Slashdot Mirror
Quake Bots Rock The Prefrontal Cortex
0x4B writes "Some researchers from Vanderbilt University have used id Software's Quake III Arena to test a model of the human prefrontal cortex. The model was injected into the control systems of a Quake Bot, allowing it to flexibly adapt to changing enemy characteristics. The bot was required to identify the vulnerability of its enemies to different types of weaponry through repeated combat trials. Not only are the bots busy shooting each other relentlessly, but you can catch the action by joining the battle as an observer. Source code and Quake virtual
machines are available for download."
Sounds like they were caught playing quake during work hours, and had to make up an elaborate story :)
/. blurb, it's not that cool if all they can learn is what weapon is best.
nah just kidding.. But from the
they day the bot switches to grapple and follows PCs round laughing at their hopelessness all the while grappling them to death I'll know their finished.
Or shooting someone with shotgun cos they'll fall backwards into the lava.
Or camp under the stairs when their bored
etc.etc.etc.
There are places where the networks are not touching,and there are places where they are-Boeing's Lori Gunter
The output from this mod, after it has been running a while:
I frag, therefor I am.
How long until Valve adds this to HL2?
Perhaps this is what is holding up Duke Nukem Forever?
www.eFax.com are spammers
I just tried it for about 5 minutes (seems to me that this should be long enough for beasic learning), but the bots (with Quad colours) just seem to run around and do not react at all, no mather how many slugs I shoot at them.
:( Am I mssing some README?
I tried it with various settings of difficulty
Genius doesn't work on an assembly line basis. You can't simply say, "Today I will be brilliant."
An interesting development, but if they really want to replace human players they still need algorithms for name-calling, complaining and cheating.
Can anyone comment on what types of AI algorithms they use to do something like this (neural networks, markov, etc.) or know of any good resources on the net?
I am so ready for a thinkbot.
When we left our clan Unreal Tournament server going for a week with godlike bots with no timelimit and forgot about them...
Came in for a visit, and it was loque, tamara and another versus poor old Desloch by himself. 769 caps to 212.
Evened the teams a bit and Deslochs team had picked up another 400 caps a day later. What a fighter, eh?
"And that was how I spent my childhood", said the T1000.
If tits were wings it'd be flying around.
We must create an Arnold bot to defend ourselves!!!
Ya, RTFM, the point of this isn't to make them good at kicking your butt. This bot is supposed to be pitted against two special dummy bots, name one and two. Then it can react to the bots chaning their sheild. blah blah blah, it's actually a very cool use of nueral nets, but you wouldn't know that unless you read it :)
http://monkeyserver.com --- weeeeee
Yeah, I figured that one out already after reading the sites.
/. material if you ask me.
I was a bit more enthousiastic, but now I think it's basically a first try, lacking an implementation, not really
Sigh, again one of those decent ideas, but once again ideas do not implemented by themselves.
Move along, nothing to see.
Genius doesn't work on an assembly line basis. You can't simply say, "Today I will be brilliant."
It seems like it is a little unclear to some people how to correctly run our mod (largely due to me leaving out significant details on the download page). It is intended as an experiment involving one learning bot utilizing the PFC model and two special dummy bots. You need to make sure you run the game using the map I made so that the bots have constant interaction. Using the command line listed on the download page, start the game, and then switch yourself over to be an observer. Run the following console command: /idscenario
/addannbot just like you would /addbot to put a learning bot in the mix. For best results, use a bot that doesnt suck. Now switch over to the learning bot's first person view to see the status of its Neural network and PFC layers changing as a result of its perception. Igor, its aliiive.
Now you have some very colorful dummy bots with unoriginal names running around doing nothing. But it gets better....
Use the console command
Now, there is no damage or death in the mod because we didnt want this to complicate the experiment. What you should see is a blue icon appear above an enemy that is hit. A deflected shot will bounce off like it hit the invulnerability sphere. When the bot hits, you will notice the little white box at the top of the network status overlay (upper right corner of the screen) go solid. This signals that the bot got a reward.
Directly under that are two yellow boxes, these represent how much the bot wants to choose each weapon (full is highest). Once the bot learns something you will notice these switching dramatically in response to the characteristics of its enemy. The bottom row of red boxes shows the characteristics of the current enemy (shield color, gun color, position, ID).
Now with all this information the bot tries to figure out what about its opponent is important in deciding how to kill it. The top row of yellow boxes at the left of the screen encodes what dimension the bot is considering, shield color, gun color, location, or ID(name). When the bot picks the right dimension, it can reliably slam its opponent with the right beam color. When it chooses the wrong dimension, it performs miserably until it gives up and explores something else.
Our experiment is set up such that the first correct dimension is shield color. After the bot figures it out the experiment will autoswitch to the ID dimension. When this happens you will see a message appear at the top of the screen in red. When its behaving well, the bot will catch on quickly.
Thanks for checking out the mod, and sorry about being late with this info. If you've got questions a lot is explained in the code walkthrough on the site, otherwise just ask me here. cheers,
-Tamer
Hi dan,
I was wondering what gave you that impression about our mod. Did you run it correctly? Please refer to my post about seting up the experiment. I realize now that I should have put clearer instructions up on my site.
-Tamer
Well basically I read the article with interest but messed up in one particular point: I used /addbot instead of /addannbot. k. I only noticed this after reading the article again this morning.
/idscenario /addannbot
:)
so
seems to do the job nicely
I am not that much interested in the gaming applications, rather than the experimental setup in itself. Basically a hobby from past research.
Being more verbose in the output would really be interesting.
Do you plan to put serveral ANN bots opposing each other? As I can see now, several ANN bots converge on shooting one and two (with some exceptions). I guess it would be interesting to see how 'characters' would be forming in bots when there is e.g. negative feedback.
(e.g. avoid punishment at all cost vs. punishment does not make up for potential gain).
Genius doesn't work on an assembly line basis. You can't simply say, "Today I will be brilliant."
One of the things that made our task hard was the unpredictability of the Quake world. In other words, when the ANN bot misses a shot completely it is unable to get any useful info from it, potentially confusing it. The situation where it shoots the right gun at its enemy and hits the other bot sometimes arose as well. This made locking on to the right dimension difficult for the model because of this unreliable reward schedule. If we gave too much reward for a hit then the ANN bot could get stuck on something random and never give it up. If we didnt give enough reward it would never be convinced that it was doing well. Also not knowing when the ANNbot will score its next hit made things tricky.
We havent given much thought to applying the model to a different task yet. We simply wanted to see if the model could overcome the complexity of the QIIIa world. Since i recently graduated I'm no longer working on research projects as much as trying desperately to get a job.
Speaking of positive and negative feedback, one of my earlier forays into NN's in QUake IIIa was a simple mod where I controlled where the bots aimed different weapons with TD learning (a technique where a NN gets trained through rewards based on the action the agent took). Not having the right rewards produced some interesting bots: When I only gave positive reward for hitting an enemy the bots learned to spin in circles either to the left or right because this behavior guaranteed they would at least get a shot in on someone once a rotation since there were so many bots on the map. To fix that behavior I had to give small negative reward for turning away from an opponent and larger positive reward for killing the enemy. This encouraged the bots to finish the job instead of spinning around whipping rockets at enemies across the room.
Another thing that was wierd is I would start like 10 of these bots training, all equals with the same untrained network, and after a few hours some of them were completely hopeless and others were brilliant. It seems that their experieces during training were responsible for the difference. Some bots learned themselves into corners, eventually expecting their own failure and essentially giving up. Others would get better and better the longer I trained them. It's difficult to try to think of what these experiences were that caused some bots to get depressed and others to succeed.
-t