What is Digg?Sponsored by Dragon Age: Origins
Can't get enough Dragon Age: Origins? Check out new footage. view!
DragonAge.BioWare.com - EA presents BioWare's new dark fantasy epic Dragon Age: Origins. '9/10' from Game Informer.
60 Comments
- GCarden, on 12/02/2008, -2/+31If we use video game AI in our robots, then I think we're going to have a lot of robots that keep walking into walls.
- john2kx, on 12/02/2008, -2/+20OK, thanks, I'll be waiting here. In the meantime, learn how to use the word "your".
- offrdbandit, on 12/02/2008, -0/+16Or worse, robots than can stick a gun around a corner and headshot you from 100 yards without looking...
- inactive, on 12/02/2008, -1/+14Allen Iverson is washed up. Why would we research him?
- ShyGuy91284, on 12/02/2008, -0/+11Things are better than they used to be. I remember back when they had the bot mods for Quake 2 that you'd download it, and have to run around the map a bunch and let the bot learning utility learn the map from that.
- Yazilliclick, on 12/02/2008, -1/+11No, games certainly aren't where more advanced AI is really going to show itself any time soon. Developers have made it abundantly clear that they do not believe it's worth the time to invest in. It's not a very good feature to sell the game with. The only way this could possibly come about would be a third party developer create AI libraries that are generic enough to be used in games, but that would be a monumental task to create a good AI that's generic enough and valuable enough to make money back off of licensing.
- jawshoeuh, on 12/02/2008, -1/+10BURN
- Iwantawii, on 12/02/2008, -0/+9What if the AI in video game characters became aware of the "Matrix" in which they were programmed and allowed to be?
- sponeil, on 12/02/2008, -0/+7When it starts camping spawn points and insulting your mother, you'll know it's beaten the Turing test for FPS games.
- Sornos, on 12/02/2008, -0/+7Well yeah. I've seen some great AI in games. Specifically, in RTS games. An AI enemy in Command and Conquer 3 really threw me off a while back. I had stationary gun placements guarded by several units by a bridge that was the only way into my base. The AI tried a full frontal assault on my units but they easily killed. So it sent units single file around the outer edge of the bridge where my turrets couldn't get at them. These units ignored everything it passed by until they reached the rear of my base where they started to tear up my power plants to disable the turrets. I was shocked.
Granted this is just a simple example, but it shows problem solving. AI in games is a practical foundation because there is a real world demand for functional AI. Once the foundation is laid, you can set out to really fine tune the process. - austin63, on 12/02/2008, -0/+5No, but good AI research would make for better games.
- radu79, on 12/02/2008, -0/+5I think investing time in the AI is a good thing, and players like it.
I am the developer of a small, indie MMORPG, without about 4K players.
As in any other game, most of our AI is dumb, just walking around, and attacking players or not based on some simple tests.
However, a while ago I added a different monster, that hunts players at night. This monster is much more intelligent than the average monster, for example will try to stay in dark areas, will not attack people that he thinks are too strong (unless if he is hungry), will try to flee if hurt and so on.
Well, it turns out that many months after implementing it, he is still popular, and players still talk about it and try to hunt it. - Yazilliclick, on 12/02/2008, -0/+4Do you have a point or even know what would be an advance in AI ?
- iDoraemon, on 12/02/2008, -2/+6No. The last thing I need is for A.I. to emulate the behaviors of gamers I come across in online gaming.
You know what I'm talking about, screaming pubescent boy yelling racist profanities while teabagging the corpse of my online character. - AncientWeird, on 12/02/2008, -0/+4I want it the other way around. I want AI research to breathe new live into video games. I'm sure you all feel the same.
- Miketwo345, on 12/02/2008, -0/+3AI, for a long time, has simply been advancements in search algorithms. This is important, but the problem is that it only works in a well-defined space, like Chess or video games.
The real world is noisy. Sensors are noisy, things move, actions are not precise. The fundamental assumption of traditional AI was that computers would eventually be fast enough to model the real world, and then the model could be searched with advanced AI to find the correct actions.
It turns out that modeling the world is a much harder problem. So a new wave in AI, called behavior-based, was started by Rodney Brooks a few decades ago. "The world is its own best model" was the catchphrase, and most advancements in AI since then have been in using the complexity of the world to make your robot complex.
(What's the point of this background info?) I don't think AI in video games will lead to better robots, because video games are a closed model, whereas the real world is infinitely more complex, and requires a different "type" of AI. - Terasiel, on 12/02/2008, -0/+3Can Hockey breathe new life into cold resistance research? Can Football breathe new life into blunt force trauma research? Can video games breathe new life into AI research?
...What do you think they've been doing all this time? The basics people. They've been further developing the basics and foundation-level of this technology. Duh. - Shakermaker, on 12/02/2008, -1/+4SICK burn!
- nbcaffeine, on 12/02/2008, -0/+3Yeah, but was there ever more fun had than Rockets and Rail Guns in Q2DM1?
- redpixie, on 12/02/2008, -0/+3Or hit a "glitch" in the environment and you get stuck with a stick, trying to get it out of that tree it's just shot up.
- inactive, on 12/02/2008, -0/+3yes
http://www.youtube.com/watch?v=YnbKOboT5wQ - gnotDigger, on 12/02/2008, -0/+2YES WE CAN Arificial/Intelligence 08!!!
- phoi, on 12/02/2008, -0/+2Probably the same thing that would happen if a fish became aware of the fact that it was in water.
- tm13lke, on 12/02/2008, -0/+2What game is it? I'd like to check it out
- inactive, on 12/02/2008, -1/+3Every FPS fan knows player vs. player is where the real intelligence is. That said, all robots should be controlled live over the internet by geeks wearing headsets.
- hinmanj, on 12/02/2008, -0/+2Hah, I just did a research paper on this. My take on the topic: Game AI will _always_ be about smoke and mirrors. On the other hand, as processors keep getting more and more cores, developers are starting to get more power than they know what to do with, so they can actually develop real systems that mimic academic AI. In a few years I think the 2 will entangle, and contribute to each other. But game AI will never really develop some crazy technique (before academics) that will be crazy useful for academics.
- Wisgary, on 12/02/2008, -0/+2I remember that stuff... that was when the whole concept of multiplayer gaming against the AI was unheard of, bots were being created left and right for the same game, some with better AI than others (Eraser kicked ass).
- Yazilliclick, on 12/02/2008, -0/+2I agree that good use of AI can greatly add to gameplay however your example is rather simplistic form and decision model by the sounds of it and does not really incorporate any form of learning. I applaud that you took it further than most developers see fit to but unfortunately most don't see it worth the resources as AI doesn't show well in marketing media and thus resources are spent on fancy graphics instead.
- PleaseJustDie, on 12/02/2008, -0/+2When I play against bots its because I want a challenge, without having to deal with lag/latancy/children, and not feel like I'm playing an army of 12 year olds.
- DragonForce4, on 12/02/2008, -0/+2as long as they don't take examples from Call of Duty: World at War
- morninglorii, on 12/02/2008, -0/+2I'm not disputing what you're saying here (and if you're right, I think that's awesome), but unless the AI enemy did something like this in multiple games, it may just show a luckily-picked random strategy, rather than problem solving.
- DrNafork, on 12/02/2008, -1/+3What if humans became aware of "God" and allowed to be?
- DragonForce4, on 12/02/2008, -0/+2what's your game called?
- inactive, on 12/02/2008, -1/+2AI for practical application at least is only good for getting things done we want done that we don't want to do ourself - to get things done faster and more accurately. It's not so much a need for AI as it is for a mindless drone.
- Scrappy1850, on 12/02/2008, -0/+1i play CS:S and TF2 and i dont notice it. i play L4D also, dont hear it. not that it isnt there, its just not like halo.
- toasterweasel, on 12/02/2008, -2/+3jo momma
- FredFredrickson, on 12/02/2008, -0/+1I might have bought into this kind of thing when I was a kid reading Game Players magazine... but 15 years later, I think it's pretty safe to say that top level AI research is going to come from a lab, and not a bunch of game programmers.
- radu79, on 12/02/2008, -0/+1Well, I do plan to add some basic learning to that specific monster. For example, to keep track of the players that kill him often, so that when he is fighting one of them to try to use more healing, or just run when he sees them.
Now, if we are talking about real AI, such as neuronal networks, that's a little too hard to implement (properly), and I think the players would rather have me focus on other things. - SirChasm, on 12/02/2008, -0/+1AI research has stagnated because AI cannot overcome a very fundamental problem - it cannot scale well. AI that performs well in a computer model fails terribly when applied in a real-world situation. No matter how good you make a computer game AI, it's going to run into this very same problem when you pluck it out of its well-defined computer world and into our ill-defined world.
So no, video games can't breathe new life into AI research because they're not addressing the fundamental problems that are choking the AI research in the first place. - PleaseJustDie, on 12/02/2008, -0/+1I was always partial to the Reaper Bot from Quake 1.
It would learn any level by actually running around and every time you killed it, it got harder, it would play very similar to many human players, jumping, even grab the invulnerability then jump in the water with the lightning gun.
I just wish it could have handled Team Fortress. There were no good Team Fortress bots. The one I had kept drowning, maybe 2 out of 3 times, trying to follow the waypoint in 2Fort5 that enters the enemy base through the water. - sslemon, on 12/02/2008, -0/+1Entertainment needs more fake things?
- Vanor, on 12/03/2008, -0/+1First off, calling what a computer program does in a video game "artificial intelligence" is an insult to the term and misses the mark by a wide ***** margin. *****, it doesn't just miss the mark, it's like the arrow left the bow, did a U-Turn and then flew back and hit the bowyer in the chest.
AI isn't really a matter of "intelligence" but of sentience. For an AI to be what it is abscribed as, it must be self-aware. The thing has to be aware of what it is, and has to endeavour to learn more about itself and the world around it indenpendent of any outside influence. While I won't rule out the possibility that video game programming might have some minor contribution to AI research, it's very slight.
It's also equally likely that with the advent of holographic harddrives and the eventual research that will be done in that area, is what will ultimately give rise to true AI programming. This is mainly because Dr. Karl Pribram's research into how the brain works concluded that it functions in a holographic manner. This would explain how our brains are capable of storing so much memory, and more importantly how we can access it so quickly and so totally. It also explains how people who lose portions of their brain can still relearn some the functions that they may have originally lost, since nonlocality is also a property of holograms.
Long-story short: Don't expect video games to make any significant contributions to genuine AI R&D. The real progress is gonna be made with holographic research and applications. - lennybird, on 12/02/2008, -0/+1Then came PodBot...
- radu79, on 12/02/2008, -0/+1BTW, here is a video of a player trying to engage it. Notice how he runs after he determines that he can't win:
http://www.youtube.com/watch?v=iQuP5XPjhsk (the video is made by a player, so it's not official in any way). - yamikaze, on 12/02/2008, -0/+1Dugg for proper use of "lie" and "lay" in the article.
- radu79, on 12/02/2008, -0/+1It's called Eternal Lands:
http://www.eternal-lands.com - heavystone, on 12/02/2008, -1/+2No. There is no such thing as AI in games. They are all mapped to behave in a strict pattern in different circumstances. In other words it's simulating AI, but it's not really AI. "Real" AI would need to adapt and find solutions on it's own according to situations.
- Aurabolt, on 12/02/2008, -0/+1Do you think God ever says that to himself, about us?
- LukeD, on 12/02/2008, -0/+1FEAR is, as the article says, quite an interesting landmark in game AI, and it could well be a contributing factor in the recent interest from academia in feeding ideas back and forth between ourselves and industry. In Europe, we have the AI and Games Research Network (www.aigamesnetwork.org), and later this month the IEEE 2008 Symposium on Computation Intelligence in Games will be held in Australia. I wouldn't go as far as to say that new life is being breathed into the field, but certainly from the researchers point of view, we are increasingly starting to use games as good sandboxed environments to test theories and run experiments, so from that angle games are starting to come into their own for AI research.
- tm13lke, on 12/02/2008, -0/+1Really? Go play CS or Team Fortress and try to say that again. You'll get those kind of gamers everywhere.
-
Show 51 - 62 of 62 discussions
© Digg Inc. 2009
— Content created and posted by Digg users is