GDC 2004 Impressions?

I presume nobody got turned away from the dinner since they had a fixed number of invitations this year instead...

I got back and was immediately put in crunch mode. In fact, I have to compose this post across several builds.

My favorite AI session (other than my own roundtables of course) was the "Beyond Finite State Machines: Managing Complex, Intermixing Behavior Hiearchies" by Michael Mateas and Andrew Stern. You can go to http://www.cmpevents.com/GDx/a.asp?option=C&V=11&SessID=2181 or www.interactivestory.net or egl.gatech.edu for more information.

Actually, we had to turn a number of people away (unfortunately) from the 7th Annual AI Programmer''s Dinner. We sold out the very first day, and probably turned away half-again as many people (which would have brought our total up to about 70) over the following days. And then, we actually had one no show. So, we had 49 seated for an outstanding feast, coupled with great AI discussion over wines and stuffed chicken, baked salmon and prime rib roast, not to mention devilish deserts.

Since the GDC will be moving to San Fransico next year, we will have to start anew, the process of finding a venue for the dinner.

More later, I have to get back to work.

Eric

Molyneuex's stuff was really cool. I definately want to go grab Fable and The Movies just so I can dork with them. The story about the wedding and tatoos in Fable was a riot!

The Mateas/Stern "game" is, indeed, more of a test bed. It could possibly lead to some more interesting things down the road. I would have liked to see them get to more of a multi-layered level with the behaviors... perhaps if they had 2 hours instead of one?

Brian Reynolds did a cool talk that I hope resonates outside the AI domain... the one on "How AI enables designers". Using the progression from the original Civilization through "Rise of Nations", he showed one particular track of how solving an AI issue generated new design potential. That has been my mantra - that AI should be done VERY early if not coupled with the design process. Unfortunately, when games are designed by artists, that isn't always the case.

The roundtables were interesting but they have really begun to attract a lot of noobs. I would like to have one really high-level roundtable and see what shakes out. I would have to agree, however, that a lot of the changes of late seem to be scope-based rather than technology based. That just falls into more of an engineering thing rather than pure AI discussion.

One funny point was the guy who talked about leading soldiers from Far Cry into the water. They get stuck in a "tread water" state and just stay there. He would lead whole squads into the water, leave, and then toss a grenade in. Wow... the things that sneak through the testing department, eh?



Dave Mark - President and Lead Designer
Intrinsic Algorithm - "Reducing the world to mathematical equations!"

[edited by - InnocuousFox on April 1, 2004 2:57:02 PM]

Bob? Is that you?

Anyway, that was a good point about Beyond FSMs. One of the things that we were talking about is how they came up with arbitrary "priority" numbers for deciding what trumps what. While some of that can be seeded, we all felt that there should be some fluidity to it. Using 10 numbers limits the designer at some point... eventually you will paint yourself into a corner of deciding which "8" is more important.

Nifty concept but incomplete.

Dave Mark - President and Lead Designer
Intrinsic Algorithm - "Reducing the world to mathematical equations!"

Perhaps someone could post a precis of the concept in a new thread and we could discuss its pros and cons, with the possibility that someone might come up with a novel (and better) solution?

Timkin

Speaking of the GDC...and the roundtables... I found out tonight that I somehow rated a mention during one of the roundtables... but that it might have been decided that I don''t actually exist! Okay, you''ve found me out... I''m just a well written perl script and my try pseudonym is in fact, Timkin.pl!

Just call me Timbot!

quote:

---

Original post by Timkin
... but that it might have been decided that I don''t actually exist! Okay, you''ve found me out... I''m just a well written perl script and my try pseudonym is in fact, Timkin.pl!

Just call me Timbot!

---

Perhaps its time for Timkin to show up in person at one of these conferences and have drink with us, so that we can finally put these rumors to rest?

Eric

Heck, Eric... he came up once or twice in the ones YOU ran... but we never doubted his existance. That means he came up ELSEWHERE as well!

Dave Mark - President and Lead Designer
Intrinsic Algorithm - "Reducing the world to mathematical equations!"

quote:

---

Original post by Timkin
Speaking of the GDC...and the roundtables... I found out tonight that I somehow rated a mention during one of the roundtables... but that it might have been decided that I don''t actually exist! Okay, you''ve found me out... I''m just a well written perl script and my try pseudonym is in fact, Timkin.pl!

Just call me Timbot!

---

That was one of our roundtables, Timkin! We took special pains to mention you at the "AI for Beginners" roundtable and this forum too.




Ferretman

ferretman@gameai.com

From the High Mountains of Colorado

GameAI.Com

quote:

---

Original post by Ferretman
We took special pains to mention you at the "AI for Beginners" roundtable and this forum too.

---

So I''ve been promoted from Noob to Beginner? Awesome!

quote:

---

Original post by Timkin
So I''ve been promoted from Noob to Beginner? Awesome!

---

It''s conditional on you actually showing up one of these years.



Dave Mark - President and Lead Designer
Intrinsic Algorithm - "Reducing the world to mathematical equations!"

While I enjoyed the roundtables I attended as a whole, I found the AI ones to be frustrating.

It seemed a lot of the things being said were argumentative, and its not really a practical medium to hold debates as such. Well at least for the size of the roundtables.

Perhaps the smoothest roundtable I attended was one on multiplayer tips & tricks of the trade.

- - - -

My impression is similar to InnocuousFox''s regarding the level of knowledge about AI that attendees had. Except I would say that of those who spoke, only two or three seemed to have studied AI in-depth. However I think it paints too bleak a picture to say that the roundtables attracted a lot of noobs.

Rather they attracted a lot of people who weren''t up for discussions, and so listened instead. I wouldn''t guess where they are at in the AI arena.

Although perhaps the environment would have been friendlier to discussion if a couple of the more outspoken individuals didn''t dominate the floorspace and circumvent the moderator in acquiring the floorspace to speak.




regards,

GeniX

www.cryo-genix.net

Most of the really interesting discussion actually happened at the AI dinner :-)

That was quite enjoyable, and the organisers should be thanked again. I hope next years venue proves just as good.



regards,

GeniX

www.cryo-genix.net

quote:

---

Original post by GeniX
Although perhaps the environment would have been friendlier to discussion if a couple of the more outspoken individuals didn''t dominate the floorspace and circumvent the moderator in acquiring the floorspace to speak.

---

No kidding. We are, perhaps, thinking of the same person in one major instance? While he was knowlegable, it was a bit over the top at times. *sigh*



Dave Mark - President and Lead Designer
Intrinsic Algorithm - "Reducing the world to mathematical equations!"

quote:

---

Original post by GeniX
Most of the really interesting discussion actually happened at the AI dinner :-)

That was quite enjoyable, and the organisers should be thanked again. I hope next years venue proves just as good.

---

I have to agree there... although I''m trying to figure out who YOU are.



Dave Mark - President and Lead Designer
Intrinsic Algorithm - "Reducing the world to mathematical equations!"

I think I would be thinking er.. of that person, yes.

I would probably be one of those noobs you speak of. While I have implemented neural networks, some pathing algorithms and dabbled in some other stuff, I certainly don''t keep up to date on AI. Even though I work on a game which could involve quite challenging team AI, I haven''t had much to do with the AI in it either.

- - - -

One particular discussion I had, which was quite good, was with a student on the way over to the AI dinner. Having been to the talk on APF''s, I suddenly started getting some vague ideas on how they could be applied to individual agents when some group behaviour is nessecary.

We had quite an interesting discussion on whether it could give good flocking dynamics (with birds still flying through each other, but its the general look of it which IMO counts the most for games). I think some good techniques can be used to make the flock, or sections of it swoop down in nice arcs like birds often do and then rejoin the flock. The flock can easily be given direction and a visual sense of purpose with well-placed APF''s.

Of particular interest were some of the sneakier techniques proposed at the end of the talk on APFs with tangents to the field, agents possibly creating small temporary APF''s around them or in their wake.

The discussion quickly moved on to RTS games involving groups of units and what sorts of results might be gotten from either the groups, the world, or the enemy being able to generate APF''s that may not even be based around a point.

I was meaning to do some test implementations when I got home, but you know how it is. The energy and enthusiasm wanes as normal life returns





regards,

GeniX

www.cryo-genix.net

quote:

---

Original post by GeniX
Most of the really interesting discussion actually happened at the AI dinner :-)

That was quite enjoyable, and the organisers should be thanked again. I hope next years venue proves just as good.



regards,

GeniX

www.cryo-genix.net

---

Thank you! As one of the organizers I appreciate that...the dinners are just plain fun and it''s a pleasure to put them together each year.




Ferretman

ferretman@gameai.com

From the High Mountains of Colorado

GameAI.Com

quote:

---

Original post by GeniX
While I enjoyed the roundtables I attended as a whole, I found the AI ones to be frustrating.

It seemed a lot of the things being said were argumentative, and its not really a practical medium to hold debates as such. Well at least for the size of the roundtables.

Perhaps the smoothest roundtable I attended was one on multiplayer tips & tricks of the trade.

- - - -

My impression is similar to InnocuousFox''s regarding the level of knowledge about AI that attendees had. Except I would say that of those who spoke, only two or three seemed to have studied AI in-depth. However I think it paints too bleak a picture to say that the roundtables attracted a lot of noobs.

Rather they attracted a lot of people who weren''t up for discussions, and so listened instead. I wouldn''t guess where they are at in the AI arena.

Although perhaps the environment would have been friendlier to discussion if a couple of the more outspoken individuals didn''t dominate the floorspace and circumvent the moderator in acquiring the floorspace to speak.




regards,

GeniX

www.cryo-genix.net

---

I have to agree wholeheartedly and as one of the moderators I''ll even come right out and say it--I was dang close to losing control of the discussion a couple of times, especially that first day. I apologize for that; the folks involved were , as you say, rather vocal ! I haven''t usually let that happen in the past, and I promise that should we do the roundtables again I''ll keep a tighter rein on things.

That said I thought the roundtables were fascinating this year...we really covered a lot of topics. The potential of hardware actually hitting the market was interesting, and I was quietly surprised at the lack of the middleware discussions that there have been in the past. Either developers are evaluating the products out there and keeping it quiet, or they''re just not interested at present.....






Ferretman

ferretman@gameai.com

From the High Mountains of Colorado

GameAI.Com

I attended your tutorial, Ferretman, and will agree there were some very interesting things mentioned.

I dont know if there has been a thread around here on it, so I will mention it anyways. Those interested in the AI hardware should keep tabs on http://www.aiseek.com for when it comes online.

I am most interested to see how they get around some of the obvious implementation issues - are there many AI algorithms that can be abstracted enough to be a useable generic solution across the board?

Some objections were cited by one of the more vocal participants in Ferretmans tutorial to a representative of this Israeli group doing the AI acceleration hardware. Amoung these was the need to transfer info to the card in great quantities, and then retrieve the info at a fast enough speed. At the time I remember thinking that this objection had sufficient grounds to probably render the hardware impractical on current systems.

However, I have recently been awakened to the up-coming PCI-Express technology. With some of the stats given for the x16 slots (initially one on a board - for the gfx card), I dont think the data transfer for and AI accelerator is an issue (provided it can be accessed via an x16 slot).

Anyhow, I am quite keen to see what solution is proposed. Whatever it is, it will be slow to catch on if anything comes of it. The industry (heres my great generalisation) seems to be quick to optimise and improve hardware solutions, but slow to welcome new ones.




regards,

GeniX

www.cryo-genix.net