AI

Comment

Comment by Kristján Friðbertsson on November 10, 2009 at 12:10pm

In case you find yourself in London next month...why not check out AI2009 ?
http://bcs-sgai.org/ai2009/

Comment by Kristján Friðbertsson on November 1, 2009 at 5:38pm

What a great upcoming opportunity for all of us to try that new A.I. variation we've been contemplating, not to mention a new implementation of some AI approach we have just come across.... the IGI Awards. Have you heard about this? This could be a really good chance to either create an AI heavy game concept or to offer AI expertise to some other group(s) creating games to take part in IGIA10.

In case you've missed it, check out the IGIA10 group:
http://igi.is/group/igia10

PS seems I hadn't already done it so I just wanted to thank Hannes Högni for posting here about the Fanfest AI Roundtable, much appreciated :)

Comment by Hannes Högni Vilhjálmsson on October 5, 2009 at 2:15pm

I attended the round table discussion on AI in EVE. As I expected, the focus was on fighting tactics. The opening question was what the players thought of the new Sleeper NPC AI, which are a new breed of opponents, lurking on the other side of the new wormholes. CCP explained that they prefer to deploy new technology (in this case new fighting behavior) through the introduction of new content, rather than by upgrading existing content, so the "old" NPCs, such as various pirates, are still pretty "dumb". One of the things that the Sleepers seem to be good at, is to observe what effect their attacks have on your ships, and then gang on those ships that have the weakest defenses. The players were agreeing that they were impressed by this new AI. However, they also demanded more. Many of them were hoping for NPCs that would start chasing them down through the universe. CCP pointed out that this would introduce an exploit where you would essentially hoard the NPCs into an ambush, harvesting them like crops. This comment brings up one of the biggest questions when deploying any kind of AI for MMOs: Can they be exploited? (and should they be allowed to be?). Only towards the end of the discussion did CCP touch on other kinds of AI, and there was general consensus that more "background / wildlife" AI would be fun to have. Especially with the upcoming planet improvements, you should be able to observe more life around inhabited planets (space shuttles, satellites, lower orbit space stations, etc). This kind of populating of worlds would need to be accomplished through procedural means, so one could speak of an "intelligent" system for placing such activity around the world, in addition to a system for controlling its behavior. Perhaps there is an opportunity here for some cool projects. There was no discussion about AI in Incarna (previously Ambulation and Walking in Stations), so I did not review CADIAs work on social behavior at this time.

Comment by Kristján Friðbertsson on October 2, 2009 at 10:38pm

Just pointing out to those not already in the know.. :-)

Fanfest at 11am tomorrow, Saturday 3rd Oct, has a roundtable discussion on A.I. in Eve. Should be very interesting and enjoyable and if I find myself still under the weather I will of course demand all your transcripts and recordings from the event ;o) otherwise - see you there.

Comment by Ævar Örn Kvaran on September 7, 2009 at 9:51pm

This should interest you all if you haven't seen it. It is really cool, and well I must admit a bit scary.

http://www.engadget.com/2009/08/22/video-robot-hand-shows-off-amazing-dexterity-speed/

Comment by Ævar Örn Kvaran on September 7, 2009 at 9:40pm

Of course as with other technologies, it can be used for good and evil. But of course if your definition of evil is total lack of empathy and blindly following of orders regardless of what they are... then yes robots and AI is evil. But you really have to have the capacity of intension of your own free will to be concidered evil.
We have to remember as well to divide the definition of AI into multiple fields, and seperate true AI from other fields of AI.
For example decicion trees do fall within the field of AI and can be achieved in excel.
I do not believe that true AI can be achieved without studying neurons in the brain and the von neumann model might not be capable of replicating this functionality properly.
I know that allot of AI funding has been military founded and in fact the military has been really pushing AI for military purposes, I.E. Killing people and blowing up stuff, but that doesn't make the field necessarily dangerous and definitely does not make AI any more evil than airplanes, spacecraft and the like.

I know you are not if not all being a bit sarcastic but because of allot of movies and science fiction a whole lot of people are afraid of AI.

The only thing we can hope for and fight for is to minimize military use of advanced technology for killing, and hopefully a world without military all together in less than 100 years.

Comment by Deepa Iyengar on September 7, 2009 at 4:22pm

hehe so did you make your evaluation of "Kismet equals cute" in that state?

Comment by Finnur Magnusson on September 7, 2009 at 4:19pm

Don't remember, I was quite drunk at the time

Comment by Deepa Iyengar on September 7, 2009 at 4:10pm

Creeeeepy robot, IMO, from the looks point of view. It´s hard to make neoteny convincing. Hannes has collaborated with Cynthia Breazeal (creator of Kismet) in the past. Who did you meet?

Comment by Finnur Magnusson on September 7, 2009 at 3:53pm

I once met a guy in Boston who worked on Kismet. I think he's cute (the robot that is)

• ‹ Previous
• 1
• 2
• 3
• Next ›
• Page