoes" - Bill Hicks
... about the fact that robots programmed with algorithms created by humans should solve the human way. Since they don't know the outcome of these algorithms how did they find them? Pure luck?
Why simulate algorithms already created, when it's possible to predict it through calculations? The only reason I can think of is fun. But simulating something is just visualising something we already know.
Look a monkey!
http://www.me.iitb.ernet.in/yantriki/
IIT Bombay has a robotics festival called Yantriki.
They've had games like TUG-O-WAR in '94, moving on to SOCCER, BASKETBALL, BUNNY WARS, CARROMINES, SUMO WRESTLING and WATER POLO, since it's beginning in '94.
-Shaunak.
They should replace the fans with robots. That way they might actually be able to sing in-time and tune. Also, they could smash stuff up after the game _carefully_ - cutting tables into neat little bits before throwing them at the opposing teams fans and not at passers by or camera-crews. You could deactivate the most violent ones.
I think this experiment could teach us allot about stupid people, and how they behave in packs.
This comment does not represent the views or opinions of the user.
I found the article very interesting in so much as you got more diversity when the team as whole was rewarded/punished but I really don't agree that the way the the robots played has anything to do with capitalism.
I would also strongly disagree with the idea that capitalism is just about individuals seaking one goal (to obtain wealth) when in practice it is groups of people working together to obtain that goal. For example in a company (of any size) nothing works if people are just blindly following their own goals, but it does work if people are all working for the greater good of the company. People do not just sit there thinking about whether a decision is going to be benfical to them based on "My Wealth" vs. "Company Good", because if that was all that they did they would obtain neither since their capitalist ideal is at that point linked with the success (or not) of that company.
Of course there are exceptions to this... But IMO the essence of capitalism is that people as a collective, not just as individuals, can strive towards those goals of accruing wealth, and it is this that the robots are doing. The team that works as a team is the one that is consolidating its wealth and is the one that I think is being more capitalist simply because if they do well as a team then they are doing well as individuals in a team (For example if there are 5 bots and they get 10 goals then each of the bots is 'worth' 2 goals)
Out of interest I looked up capitalism also. The definition that I got from Websters is as follows:
An economic system in which the means of production and distribution are privately or corporately owned and development is proportionate to the accumulation and reinvestment of profits gained in a free market.
I don't see anything in that that says that capitalist is just about getting rich....
Whats the big news? See this( at Uni of Stuttgart ) they even have ROBOCOP matches!
Robocop
This is Multiagent theory in the context of AI.
Voltaire: God is dead.
God: Voltaire is dead!
The general technique is to use an approximation of search (finding based upon luck) which tends to find good solution. Generally, it is based upon luck, and the only way to guarantee that its human-like is to restrict the search space to human-like possibilities. Whatever human-like means.
This is NOT the simulation of algorithms already created always, its often the creation of totally new algorithms, which can be extracted after development - a learning approach.
How? Straight statistics, support vector machines, decision trees, neural networks, fuzzy logic, and simulated annhealing are all common techniques to lead towards the goal. Who knows what they actually use.
Mod me down and I will become more powerful than you can possibly imagine!
I don't know what Mr. Blach is exactly doing but the Robocup is already there for years.
Owner of a Mensa membership card.
A poster above had commented on Logical v. Illogical Actions. I'd agree with him, but what actually sets us apart is emotions, or instinct, however you may choose to call it since the two are subtly connected.
Robot emotions play a much larger part than most people know. In fact, any serious researcher into AI would know that emotions are nothing but another decisive factor, except that they are not well understood.
I'd seriously advice looking at Arthur T Murray's research into this area available here. It has a lot more to it than mere `entertainment` value.
And if you really are worried about robots having souls or a conscience, you should read this!. It is a pretty insightful article into what robots may have to do to qualify as humans.
You will get a lot more useful information if you do social experiments with people instead.
I agree that group performance as well as individual reward should be rewarded, and in most large companies it is. Annual bonuses are often based on company performance multiplied by individual performance. Rewarding a small group/division is useful where its possible.
Soccer is individualistic. You will make more money (or receive more fame if amateur) as a striker than as a defender, so if you have the talent, that's the position you want. Economics also alow for people to realize that their talent level may be better suited to being a defender.
The robot social system may also evolve into letting better robots be the strikers, but its incidental. The motivations of individual fame and wealth are not being accounted for, but is what causes teams and players to stabilize into their positions.
If you don't want to register, just visit this url instead.
-J
Alexis 'jeriqo' BRET
Think this capitalism argument is based on a false premise...
The results may be surprising to those who believe that the pursuit of individual rewards -- as in capitalism -- encourages people to develop a diversity of ideas, points of view, goals and strategies to achieve them.
To summarise quickly - he says when robots are only rewarded for scoring, they all go for goal, whereas when all on a team are rewarded when one scores, they work effectively as a team.
Now - last time I checked, when my team (company) scores (gets a good deal), we all benefit, not just one of us. So it's a bad analagy, which the guy can stick up his arse (or ass, if we're in socher-land).
'Cause you can't leave out the influence of the crowd on a match!
Blar.
Pretty picture next to the story. I don't think I've ever seen it before.
--- At my sig, unleash hell.
I think we need more than a grain of salt, here, we need a whole bag. I've seen many such pieces of research, and the pattern is this: 1) Code up some sort of multi-agent simulation; 2) Note a behaviour vaguely similar to something seen in humans; 3) Set your Extrapolater on Max; 4) Hand it over to the university's PR department to write press releases and jump up and down about it.
Note that Balch himself is much more cautious (than the journalist) about the applicability of his findings to human interaction. Who knows, he may even be correct that using artificial agents could be useful in testing social hypotheses -- at least they're controllable and repeatable, though of course that's both good and bad when it comes to trying to model humans. But you'd be crazy to conclude anything about the properties of capitalism or socialism based on individual vs. global feedback schemes.
First one male robot scores a few lucky goals.
Then some female robots throws themselves at him.
Soon he believes the team can't win without him.
Large salary demands and endorsement deals follow.
:-)
I'm training my Bio Bugs to play a wicked game of Tetherball.
------
Let me give you the lowdown
Robosoccer has been going on for years and years, and there are frequent cups in it. I fail to see the news here. Besides, it's been slashdotted lots of times
Isn't that the word Yanks use for football because they've stupidly reserved the world "football" for their childish, dumbed-down version of rugby?
All problems as far as we know can be placed into a few general categories based upon their complexity. It can be shown (though I won't be doing it) that both making a robot do what you want and making a human do whatever are decidability problems: both have the same degree of complexity in the general sense - one task may be harder, but only because its "bigger" - containing more dimensions or variables that must be tested. HOWEVER, if we can find a technique that will describe/model the robots, then it is extremely likely that we can use a morphologically equivalent technique to describe/model human behaviour (based upon the fact that both problems are decidability problems).
This is one of the basic ideas of modern artificial intelligence.
Mod me down and I will become more powerful than you can possibly imagine!
As a youth soccer coach, I could have predicted this finding:
"Under the first scheme, a reward signal is sent only to robots that score a goal. As the match progresses, every team member ends up learning the same sequence of behaviors -- going after the ball in a solo effort to score. As a result, the circles on the screen bunch around a single point -- wherever the ball is -- leaving the rest of the field open to attack."
Apparently this guy has never watched six year olds play soccer with Mommy and Daddy cheering from the sidelines for them to "kick it!"
This reminds me of a book:
5 4/ qid=1006962846/sr=2-2/ref=sr_2_79_2/104-6700944-68 20768
No Contest : The Case Against Competition
Here's a blurb:
Updated for the 1990s, the award-winning book that stands as the definitive critique of competition. Kohn says that contrary to accepted wisdom, competition is not basic to "human nature," but actually poisons our relationships, damages our self-esteem, and holds us back from doing our best. Includes new matial on "cooperative learning" in the classroom. "By perceiving that cooperation is the answer, not competition, Alfie Kohn opens a new world of living."--Dr. W. Edwards Deming.
Here's an Amazon link:
http://www.amazon.com/exec/obidos/ASIN/03956312
Allright, I admit it, I'm a sucker for a good story. This has the makings of a great story. How the heck did you almost get run over by a 100kg research robot?
(Yes, this is off topic)
From www.m-w.com
m ent/PROGRAM.htm for one historically influencial socialist group and their platform.
Main Entry: capitalism
Date: 1877
: an economic system characterized by private or corporate ownership of capital goods, by investments that are determined by private decision, and by prices, production, and the distribution of goods that are determined mainly by competition in a free market
There is nothing in that definition that says only one individual may benefit. Indeed, it's already been pointed out that private individuals in a capitalist system already voluntarily cooperate for their mutual benefit. Capitalism just means that individual decide rather than the state.
A better political comparison would be between a robot team where each robot has control over its own robot body, and a robot team with a central master control, telling each robot what to do. Indeed, its interesting that his robots evolve individually where there is a sort of distributed intelligence, rather than under some central authority as in socialism or communism.
Just for comparison,
Main Entry: socialism
Date: 1837
1 : any of various economic and political theories advocating collective or governmental ownership and administration of the means of production and distribution of goods
2 a : a system of society or group living in which there is no private property b : a system or condition of society in which the means of production are owned and controlled by the state
3 : a stage of society in Marxest theory transitional between capitalism and communism and distinguished by unequal distribution of goods and pay according to work done
I think its interesting that the third definition allows for unequal distribution of goods and pay.
Also look at http://fcit.coedu.usf.edu/holocaust/resource/docu
It seems to me that this is an obvious result. The rewards being given are very one-dimensional...they only reward scoring. If rewards were given individually, but also given for, say, blocked shots, or breaking up passes, the results would change. You would be able to condition "players" through individual rewards to play offensively or defensively. Variable magnitudes of reward would help to, since you could reward a prospective forward more than a prospective defenseman for scoring a goal and vice versa for defensive play.
It is somewhat interesting that rewarding the whole team when goals are scored produces some robots specialized for defense, but really, that result is obvious too. Basically, what you are saying with the reward is that whatever position that robot was in, or sequence of actions they were taking at the time of the goal was a good one. If they were in a defensive position at the time of the goal, they'll become slightly more predisposed to defense (assuming that's more or less how their programming works). In other words, it seems to me that all these experiments demonstrate is that their algorithms work.
Which one of the robot soccer moms is bringing the robot oranges and robot water in their robot minivan this week?
As someone who dropped out of high school because of stupid ideas like cooperative learning and "team teaching" and block scheduling which were the antithesis of learning, I'm so tired of seeing this drivel being accepted and put into practice.
The Article quotes a 'cooperative learning' researcher extolling the joys of diversity that it produces. It doesn't mention that while kids are learning diversity, they're not actually LEARNING anything.
Group work in a school setting produces smart students who don't excel because they're holding up the rest of the group, mediocre students who can slack because the intelligent ones will do the work for them, and slow students who never get the attention they deserve.
I'm tired of these education theorists taking their insane pet theories and putting them into schools. And its even worse when they use some sort of silly robotic experiment to back them up. Isn't there some way to stop these wackos?
I agree with the opinion of many others that the connection drawn between this simulation and capitalism is a little far fetched. The modivation behind human behaviour is far more complex. Besides group production is rewarded through profit sharing, bonuses and stock options. However, I do think this simulation shows something useful on a far more obvious level. That is that rewarding group behavoir of autonomous agents may produce better group behaviour than individualistic behaviour. As opossed to some behaviour such as bird flocking which has been shown to coorelate well with a purely individual reward (motivation) system. I think one fundemental flaw in his reward system is that rewards are only based on goals. In an individual sense, this will never produce a goalie. So without the proper reward stimulation you can not expect the team to perform well on all levels. A useful experiment to show what's really happening here would be to set up two sets of 50 (or more) teams. Then keep statistics on each player such as scoring, saves, take-aways, etc. For the first set of 50 use an individual reward system. For the second set of 50 use a team reward system. Then draft two teams consisting of the top players from each division and pit them against one another. This might give a better understanding of which reward system really produces better players. Perhaps the individual reward system would produce some real supper stars, where the group reward system only produces good teams, but no stars. Jettra
Six year old soccer players is what you get when you give the MIT AI Lab folk one day to write a soccer-playing team in Java:
AI Olympics Sockey
(note: this applet was written independently of the Robocup tournament and doesn't share the same rules or physics.)
Although rugby is less of a wussy game - argh! Stop being a hippocrite! Naughty Shade, naughty!
It's very astounding how many faceted threads this article brings up.
The author of Logical v. Illogical Actions addresses some interesting points concerning the relevance of using robots to test human-based theories. To continue rmo6's statement, it's important to remember that in the field of nonhuman studies applied to human research, the general rule of thumb is that, as you go up or down the evolutionary tree, you usually end up trading control for relevance (and vice versa). That is to say, rats will do just about anything for the pellet, and monkeys will throw their shit at you. I'd love to follow where and IF robots fit in that scheme. GodSpiral brings up a rather Skinnerian Complaint - that individual performance should be focused on; yet, to answer that, aren't groups composed of individuals? The designers seem to be saying that robot behavior possess tremendous relevance, especially within the field of industrial psych, but like halftrack points out, how do these robots end up going along these paths - serendipity or predilection?
Also, if you want a fascinating survey of AI development (through communication, technology, industrial, etc) try Janet Murray's "Hamlet on the Holodeck."
Of course they can. And the answer is 42.
You might get some interesting results by rewarding the robots individually for scoring, but assessing penalties across the group for being scored upon.
That way, maybe you could start getting to some of the more complex issues people on the board are describing. Particularly once you started making certain teams -- or even computers -- more or less "selfish" by varying the degree to which teams or robots were affected by "group-wide" penalties as opposed to individual rewards.
Cooperation can be explained as a second order genetic effect, i.e. helping some close relatives survive might get more of my genes into the next generation than if I survive and they die. But it can also be explained by the handicap principle: Taking on a visible handicap can increase an organism's chances of attracting a mate. The Handicap Principle, once a controversial theory, is now accepted as our best explanation for all sorts of risk taking behavior.
Taking risks (cooperation, a big flashy tail, antelopes 'stotting' in front of cheetahs, etc.) proves genetic fitness to potential mates.
So cooperation is really a form of 'persuit of individual rewards.' A roboticist can simply choose the reward parameters they like for their experiments, but in nature, reward systems are products of evolution favoring winning strategies.
Cooperation is a better strategy than selfishness because the chicks dig it, and it might help your relatives pass on more genes.
- None can love freedom heartily, but good men; the rest love not freedom, but license. -- John Milton
robots are as bad as mexicans. they will come in and steal our jobs. and thereby putting tons of hard working americans out of work. i compeatly agree with the fact that this is some cool stuff. i know this is off topic but the advancment shouldn't come without a pirce. The powers that be have already put out alot of blue collar workers out of work with the robotic arm. will this happen in other fields. what is to stop them. im not anti-tech. i just belive that there is a line to be drawn. if these robots are to be made. they should not be able to be used to do the jobs that others deserve.
If they had robots on the sidelines who did nothing & were rewarded when the team scored, they would probably continue to do nothing when put in play.
That seems to be a acturate simulation of Socialism IMO.