Slashdot Mirror
AI Researchers Produce New Kind of PC Game
Ken Stanley writes "In an unusual demonstration of video game innovation with limited
funding and resources, a mostly volunteer team of over 30 student
programmers, artists, and researchers at the University of Texas at
Austin has produced a new game genre in which the
player interacively trains robotic soldiers for combat. Unlike most games
today that use scripting for the AI, non-player-characters in NERO learn
new tactics in real-time using advanced machine learning techniques.
Perhaps projects such as this one will encourage the video game
industry to begin to seek alternatives to simple scripted AI."
Slashdotted before it even went live. Here is a working link. Downloads are currently at 511, I hope their counter has more than 9 bits...
If it's UT anywhere but Austin, you say where.
If this technique provides for fun gameplay, or more importantly, a notable difference in the experience, then sure, it might become more common.
Keep in mind though - entertainment is meant to be entertaining, not neccesarily realistic or academically advanced.
---------
Get back to me when my brain starts working.
One of the earliest forms of AI I ever learned about was MENACE. A pre-computer means of training a system to play and win Tic-Tac-Toe. I will confess to loosing more than a little time "training" my system.
What do you know I wrote a novel
Imagine a beowolf cluster of these AI robots!
This is a neat concept, with or without the "neuroevolution" approach (evolving artificial neural networks with genetic algorithms). Including human brains in the training loop for algorithm development is key. The reason so many AI algorithms have found limited application in fielded physical systems (such as weapon systems) is because the competing approach--dozens of smart engineers, working long hours, tweaking human-readable algorithm code and Monte Carlo simulating the tweaked designs over and over for years--is so effective.
The DOD will get interested, and use a similar technique to train -real- robots?
"If anything can go wrong, it will." - Murphy
Perhaps the robots can teach us geeks to get on up and shake that fatty booty!
I could have sworn there was a Japanese game already like this. You could put different parts on the robot and teach it techniques in combat to fight other user created robots. It is probably no where near as complex but the same idea.
Maybe someone has the name?
This isn't entirely a new idea. CROBOTS, for example, put one in the position of designing AIs that control tanks and then pits them against one another in an arena.
Better download it now before their server learns to resist the slashdotting
403 Forbidden. Nice try, maggot
The grass is always greener on the other side of the light cone.
Anyone have a torrent for the game?
You are such a nerd.
The drivatars in Forza Motorsport (Xbox) seem somewhat similar to this innovate technique. AIt's an AI that is taught by example. Check drivatar.com for more details.
Oh hell, you know this will taken over by /.'ers AND do /.'ers know a damn thing about soldiering?
Probably not, but beware -- you may just create a robotic system administrator/repairman. Don't put yourselves out of a job!!!
Ignorance is not a crime; neither should it be a way of life
Congress control $ = inmates run the asylum
Joshua: Greetings, Professor Falken.
Stephen Falken: Hello, Joshua.
Joshua: A strange game. The only winning move is not to play. How about a nice game of chess?
For those of you who actually look on a user's history of posts, yes this is a variant of another post I did, however it's apropos here as well.
Sorry about the writing. Robot fingers, you know? Cliff Steele in DOOM PATROL #23
> Perhaps projects such as this one will encourage the video game
> industry to begin to seek alternatives to simple scripted AI.
hopefully it will encourage the video game industry to begin seeking alternatives to Yet Another High Resolution First Person Shooter.
Imagine a beowolf cluster of trained anonymous cowards imagining a beowolf cluster of these....
Maybe, after all, such a cluster exists because there is such a post on everything remotely clusterable.
"may seepeeyou iz ay newral ned prooozezzor... a laaarning compooota"
I implemented learning AI in a couple of popular video games (including at least one multi million unit PC title) more than 5 years ago, and I'm pretty confident I wasn't breaking any new ground.
"Who is the Journal of Quantum Physics going to believe?" --Stephen Hawking
Perhaps projects such as this one will encourage the video game industry to begin to see alternatives to simple scripted AI.
Not unless companies like Sony start making their hardware with ideas like this in mind. 2 TFLOPS is impressive for insane real-time graphics operations, but floating point operations aren't exactly optimal for things like AI.
(I know TFA was referring to PC games, but so many games are multi-platform these days that often developers just work with the least common denominator.)
Mods: Do you disagree with me? Go ahead and mod me down. Meta-mods will sort it out. Good luck!
It would be a lot easier to train a robot to train the other robots to fight (in the long run)...Wouldn't it?
"Galapagos" by Anark had a robot creature with some kind of neural net, and you had to teach him to navigate around by providing him with appropriate stimuli and rewards.
It could get frustrating--sometimes if he hit a particular deadly obstacle too often, he'd become traumatized, and would then refuse to go anywhere near it, which could make the level impossible until you had allowed him to wander around and petted him and calmed him down.
Great game, though. I wish there were more like it.
GCHQ Quantum Insert installed. If only our tongues were made of glass, how much more careful we would be when we speak
If it is anything like David Lightman little game with WOPR's Joshua then I am out of planet earth with the next shuttle flight.
Linux port is comming soon :D And it's gonna use GTK1!?!?!
Cheers,
RoadkillBunny
There's no Linux version, although it seems that one is planned, and the Windows installer won't run under Cedega :(
Respectfully, the flagship institution UT Austin is simply "UT" everywhere in Texas except UT-El Paso, UT-Arlington UT-Muleshoe, etc. TH UT '91
Ah, fill in your own punchline here.
Please dupe this article when there's a torrent available and the Linux version is finished.
Until then it's quite useless.
I'm a chainsmokin' alcoholic sociopath, so-ci-o-path
I am greatly confused.
What is a bittorrent magnet link and how do I use it?
Also, please make the wikipedia page for this better.
http://en.wikipedia.org/wiki/Magnet_link
Thank you.
Although Azureus is mentioned, I do not have a lot of ram, so I dont want to use that.
Didn't rtfa, site down. Wonder if they trained their webserver this way as well.
It's been done before.
Apologies to the submitter!
Only for the purposes of helping distribution, and for a limited time, torrent available at nerogame.exe.torrent
Visit Lockjaw's Lair. He won't bite.
But you expect ME to finish CODING the game for you?!?!? Will they stop at nothing?
Isn't "scripted AI" a contradiction in terms? Can't we start using more correct jargon when referring to computer controlled enemies/allies until AI is finally perfected?
How about this:
Artificial, Non-Intelligent Matrix Associated To Individual Object Nodes
Or ANIMATION for short...:)
I was in the park the other day wondering why frisbees get bigger and bigger the closer they get - and then it hit me.
Is it fun to play?
The problem with expensive investments in AI is that the publisher must have a series of successful games built on the fruits of that labor before there is any profit. This could possibly be mitigated somewhat by licensing this engine for use by other companies, but this is also weighed by the fact that your competitors are now using the same or similar types of advanced artificial intelligence in their games which may hurt sales of your own games. Large publishers, such as EA and Microsoft, have the resources and wherewithal to make these long term bets, but the smaller boutique firms have neither the willingness nor the ability to finance the development of these types of advanced engines in house. It may be useful to look at some numbers from 2004, courteously compiled by the http://www.shrapnelcommunity.com/blog/2005/02/24/" >shrapnelgames blog.
The total revenue for the game industry in 2004 was 1.2 billion dollars which was down 100 million from 2003. During this same period only two games had sales of over 500,000 units, but there were 18 games which had sales of 250,000 or more. Based upon the varying definitions of what constitutes a "new release" there were roughly 1,100 games released in 2004 of which maybe 6% earned a profit. The average budget for a competitive game is said to be around two million dollars with an average break even point of around 110,000 units sold. The average retail game price is $24.45 with only 5,000 total units sold.
Clearly, the open source community is willing to undertake these efforts on their own initiative or for other reasons related to research, as was the case with the student produced game. I am in no way denigrating the efforts of these students, what they produced with the resources available to them was simply amazing and of surprising quality. However, in the world of retail games it takes a certain amount of marketing, advertising, and Wal-Mart end caps to rise above the background noise, unless you are like the aforementioned established game companies and the reputation speaks for itself, at least until they release a real stinker. At the end of the day, when all things are factored in, there is simply not enough money in the budget of the average game to make this type of advanced artificial intelligence worth the risk and expense, at least right now. However, if there is any constant in the game industry it is change and this will probably change in the years to come. I would like to see some new and innovative games too, instead of Madden 2017, but it looks like we will have to wait a while yet.
Why is it that folks throw around "AI", when all they've done is teach computers how to switch around and use pre-defined strategies to deal with a tightly situation? Real AI is when you can take pre-learned strategies, adapt and apply them to a situation that are only minimally like situations you've faced before?
:)
Of course, using that definition, most folks aren't intelligent... which makes me think my definition must be close
See my journal for slashdot ID's by year. Mine created in 2005. http://slashdot.org/journal/289875/slashdot-ids-by-year
> Not unless companies like Sony start making their hardware with ideas like this in mind. 2 TFLOPS is impressive for insane real-time graphics operations, but floating point operations aren't exactly optimal for things like AI.
For AI based on neural networks, 2 TFLOPS might be just the thing.
Sheesh, evil *and* a jerk. -- Jade
..and he says that's what the Marines are. But really, the DoD does fund a lot of machine learning; however, the current state of the art only allows machine to solve specific problems. You need a traning metric, etc. and that's not trivial.
Repeat after me.
7 SPEs for vector processing = bayesian learning goodness
Those 7 SPEs with their bandwidth will be able to take inputs like video, sound, even EEG data from the brain. Combine this with bayesian learning techniques and the machine will infer what factors in the raw data correlate with its advantage in the game world. Imagine a game that can sense your fear with the right "helment" perhiperal containing active electrodes.
All the people who are saying SONY/IBM wasted die space on the SPEs don't realize why they called it the CELL.
It acts like the human brain.
The SPEs will be used for inference. Much like freud described our ego, the cognitive part of our brains.
The PowerPC chip to model the game world and to work toward certain ends with respect to the game, based on results inferred from data going through the SPEs. This sounds like the behavioral part of the brain. What freud called the id, if you are looking for a biological analog.
Hrmmm... Just some thoughts.
False.
FLOPs are not generally useful for things like scripted AI which are very branch heavy with a lot of indirection, and many possible branch targets and data requirements.
The techniques described in this game are highly mathematical in nature with a small memory foot-print, (adaptive neural networks and genetic programming via Kenneth Stanley's NEAT algorithm) and would benefit hugely from parallel vector proccessing.
Additionally, at the end of the day, the AI decision making is not nearly as expensive as the proximity-query and pathfinding routines that affect the decisions. These routines also benefit hugely from vector processors and high bus-bandwidth.
So fittingly, the AI will only suffer if the human intelligence can't adapt and make the fairly obvious decision to move toward more mathematical AI routines.
..a game where you used preset commands and scripted them into a run routine for robots and then pitted them against one another?
r eview.html
http://www.gamespot.com/ps/strategy/carnageheart/
I was never really very good at it but it intrigued the heck out of me back in 98/99...
There is simply too much glass..
if referer=slashdot
return 500
Simple yet efficient way to prevent DoS.
TODO: 753) write sig.
You do realize that there's much more to A.I than statistical methods (Bayesian)?
"It acts like the human brain."
Only for those who don't understand the human brain.*
*And yes I'm not implying we know everything about the human mind. Just that we know enough that the CELL isn't it.
I liked this series of games... I think I spent the most time with ATRobots, where programming was done in a sort of assembly.
-------
Incite and flee.
http://nn.cs.utexas.edu/NERO/download.php?go=yes
I'd like to know if the NEROs can evolve more advanced tactics such as:
When its health is less than 5% and likely to die, make a final kamakaze run at a tough enemy to deliver a mega bomb, draw fire, etc...
Gang beat downs - Even though the NERO is closer to enemy tank B, focus your fire on enemy tank A since its damage is critical and about to be pushed over the edge.
Unload power ups - Before picking up a weapons upgrade that would replace my super grenade, go ahead and lob all of my super grenades before picking up the power-up.
Waiting for power ups to cycle - In some games, a power-up changes every few seconds. Could the NEROs learn to wait for spread-fire on one level versus lazer fire on another level? Okay, levels is too easy, how about depending on the situation, what my friends have, etc...
And most importantly, could NERO's be taught to perform "ethical cheats"? By ethical cheat, I mean take advantage of the game engine or environment in a way not intended by the developers. -Not by patching code or using network sniff bots.
Sure, these seem like pretty simple tactics, but YOU try programming this kind of AI. It's next to impossible!
This one gang kept wanting me to join cause I'm pretty good with a bo staff.
MindRover is a PC game where you design the AIs and let them compete.
And before that there were Omega and Carnage Heart (Playstation).
None of those were "learning" games, though.
Galapagos, for the PC, was a game where you had to train an object to navigate challenging terrain. Negative reinforcement came from both the user and from "losing". It was *wayyy* too easy to cause your robotic bug to stand quivering at the edge of a cliff, unable to decide to move anywhere.
All human beings should try to learn before they die what they are running from, and to, and why. --James Thurber
Seaman move rock.
Seaman attack base.
Seaman stop dropping artillery on our own guys!
This will probably be integrated into existing genre's rather than creating a "new" one(and it isn't a new genre either, just an old one taken to a level of sophistication far beyond what we are used to).
This sort of technology would be ideal for a learning AI for real time strategy games and other military genres. It actually could be pretty cool. The computer would basically be learning to play the game with you, thus providing an increased challenge as you progress. Then later, you could just have your AI fight versus another player online and just watch the fun. This too, has already been done. A more recent example would actually be Virtua Fighter 4(a fighting game).
I think we are pretty much reaching the point where genre's are falling away. While they will always have their place, going forward instead of new genres, we're just going to see more an dmore technologies added to existing ones until we eventually get the "uber game" that simulates real life perfectly minus a few "boring" things like laws, etc.
The greatest game of all time will be the one that lets you create a fantasy version of your own life to every detail, and live it. Bit by bit we're getting there, and probably will achieve that within this century.
You are who you are, let no one tell you different. But, never close your mind to a new point of view.
It just hit 1337 downloads ;)
Hey, I ahve a great idea, let us continue to be mired in this mindset of games that teach killing so we can do more real killing & waring then get on our high christian horse and talk about morality as we kill more and give our kids more video games that teach them how to be killers. Great idea.
"In the far-flung future of the year 2000, functional programming has taken over the world and so humans live in an almost unimaginable luxury. Since it's so easy, humans have used robots to automate everything, even law enforcement and bank robbery -- the only job left to humans is to write their robots' control programs." http://icfpc.plt-scheme.org/
I've been thinking for a while that we need some innovation in game AI. It's the one area that really hasn't progressed very much lately. Sure we've got a bit more processor time to throw at it, but that doesn't really achieve that much.
/me has been fragged by a dumb bot on his own team one too many times.
New approaches are needed, and I think machine learning is the way to do it.
Also note that most machine learning algorithms require lots of floating point arithmetic. I'd gladly sacrifice one of my pixel rendering pipelines to it in order to get better gameplay.
Meant HotU (Home of the Underdogs), sorry about that.
For any SF Fans out there NERO was the name of an AI program in Saturn's Race by Larry Niven and Steven Barns.
This sounds like a search problem, (as it always is in AI), whether that be to search for the optimal architecture and weights in a neural network, the most rewarding sequence of state-action pairs, or simply the best strategy out of an endless combination of playing strategies. It does not matter so much how the "agent" learns, but how fast it learns, or adapts to new situations.
However, the problem with training an "agent" off-line, as in for instance Monte Carlo simulations, or on-line as well for that matter, like with temporal difference learning, is that the learnt strategy/policy/whatever is brittle across domains, that is, it fails to generalize to new situations. Of course there are also various degrees of success in generalization when it comes to AI techniques, but usually what you gain in generalization you lose in accuracy.
I hope this is better than Black and White, although I must say a squad of US Marines flinging shit and eating the villiagers might be more entertaining than Yet Another Doom Monster.
Jesus was all right but his disciples were thick and ordinary. -John Lennon
How do I put them together?
Tried putting the magnet link into the location box.. It sits for a while and seems to be searching, but fails to return.
Tastiest. Troll. Ever.
Does anyone remember a research 'game' which was sort of like Pacman but with real motivation. IIRC, the Pacman character was programmed to seek pleasure and avoid pain. Certain pellets were considered positive reinforcements and others were considered negative reinforcements. It ended up having some almost spooky emergent behavior, like hiding in a corner if there were too many negative reinforcement pellets. It seemed to develop responses almost like fear. Stuff like that. I can't recall the details unfortunately. I think it was done as a university project or something, maybe in the late 80s. The idea of generating unpredictable emergent behavior from a relatively simple computer program has stayed with me.
I think that will be the next stage of computer characters: to make them unpredictable even for the programmers. Rule-based learning can get you somewhat complex behavior, but it is all predictable. What we need is genuine example-based learning. So that the resulting behavior would be impossible for anyone to predict and constantly changing and evolving. Of course I am thinking along the lines of various neural network, connectionist architectures. Their unpredictability is generally considered a downside, but for a game the black box aspect seems perfect.
Quite an experience to live in fear, isn't it? That's what it is to be a slave.
For all of you taking about the branch prediction issue AnandTech has brought to light, that is in light of modelling physics in the game world. That can be pretty chaotic all the collisions and what not, but that is not what I am talking about.
I am talking about extracting data from streams of user input. Not modelling physical consequences in the game world, which does involve lots of branching.
Imagine an SPE taking raw video and running a tight algorithm to detect particular movements. This could be an incredible form of control. The PS3 could for example analyze video of say someone playing a fake set of drums and synthesize the music in real-time.
Now imagine another SPE taking that video feed + an EEG feed, and independent of the algorithm to read the player video input, studying it for correlations with respect to when the user makes a mistake with respect to its world, the game rules. It could use those inferences to screw with the player!
The control of neural response should be pretty rapid with respect to changing game factors. You would have to force the computer to make false inferences the way you try to fool a real human player.
I bet if such an EEG input device came about there would be some debate over whether to let console "condition" kids at this level.
Imagine training gnomes in World of Warcraft to go out and do battle! Or better, massive multiplayer online first person shooters that give you battalions to train. If you train them right you win, if not you lose.
Gives people the ability to have AI on their side for once.
Get your Unix fortune now!
In Forza Motorsport for the XBox, I heard that you can have something called a "drivatar", that learns your driving style and can race for you when you don't feel like playing on a certain track.
Circumcision is child abuse.
So, the main site is down. The first working download link posted in the comments resulted in a corrupt file twice, and for some reason I can't connect to the torrent. Any working download links?
> At the end of the day, when all things are factored in, there is simply not enough money in the budget of the average game to make this type of advanced artificial intelligence worth the risk and expense, at least right now.
More to the point, it's in a game company's best interest to ship a brittle AI that people will learn to beat handily after a few weeks of play, so they'll be back to the trough for the next offering.
The game industry's worst nightmare is a game that stays fun for two years. An AI that learns the game with you, and adapts its strategies to yours so that you have to keep innovating, might make that possible.
Sheesh, evil *and* a jerk. -- Jade
I remember reading about how the npc's in this game go on with their daily lives and one of the guards started a fight or stole something and then faught all the other guards. So while the guards were distracted things with the other npc's started to get out of hand. All very interesting.
To all those who haven't graduated already.. =) Wow. This is great. I've been contemplating going to UT Austin for a little over a year now. Seeing something like this from UT students makes me want to go even more..and only a 4 hour drive from home!
I'd like to see a football game that does that - where the other teams in the conference evolve to use strategies that counter your style of play.
A fine is a tax you pay for doing wrong and a tax is a fine you pay for doing all right.
Robot Wars? Program a robot to scan, move and/or shoot via an if-then chain. Then run the program - turn based - against another bot on the battlefield. Also, whatever happened to Bilestoad?... never mind, google had the answer.
The problem with this is that a game is supposed to be fun. How much fun is a game in which you play by turning it on and then leaving the room for a few hours?
I always get interested in "new genres" of video games, especially because most video games these days involve carrying big *cough* guns, shooting people, and having the opportunity to hear abusive one-liners said to women.
Needless to say it's pretty boring for anyone who isn't all that macho. Even Vampire: Bloodlines was spoiled by the offensive scenes and the dull FPS combat gameplay... and that's hard for me to say because there's nothing I like more than sneaking around in shadows and sucking out people's blood.
So when this revolutionary new genre involves... training soldiers to shoot each other with big guns. I mean, wow, how "new"!
If they had more time, the "researchers" would have probably added a few more "sexy" woman screens.
New type ? buhahahaha
Anyone remember Colobot from 2002[?] ?
A story about a learning AI fighting "virtual" wars and no Ender's Game references yet? You're disappointing me.
Justice is the sheep getting arrested while an impartial judge declares the vote void.
Not retarded, just greedy. A lot of news sites these days (especially ones involving video games) will generate ad pages, and insert some variable into the PHP to indicate that you have passed through the advertizement. I don't know if it is the case for this particular page, but it doesn't seem unlikely.
the NeuralBot for Q2? Thats what i thought of as soon as i read this. If anyone wants it, email me (since the authors site went down, it became difficult to get hold of). If i get too many, i might just put a torrent up somewhere. BTW some of the src is included.
"Why can't you stay outside of enemies' firing range? Why do I have to click to keep you at safe distance?"
"Why do you have only three modes of operation; defensive, neutral and aggressive? Can't you make up your own mind when 10 levels smaller, slow, unarmed enemy utility vehicle approaches ALONE.. Could you please destroy it, please?"
"Don't go after units which are faster than you. Shoot them but don't chase them and get yourself killed."
"Hey! There's our unit getting killed near you and you're just standing there! Do something!"
"Isn't repairing your gear like standard thing to do when you don't have orders? Do I really have to tell you each and every time again when you should repair and resupply yourself?"
"Ok, so, let's take a five minute coffee break, smoke if you got em. At ease, soldiers... AT EASE I SAID! Dumbasses! Can't you do anything recreative when you're not fighting?!"
For average gamers a really good AI with combined CNN/SVM/GA/BI/HMM/A* could seem like magic. For others it's just a time to wait until more efficient implementations pop up, or more suited/powerful hardware comes around so as every NPC of a game could be controlled by its own "brain". Until then, there are many dozens of areas of application of these and similar methods.
I am putting myself to the fullest possible use, which is all I can think that any conscious entity can ever hope to do.
There is no need for there to be any real AI in games, only the appearance of intelligence.
When the makers of Halo were testing the original game they discovered that testers consistently rated the AI to be "smarter" if they changed the game such that it took more shots to kill the aliens.
No changes were made to the way the AI engine worked, just that you had to shoot the bad guys an extra couple of times before they fell over - the testers were fooled into thinking the AI was cleverer than it was and they enjoyed the game more.
Very interesting post, the parent.
:)
However - EA has produced lots of massive stinkers for quite a while now.
Yet people are still buying lots and lots of them.
I remember a time (veeeery long ago, yes, I'm an Old Fart) where Electronic Arts games were hits, nearly all of them... sigh.
I'd suggest looking at (and acting like) companies such as Introversion. They produced really good games: Uplink and Darwinia. Both are cheap (I am NOT spending 50 EUR on YAPOS (Yet Another Piece Of Shit)), and something quite new and interesting. Cheap to produce, too: they're just three people
Two million dollars budget for a game... no wonder profits are so low.
Free PC version of ChipWits at http://www.breueronline.de/klaus/chipwits/
The University of Texas is different from Southwest Texas (aka Texas State University) in San Marcos. The only argument about the phrase "Texas University" comes from Texas A&M who claim they were the original Texas university. The A&M students also refuse to correctly reference UT correctly and instead call it TU for "Texas University."
They've got a notice: "Welcome Slashdot" -doesn't that mean that it specifically *wasn't* slashdotted? So none of us have read it? I'm confused.
Don't you mean.. BIZZARO!
This game is based on realistic models of learning and evolution. What is interesting is the tatics, I see it even more as something as a useful model of psychology or sociology. You can make aggressive teams and see how they do against sniper-like teams. (When a very agressive team fought a team trained to avoid the enemy, the teams were about equally matched.) You can see what affect it has to reward soldiers for shooting at each other, to see if they exhibit more "deviant behaviors" such as spinning around in circles. This type of game is definately different than some stagnantly programmed game. And this female is getting tired of people being closed minded enough to think the only way to make a cool game is to "add a few more 'sexy' woman screen."
The Sims with combat robots.
Slashdot social media options: AIM, ICQ, Yahoo, Jabber and Mobile Text. Why no MySpace?
Please say I can run this over the network and have it kill my roommate's retardobots!
It is no longer uncommon to be uncommon.
"In an unusual demonstration of video game innovation with limited funding and resources..."
This sentance presumes far too much. I have yet to see a well-funded, well-resourced game that was innovative. I urge all you game players to think back to the games that truely broke ground (that is, did not simply have the most detailed graphics or the slickest 3d interface.) All of them were made by a small group with very limited resources.
I have not seen ONE big-media title that was in the slightest bit innovative in any grand sense. Slick, yes. Fun, often. Innovative, never.
Make up your own list here for innovative games. I'll get you started:
Tetris. Nethack. Myst. Bolo.
Add your own here.
Will these things be marketable? "Ma, I'm not playing games, I'm training my robo-warrior!"
D'oh people, how stupid can you get? :)
Please dupe this article when there's a cassette available and the C64 version is finished.
Until then it's quite useless.
I remember thinking (not very hard) along these lines some years ago. I was doing a PhD in machine vision and we were using Doom/Quake engines to generate simulated environments for testing robot navigation algorithms.
My thought was that you would train an entity yourself in a series of one-on-one battles or training bouts. These could be staged or otherwise constructed to make mini-games e.g. perhaps testing your entity in predefined scenerios. Once you were happy with its performance you could dump it onto a USB stick and take it around your friends house or upload it to a server for an online game. The main game would put your entity in an arena against a number of other 'gladiators'. They fight it out etc. Online this could allow for 'spectators' who watch the game and potentially even bet on the winner. This might allow for prize money or other revenue stream to be introduced.
-- "Can't sleep, clowns will eat me!"
Are the folks at the university going to give Orsen Scott Card a bit of love?
"We herd sheep....we drive cattle...we LEAD people! Lead me...follow me...or get out of my way!" GEN George Patton
I remember programming AI for a mech combat game. Me and some friends would spend hours programming bots to hunt each other down and then do battle. I'm pretty sure this was on the PS1, so it's not even vaguely new.
Your sig(k) has been stolen. There is a puff of smoke!
The X-box racing game Forza motorsport already has something like this. You can train a "Drivatar" to race just like you. Once it's properly trained, it will take generally the same line as you, take corners the same way... and it also makes the same errors as you.
More info about it here: http://www.drivatar.com/
e.g.e rtech_Talon,00.html.cz
http://www.military.com/soldiertech/0,14632,Soldi
Deleted
Does the US military, with a budget well over $500 billion dollars, really think it needs this? If we can't win wars with all the satellites, carriers, tanks, bombers, etc we have now then I think we should ask for our money back.
Seems like this country is becoming a military worshipping nation like the worst dictatorships.
It is by the juice of the coffee bean that thoughts acquire speed, the teeth acquire stains. The stains become a warning
Another semi-interesting program your robots game is RoboCode http://robocode.sourceforge.net/. It's a web game where you code your robots in Java and set them loose on each other. It's more about teaching Java (and programming in general) then about AI though.
Black and white doesnt work that way.
It doesn't use neural networks, it can't learn tactics.
The creature can only do the interactions that the programmers coded, and all you can do is decide whether or not the creature should do it on such and such target.
Neural networks allow proper varied tactics, non-programmed interactions with the environment, understanding of the geometry of the world and how to take advantage of it during combat.
A Black & White creature would be completely incompetent at Counter-Strike, for example, whereas a neural network can be trained to be competent.
On the Contrary. AI is just statistics on crack (gross simplification). All of the values in my AIs are floating point so they can be sort of true, sort of false, mostly true, etc... I guess I could scale it to integers, but that would be a conceptual pain in the ass - I'd rather just program it in a way that makes sense and throw more TFLOPS at it.
I'm writing a neural net that can learn its own weight update rules genetically. That requires more floating point operations than I can shake a stick at - sigmoids EVERYWHERE.
www.olin.edu
> Unlike most games today that use scripting for the AI, non-player-characters in NERO learn new tactics in real-time using advanced machine learning techniques.
Maybe we could use this on politicians. Even artificial intelligence would be better than no intelligence.
Curry Rice! Spaghetti! A Lunch! Curry Rice! B Lunch!
Eventually we'll turn out a swath of military officers who are very good at running teams of droids. Then we'll need some droids. In the meantime, here's a story.
3 9
http://www.everything2.com/index.pl?node_id=10823
There are so many other aspects in current gaming where this could be applied. In an RPG game it could add another aspect to an AI controlled party where, not only would the levels of the other part members matter, but their combat training/experience would matter as well. In fact, the "level" aspect of your part members could be completely removed and it would make sense. Just think of all the stupid things we do when we are first starting out in an MMORPG or RPG because we are inexperienced (well, that might change for some, but in general we do get smarter about approaching our enemies and know our strength and weaknesses better as we level).
Personally, I think it would be cool to be able to go through a series of training sessions to teach my squad of 5 AI controlled soldiers to do what I want (urban warfare comes to mind). However, it would also add to the "work" factor that is present in some games. I picture an online scenario with me and my AI controlled squad versus another player and his AI controlled squad, where he has put many hours into "training" his soldiers, and he just destroys me. This is probably fair, even in a real world situation, since the better trained troops should win, but it would take the "casual" out of gaming.
I'd like to know if the NEROs can evolve more advanced tactics such as: (list of ideas)
From my reading of their research paper, many of these are not possible because the robots do not "know" their health levels. Their paper shows a neural network where the variable inputs to the network are
1. Enemy radars (5 variables)
2. On target (1 variable)
3. Object rangefinders (5 variables)
4. Enemy LOF sensors (2 variables)
In principle, adding a "health" input is completely straightforward. However, imagine how much longer you may have to train your robots as you add more inputs. Basically, if you wanted the robot to learn to IGNORE its health in performing some strategy, it would have to be exposed to training runs with health at different levels.
And the problem compounds quickly as you consider adding more and more input variables. If you add K new input variables to your network, you've increased the dimensionality of your input space by K dimensions. In the worst case, the "density' of training examples drops exponentially as a function of the number of input variables added.
This is one of the central problems of machine learning: limiting input variables to a small collection (in practice, usually under 20), in order to avoid it taking a geologic age for it to learn something interesting.
For the record, the outputs of the network are "forward/backward", "Left/Right", and "Fire".
Mind Rover
Actually, this only happens in certain games and in different degrees. In Master of Orion 2 (MOO2), the AI generally 'knew' what technologies you had because they spied on you rampantly. On the hardest difficulty and on a huge galaxy, it was common to have over 30 spies sent against you by ONE opponent. On top of that, the only way to 'kill' enemy spies was to have your own spy army on defense and hope your spies went on a counter-intelligence purge.
Vice versa, if you had a technology the enemy did not have and the enemy had technology you did not have, you couldn't trade the technology simply because neither side knew what the other had.
Whereas in other games such as in Civ 2 and 3, no the AI did not 'cheat'. It just used the embassy system. A lot. The game never gave you notices, but if you kept track and kept checking every turn, you could figure out exactly what was being researched, their average rate of research, and even roughly guess whether they were plotting to team up against you.
As far as CS programs are concerned, I think University of Texas at Dallas is bigger and better. The CS faculty at UT Dallas numbers about 60, vs about 35 for UT Austin. Of course there are corresponding differences in enrollment. Also, UT Austin's enrollment is declining, whereas UT Dallas's is increasing, and the average SAT for incoming freshmen to UT Dallas is perhaps the highest in the state at 1230. UT Austin has to deal with all those 10 percenters.
Six score characters.
Brevity being wit's soul
I have enough space.
While UT Dallas is a good school, UT Austin is consistantly rated in the top 10 for computer sciences and UT Dallas is not.
If we accept that our current ground troops warfare model ('shoot if it may threaten you') is flawed, not because the individuals are flawed but because war is confusing and nasty, then the OPs point is made: God help us if we don't find a better way to train robot warriors, unless there are only robot warriors out there on our side.
I think you have to accept - and here's where you can flame my arse to hell and back - that the US's method does suck, as you chaps execute far more friendly fires than other (developed) nation's armies. I don't know why this is, but like the man said, it's got to be addressed, cos the average Brit on the street doesn't like their relatives 'sharing' wars with your guys any more.
Genuinely sorry if that burns you up, but I see his point really well made by your own post.
Justin.
You're only jealous cos the little penguins are talking to me.
.. i just upped the torrent for the neuralBot at demonoid, no signup needed (well, maybe).