Jump to content

  • Log In with Google      Sign In   
  • Create Account

Banner advertising on our site currently available from just $5!


1. Learn about the promo. 2. Sign up for GDNet+. 3. Set up your advert!


Emergent Intelligence


Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.

  • You cannot reply to this topic
73 replies to this topic

#21 BrianL   Members   -  Reputation: 530

Like
0Likes
Like

Posted 19 August 2004 - 05:33 AM

Back on topic:

Prozak, has your question been answered yet? If not, can you provide any more context about what you are interested in? What kind of 'problems' are you interested in solving?

Depending on the problems you are interested in, there may be a better field to look at than emergence.

Sponsor:

#22 RPGeezus   Members   -  Reputation: 216

Like
0Likes
Like

Posted 19 August 2004 - 06:42 AM

Gee, I wish I had joined in on this one sooner.

Prozak: I think InnocuousFox said it best with:

Quote:

Emergent behavior" is really just a fancy way of saying "a coincidence that really looked cool"


As far as most _games_ are concerned he's bang-on. In your game you just need a way of presenting an illusion of intelligence (or cooperation) to the player. Graphics programmers fudge lighting, Game AI programmers fudge intelligence. :)

You wanted examples, so I give you mine. :)

Not long ago I wrote a tank-battle game in which the enemy tanks choose (mostly randomly) from a set of actions. Each of these actions is comprised of a set of other actions (which are selected sort-of randomly).

For example, a tank could decided to turn towards the player, and move forward at the same time. This makes the tank look like it's chasing the player, although the tank itself has no concept of chasing. Turn and move forward are separate actions that could be applied at the same time.

When a tank bumps in to an obstacle, it can take a number of different actions-- back up, back up and turn, just turn, wait, turn towards the player, shoot, etc..

Everything had to be kept really simple because the game needed to run on a slow cell-phone-- every spare drop of processing power needed to be saved for the 3D wire-frame graphics. lol

People who played the game we're generally happy with the different tank opponents, and found these opponents very tough yet still 'tank-like'.. Many people commented on how smart the tanks were for using obstacles as cover; they would duck out, fire, and then retreat back behind the obstacle always keeping the players line of fire blocked.

This particular behaviour was coincidence. The simple rules that the tank sort-of-randomly decided to follow ended up making it LOOK like the tanks were using cover, but in reality they had no concept of cover, or blocking, or hiding.

Technically what was happening was this:
-Tank would randomly choose to chase player (or drive around).
-Tank would bump in to a wall.
-Player would turn to lock on to enemy tank because a stuck target is easy to aim at.
-Enemy tank would back away from wall.
-Enemy tank could not move forward much because of the wall, therefor enemy tank spent more time turning towards player.
-When player was within enemy tanks randomly generated firing arc, tank would fire.
-After shooting enemy tank would randomly pick a new action.
-If new action resulted in tank moving forward, it was likely to bump in to a wall.
-Bake in preheated over at 350 degrees for 50 minutes or until evenly brown and you now have one tank that looks like it's using cover.


I was very happy with the results.

Good luck!
Will



#23 tonyg   Members   -  Reputation: 284

Like
0Likes
Like

Posted 19 August 2004 - 08:08 AM

I also thought 'emergent behaviour' was behaviour not planned but a result of other rules which, basically, would be by coincidence. I agree that 'co-operative agents' would be something different altogether (i.e. some king of planning).
AI Game Programming Wisdom has a couple of articles by William van der Sterren...
Squad Tactics : Team AI and Emergent Maneuvers
Squad Tactics : Planned Maneuvers
which might be of interest.
I'm still not sure whether Emergent Behaviour, rather than genetic algos, is the right area for Prozak's query.
Unless I've misunderstood it... which is very possible.

#24 Timkin   Members   -  Reputation: 864

Like
0Likes
Like

Posted 19 August 2004 - 01:44 PM

Personally I don't believe what Dave quoted above supports his flippant and terse (his words) comments earlier in the thread and taken in the context of the literature in the field of emergent behaviours (written by many people, many of whom are academics and many of whom are not), means exactly what I wrote above.

Dave, perhaps you did mean to say (albeit in other words) what I was saying, but your flippant attitude once again came across as dismissing an entire field of research, developed by thousands of very intelligent people working inside and outside of academia. Personally, I think you've incorrectly represented the sources you've quoted and perhaps you should re-read both my post and theirs.

To those others developing this thread into a flame war of 'academia' versus 'the real world ™'... please stop. This is NOT an 'academics definition' versus a 'game coders definition'. This is simply about understanding what emergent behaviour is and what it isn't.

If it is an academic vs TRW argument, then I must be in trouble, since I have been an academic, I write games (and Game AI) and I work in industry doing AI. Does that mean I don't know what I'm talking about because I'm not in one camp or the other?

Finally, Dave, you should be careful about attacking APs assuming they're the same person. They're typically not and they weren't in this case...

Please people, keep it civil.

Cheers,

Timkin

#25 IADaveMark   Moderators   -  Reputation: 2655

Like
0Likes
Like

Posted 19 August 2004 - 03:38 PM

1) My quotes simply came from a Google search for "definition emergent behavior game". They were just samples of what I found meant to show how your allegedly crystal-clear definition of the topic is a little more muddied that you imply. (Interestingly, other people came in to post in support of my definition - or rather my summation of the definition as held by many in the industry.)

2) If I "came across" as "dismissing an entire field of research" then perhaps the person interpreting it is a bit oversensative to the subject? I can't seem to find how I was dismissing anyone or anything - just merely passing along an observation of how the term is often used in our industry.

3) A careful review of the thread will show that my "attack" on an AP was conducted prior to the 2nd AP post. Even further dilligent research will show that I agreed with the post of the 2nd AP and even lamented that he/she/it was anonymous as well. I was well aware that they were different people.
Dave Mark - President and Lead Designer of Intrinsic Algorithm LLC

Professional consultant on game AI, mathematical modeling, simulation modeling
Co-advisor of the GDC AI Summit
Co-founder of the AI Game Programmers Guild
Author of the book, Behavioral Mathematics for Game AI

Blogs I write:
IA News - What's happening at IA | IA on AI - AI news and notes | Post-Play'em - Observations on AI of games I play

"Reducing the world to mathematical equations!"

#26 JonahK   Members   -  Reputation: 122

Like
0Likes
Like

Posted 19 August 2004 - 06:02 PM

Interesting debate...

My take on it is that 'emergent behavior' is not a coincidence, but it's not some grand mystery either. The agents are simply doing exactly what they were told. The behavior looks complicated, but in reality it's not, because the rules that generate the behavior aren't complicated. The various actors in the system might appear to be working together, but unless the rules explicitly tell them to work together, they're simply working for themselves. It's just that at one particular point in time they happen to have the same goal. Not by coincidence, mind you, but by design. They're just doing exactly what the programmer told them to do, even if the programmer doesn't realize he told them to do it. It happens all the time in computer programs, we just usually call it a bug :)

As an interesting sidenote, the novel 'Prey' from Michael Crichton has some interesting things to say on the subject of emergent behavior, even though it's just a so-so book...

#27 Timkin   Members   -  Reputation: 864

Like
0Likes
Like

Posted 19 August 2004 - 06:29 PM

Quote:
Original post by InnocuousFox
A careful review of the thread will show that my "attack" on an AP was conducted prior to the 2nd AP post.


I was referring to your most recent response to an AP, in which you made sarcastic remarks about a lack of credentials. If this wasn't an attack, then I've misinterpreted it. My apologies if that is the case.

Quote:
Original post by InnocuousFox
They were just samples of what I found meant to show how your allegedly crystal-clear definition of the topic is a little more muddied that you imply.


I made no such claims as to the clarity or exactness of my definitions. I merely stated that your definition was wrong based on the accepted literature in the field and offered a very brief description of emergent behaviours versus cooperative behaviours. Go back and read my original post and please point out where I claimed that my definition was 'crystal clear', if you believe otherwise.

Quote:
Original post by InnocuousFox
(Interestingly, other people came in to post in support of my definition - or rather my summation of the definition as held by many in the industry.)


You keep mentioning this 'many in the industry', yet you had to pull a few quotes off google - which you clearly misrepresent through your own lack of knowledge of the depth of literature in this field - and claim that these quotes support your comments because someone else agreed with you.


Quote:
Original post by InnocuousFox
I can't seem to find how I was dismissing anyone or anything -


To quote you form your first post in this thread:
Quote:

"Emergent behavior" is really just a fancy way of saying "a coincidence that really looked cool"


This is particularly dismissive of what EB is and of the years of work that has gone into understanding it; particularly since most literature describes EB as not coincidence, but rather something that arises from evolution or design. Ask Craig Reynold's how many different rules and combinations he had to try before he found 3 that gave him flocking behaviours! It certainly wasn't a coincidence that those three gave him the that behaviour. He chose those rules exactly because they DID provide that behaviour.

Quote:
Original post by InnocuousFox
just merely passing along an observation of how the term is often used in our industry


Ah, so it's your industry? It's not an industry in which academics participate and contribute? Are you claiming that all games technology is developed and implemented in games companies, or consulting firms like yours? Certainly, most games are implemented in said companies... but where do you think they get the technology from? Of course, as is often seen in the industry, someone using the technology doesn't mean that they understand it. Look at ANNs as a perfect example. Further to this, you think that anyone that doesn't put a sig in their post with a company name in it doesn't work in your industry! Now THAT'S presumptive.

Quote:
Original post by InnocuousFox
If I "came across" as "dismissing an entire field of research"
then perhaps the person interpreting it is a bit oversensative to the subject?


Actually Dave, you go out of your way to bash those with credentials, particularly those working in academic or research areas. Anyone on the receiving end of your long-standing disgruntled attitude to these people certainly has a right to feel attacked, because it's not a one-off event with you... it's something you continually raise whenever you get the chance. You're so hung up on attacking people who have credentials that you forget that having them actually means something in the real world. Of course, you also like attacking people that don't know anything (the so-called 'noobs' who visit this site). Perhaps you just like denigrating people for what they do or don't know.

This is not the first thread in which you've made remarks about people using definitions too much... or snide comments like 'fine-tuning their little definitions or theorems'. You also continually accuse people who use definitions of not actually doing anything useful in industry. Your comments about 'our industry' and the 'land of academia' make it perfectly apparent that you think its an 'us (you) versus them (academics)' war and that simply because you make games (a game) that you must be right. There is no us versus them. There are a lot of people working on similar problems from different perspectives, often with different aims, but generally with similar techniques. Definitions should be universal (but generally aren't, mostly do to misunderstanding or misrepresentation).

Since you're obviously ignorant of what each of us here has done and is doing, perhaps you should stop making such ignorant and blatantly rude comments suggesting that we are 'doing nothing'.

Interestingly enough, rather than simply challenge my initial response and ask me to produce references to back up my statements/definitions, you chose to cast aspersions about people who use definitions.

Perhaps we should turn the spotlight on you Dave and ask you what YOU are doing in the AI or Games industry, other than still making that same airport game (which by your own admission doesn't use AI but rather a few tuned response curves and mathematics). Have you published any peer-reviewed articles, or developed any new algorithms or techniques of note? Have you contributed anything useful to the body of knowledge held by the field? Given a presentation at a professional conference? Actually implemented any AI that has had to survive in the real world? Is your entire contribution to this field merely that you log into this site occasionally and offer up spurious information, flippant comments and abuse? It certainly appears that way... and according to your standards, that's enough to judge someone by. Perhaps we should just judge you as 'not doing anything useful' and make snide comments about people from the 'land of Intrinsic Algorithm'.

Or perhaps, for something unique, you might try arguing a point based on its merits, rather than denigrating your opponents.

Timkin

#28 Ferretman   GDNet+   -  Reputation: 276

Like
0Likes
Like

Posted 19 August 2004 - 06:50 PM

Quote:
Original post by Anonymous Poster
Oh for gods sakes, does this board have to become a battle between academia and industry? In game development, emergent behavior *is* a good-looking coincidence. In university, it's a non-linear whatchamacallit. Whoop de doo.

I've been a game programmer for 10 years at 4 different companies and we all consider emergent behavior to be what InnocuousFox says.

Oh no, we must not use an improper definition!


Amen.

IF is precisely right, we do consider EB as more a happy coincidence than anything else. Yes we guide the coincidences, and yes we've got a particular kind of behavior in mind, but the details are entirely up to the algorithm(s) in question. Any meaning to those behaviors is assigned by the observer, not the agents in question.


Ferretman
ferretman@gameai.com
From the High Mountains of Colorado
GameAI.Com

#29 Ferretman   GDNet+   -  Reputation: 276

Like
0Likes
Like

Posted 19 August 2004 - 07:04 PM


Quote:
Original post by InnocuousFox
(Interestingly, other people came in to post in support of my definition - or rather my summation of the definition as held by many in the industry.)


Quote:
Original post by Timkin
You keep mentioning this 'many in the industry', yet you had to pull a few quotes off google - which you clearly misrepresent through your own lack of knowledge of the depth of literature in this field - and claim that these quotes support your comments because someone else agreed with you.


Well I'm one. Eric Dybsand was another, and I'll go out on a limb and add Neil Kirby and (probably) Will Wright as still others. I've been in some of the same conversations that IF has been in with some of the true industry luminaries...EB is a happy thing that comes from assigning motivations and seeing patterns in what is ultimately random or slightly directed behavior.

No one ant knows it's building an anthill...they're just each stacking pebbles on each other.

Quote:
Original post by InnocuousFox
just merely passing along an observation of how the term is often used in our industry


Quote:
Original post by Timkin
Ah, so it's your industry?


You know, he did not say that. He said "our", as in yours and mine and his and ten thousand others.

Quote:
Original post by Timkin
It's not an industry in which academics participate and contribute? Are you claiming that all games technology is developed and implemented in games companies, or consulting firms like yours?


How could you misread something he said so deliberately? Without the academics the industry would be a much sadder place, and IF is very well aware of that. Heck, I think he may listen to the academics more than some of us who are actually building AIs all the time...they have some pretty good ideas!

Quote:
Original post by Timkin
This is not the first thread in which you've made remarks about people using definitions too much... or snide comments like 'fine-tuning their little definitions or theorems'. You also continually accuse people who use definitions of not actually doing anything useful in industry. Your comments about 'our industry' and the 'land of academia' make it perfectly apparent that you think its an 'us (you) versus them (academics)' war and that simply because you make games (a game) that you must be right. There is no us versus them.


Actually my read is that your posts are more along the "us vs them" vein than anything IF has said. He said that most of the people he's talked to in the industry don't consider EB to match the definition you presented..there was no "us vs. them" until you started wondering about the credibility of his sources, his experience, and his arguments.

Lighten up everybody, please...this is the worse kind of emergent behavior that seems to breed in forums....


Ferretman
ferretman@gameai.com
From the High Mountains of Colorado
GameAI.Com

#30 Prozak   Members   -  Reputation: 885

Like
0Likes
Like

Posted 19 August 2004 - 08:04 PM

Ok, lets see...

TeleCommunications is an industry that is allways trying to achieve the maximum potential of data transfer.

There is a whole science between the Phone Industry and how calls are routed...

...and I remember reading about using biological-mimicking algorithms to create "paths" between two nodes, in a way similar to how a group of ants creates the less-energy path in the jungle (notice i said the less energy path, not the shortest path, but the path that requires less energy to travel).

With such a hugely complex system, with so many little rules, I would expect to see emergent behaviour arise from it (or by design, maybe the engeneers added "master" rule systems that destroy any sort of "deviant" behaviour from the system, whatever that may be)...

#31 BenevolentLiao   Members   -  Reputation: 144

Like
0Likes
Like

Posted 19 August 2004 - 09:02 PM

Quote:
Original post by Ferretman
......this is the worse kind of emergent behavior that seems to breed in forums....

As I know, emergent behavior is surely defined as what the Timkin's say at academia. At this early day, industry seems still in the darkness. Maybe industry will win this game, but at this time, I select to stand with Timkin.

#32 kirkd   Members   -  Reputation: 505

Like
0Likes
Like

Posted 20 August 2004 - 02:12 AM

Just my $0.02 and why not add a little fuel to the fire. This is from a 1999 paper in the journal Artificial Life. Reference and abstract:

Ronald, E. M., M. Sipper, et al. (1999). "Design, observation, surprise! A test of emergence." Artif Life 5(3): 225-39.
The field of artificial life (Alife) is replete with documented instances of emergence, though debate still persists as to the meaning of this term. We contend that, in the absence of an acceptable definition, researchers in the field would be well served by adopting an emergence certification mark that would garner approval from the Alife community. Toward this end, we propose an emergence test, namely, criteria by which one can justify conferring the emergence label.

-Kirk

#33 RPGeezus   Members   -  Reputation: 216

Like
0Likes
Like

Posted 20 August 2004 - 08:50 AM

Anyone else care to contribute an example of their emergent behaviours? IIRC, Timkin, I remember reading a post of yours having to due with corners in mazes. Care to re-cap for those who may have missed it?

Will

#34 IADaveMark   Moderators   -  Reputation: 2655

Like
0Likes
Like

Posted 20 August 2004 - 09:48 AM

Quote:
Original post by Timkin
I was referring to your most recent response to an AP, in which you made sarcastic remarks about a lack of credentials. If this wasn't an attack, then I've misinterpreted it. My apologies if that is the case.
It wasn't the case... notice the use of the "sarcasm" tags. Those were specifically put there to show that I was rolling my eyes at the whole "credential" thing.

Quote:
I made no such claims as to the clarity or exactness of my definitions. [snip] Go back and read my original post and please point out where I claimed that my definition was 'crystal clear', if you believe otherwise.
You told me I was "wrong" with the added implication that anyone who says otherwise is incorrect as well.

Quote:
Original post by InnocuousFox
(Interestingly, other people came in to post in support of my definition - or rather my summation of the definition as held by many in the industry.)


Quote:
You keep mentioning this 'many in the industry', yet you had to pull a few quotes off google
I didn't "have to" do anything. That was an idle search just for the sake of saking.
Quote:
... which you clearly misrepresent through your own lack of knowledge of the depth of literature in this field - and claim that these quotes support your comments because someone else agreed with you.
I didn't misrepresent anything. I copied and pasted them complete with links to the original source. It wasn't a bloody thesis, Timkin. Get over it.

Quote:
Quote:
Original post by InnocuousFox
I can't seem to find how I was dismissing anyone or anything -
To quote you form your first post in this thread:
Quote:
"Emergent behavior" is really just a fancy way of saying "a coincidence that really looked cool"


This is particularly dismissive of what EB is and of the years of work that has gone into understanding it; particularly since most literature describes EB as not coincidence, but rather something that arises from evolution or design.
And again, I clarified later that I was speaking soley in terms of the game development world and how the term is used colloquially in that world. (Note that my reporting of this usage has been backed up by others in this thread... and yet you seem to be "dismissing them.)

Quote:
Ask Craig Reynold's how many different rules and combinations he had to try before he found 3 that gave him flocking behaviours! It certainly wasn't a coincidence that those three gave him the that behaviour. He chose those rules exactly because they DID provide that behaviour.
This is so incredibly irrelevant to the conversation that it is ludicrous. I never in this post dismissed any research. Notice that I didn't say that arriving at these algorithms is a random collection of events and coincidences. Sheesh... it is amazing what you manage to read into things!

Quote:
Quote:
Original post by InnocuousFox
just merely passing along an observation of how the term is often used in our industry

Ah, so it's your industry?


our: The plural nominative case of the pronoun of the first person; the word with which a person in speaking or writing denotes a number or company of which he is one, as the subject of an action expressed by a verb.

Therefore, at a minimum, YOU and I become "we". The posessive of "we" is "our". Certainly I was including far more than you and I, but the term still relates. YOU were the one who then claimed it was "your industry" (meaning "my" industry). It was YOU who removed YOURSELF from the collective posessive pronoun that I used. Why is this hard?

Quote:
It's not an industry in which academics participate and contribute?
I claimed nothing of the sort.
Quote:
Are you claiming that all games technology is developed and implemented in games companies, or consulting firms like yours?
I claimed nothing of the sort.
Quote:
Certainly, most games are implemented in said companies... but where do you think they get the technology from?
I never claimed anything to the contrary of this point.
Quote:
Of course, as is often seen in the industry, someone using the technology doesn't mean that they understand it. Look at ANNs as a perfect example.
This is irrelevant to the topic.
Quote:
Further to this, you think that anyone that doesn't put a sig in their post with a company name in it doesn't work in your industry! Now THAT'S presumptive.
I claimed nothing of the sort.

Quote:
Actually Dave, you go out of your way to bash those with credentials, particularly those working in academic or research areas.
I do nothing of the sort. (more later)
Quote:
Anyone on the receiving end of your long-standing disgruntled attitude to these people certainly has a right to feel attacked, because it's not a one-off event with you... it's something you continually raise whenever you get the chance. You're so hung up on attacking people who have credentials that you forget that having them actually means something in the real world.
Putting aside the point that you have implied a helluva lot of motive on my part, you are missing the point of the times that I have "bashed" here. Those are largely the situations where someone comes in asking for a simple, introductory answer to something which he/she/it may be attempting for the first time - and they are answered by some with the most complicated, involved, high-level, doctoral, cutting-edge techniques. They are blown away... not that the material, in and of itself, is incorrect or not useful... but that, to the person in question, it [i]IS[/u] useless. My complaint is always that people need to have a sense of scope on this board - not only of size and complexity for the issue at hand (AS MEASURED BY THE SKILLS OF THE PERSON ASKING THE QUESTION IN THE FIRST PLACE), but of what is appropriate for the genre as a whole. Among some, that level of perception is often significantly lacking.

This is NOT an academic AI board.
This is NOT an AI R&D board.
This is NOT a board where the participants have taken the pre-requisite courses.

Quote:
Of course, you also like attacking people that don't know anything (the so-called 'noobs' who visit this site).
I don't recall doing this. What I do, on occassion, is ask them (and others) to qualify their questions a little better - or to take their focus away from the "tool selection" and pay more attention to "problem definition". Attack? I don't think so.

Oh, and "Mr. Theory & Research"... surely you know that it's not correct to continue to make assumptions regarding something of which you have no direct knowledge. You would lose points for the continual use of phrases like "you also like". Unless you have some method of caputuring me giggling away as I type the alleged offending posts, you don't know a damn thing about what "I like", motives, etc. Please, since I trust that you are making an effort to approach this logically, stop making statements that only serve as an attempt to color perceptions of my participation here.

Quote:
Perhaps you just like denigrating people for what they do or don't know.

And perhaps you like typing 10,000 word responses to people with questions along the lines of "how do I make my tank move?" Shall we analyze YOUR motives in doing that or simply report on the group perception of the altitude of your nasal intakes above the mean surface level of the sea?

Quote:
This is not the first thread in which you've made remarks about people using definitions too much... or snide comments like 'fine-tuning their little definitions or theorems'.
Yep. It's also not the first thread in which it happened that people used them too much. We seem to have found a cause/effect link.

Quote:
You also continually accuse people who use definitions of not actually doing anything useful in industry.
Actually I seem to recall pointing out that endless rehashing of what a definition "should be" doesn't do anything useful in the thread meant to answer a production question rather than a theoretical one. There is a large difference.

Quote:
Your comments about 'our industry' and the 'land of academia' make it perfectly apparent that you think its an 'us (you) versus them (academics)' war and that simply because you make games (a game) that you must be right.
Since I had to point out to you the flawed premise of your interpretation of my use of English pronouns, this is no longer relevant.

Quote:
There is no us versus them. There are a lot of people working on similar problems from different perspectives, often with different aims, but generally with similar techniques. Definitions should be universal (but generally aren't, mostly do to misunderstanding or misrepresentation).
I have never contested this point. You are in error if you believe that I have.

Quote:
Since you're obviously ignorant of what each of us here has done and is doing, perhaps you should stop making such ignorant and blatantly rude comments suggesting that we are 'doing nothing'.
Again, my contention is that there is a paucity of "scope awareness". My "rude comments" simply are observations that the disussions occasionally take on an intellectual elitist air that has little to do with the root purpose of the thread. In your shock and amazement that you have gone way over the heads of the original posters, you can get quite offensive as well, sir.

Quote:
Interestingly enough, rather than simply challenge my initial response and ask me to produce references to back up my statements/definitions, you chose to cast aspersions about people who use definitions.
Unfortunately, I did not record all the roundtables, lectures, panels and discussions that I attended at GDC. That would be convenient. Thankfully, there are some who have come to this very thread to point out that their understanding and observation of the subject matches mine. On the whole, however, I didn't really give a crap about "proving" anything to you. I, for one, wasn't writing a research paper on the subject so I didn't figure I needed to have a freakin' bibliography and footnotes. Again, get over it.

Quote:
Perhaps we should turn the spotlight on you Dave and ask you what YOU are doing in the AI or Games industry, other than still making that same airport game (which by your own admission doesn't use AI but rather a few tuned response curves and mathematics).
If you like... but that is hardly relevant here, is it? Incidentally, I can't remember claiming that my game doesn't use AI. I suppose it depends on you God damn definition of Artificial Intelligence, doesn't it? *sigh*

Writing the AI to generate the best possible combination of thousands of flights to/from hundreds of cities using hundreds of aircraft of a dozen different types, speeds, ranges and capacities all the while taking into account time zones, prevailing winds, airport curfew hours, passenger departure/arrival time travel patterns, the need for connecting flight banks at hubs, taking into account limited gate space, airport arrival/departure traffic limitations AND the Great Bloody Circle Route... is no small task. Of course, some of the people here were awful bloody impressed by the first cut of AI that I had written - when the actually freakin' saw it.
Quote:
Have you published any peer-reviewed articles, or developed any new algorithms or techniques of note? Have you contributed anything useful to the body of knowledge held by the field? Given a presentation at a professional conference? Actually implemented any AI that has had to survive in the real world? Is your entire contribution to this field merely that you log into this site occasionally and offer up spurious information, flippant comments and abuse? It certainly appears that way... and according to your standards, that's enough to judge someone by. Perhaps we should just judge you as 'not doing anything useful' and make snide comments about people from the 'land of Intrinsic Algorithm'.
I'm sorry... which of those was a requirement for calling a spade a spade? Which one of those little check-boxes of yours was a requirement, without which I would be disqualified from proposing how to do the AI for shot selection in pool? Simulating stock market fluctuations and particpant buying/selling behavior? What do I have to get published/approved in order for me to have a damn good algorithm for AI player bidding strategies in Texas Hold'em poker? My stuff works. It works well. 'Nuff said.

Quote:
Or perhaps, for something unique, you might try arguing a point based on its merits, rather than denigrating your opponents.
Sorry that I have to denigrate your last statement... but now you are just being rediculous.

#35 BrianL   Members   -  Reputation: 530

Like
0Likes
Like

Posted 20 August 2004 - 10:55 AM

Apologies in advance, this is a brain dump over a few compiles.

I think part of the confusion here is due to fact that emergence has become a buzz word in the game industry. To public relations people, press, and gamers, emergence means 'cool things that happen unpredictably/randomly' or 'gameplay options'. Games which lead to this definition include Dues Ex and Grand Theft Auto.

Part of the confusion is the viewer; a game player who doesn't know what is happening behind the scene may see emergence everywhere, even though designers have a tight leash on what is happening.

Developers also frequently create artificial interactions to simulate emergence, frequently after seeing a cool event occur. For instance, in Half-Life, when an AI is gibbed, it has a 1 in 20 chance of its head flying in the players direction (instead of a random direction).

In game developement, emergence is seen as the opposite of scripting. While this is certainly true in ways, this certainly expands its definition to include anything systemic.

Several developers I have talked with (game designers and AI programmers) are trying to trim back the open ended definitions. This is resulting in the industry developing its own set of words to describe games which 'encourage emergence'. A recent term I have seen tossed around to describe the potential for emergence in games is 'possibility space'.

Developers are paying attention to self-organization, complex theory, non linear dynamics, etc. Most of the developers I have talked with want to avoid 'contaminating' a useful term by expanding its definition or contradicting established literature.

#36 IADaveMark   Moderators   -  Reputation: 2655

Like
0Likes
Like

Posted 20 August 2004 - 11:15 AM

That actually clarifies quite a bit. Shall we term this phenomenon "definition drift"?
Dave Mark - President and Lead Designer of Intrinsic Algorithm LLC

Professional consultant on game AI, mathematical modeling, simulation modeling
Co-advisor of the GDC AI Summit
Co-founder of the AI Game Programmers Guild
Author of the book, Behavioral Mathematics for Game AI

Blogs I write:
IA News - What's happening at IA | IA on AI - AI news and notes | Post-Play'em - Observations on AI of games I play

"Reducing the world to mathematical equations!"

#37 BenevolentLiao   Members   -  Reputation: 144

Like
0Likes
Like

Posted 20 August 2004 - 01:20 PM

Quote:
Original post by InnocuousFox
Shall we term this phenomenon "definition drift"?

Right. It is surely "definition drift". I agree with you.

Something on dawn, and we are praying for it.

#38 IADaveMark   Moderators   -  Reputation: 2655

Like
0Likes
Like

Posted 20 August 2004 - 06:20 PM

From "AI Game Development" by our pal Alex J. Champandard, Chapter 43, page 584:

Quote:
A Definition of Emergence

Emergence can be tricky to define, because capturing all the intuitive examples of emergence with a single plain-English definition is difficult. Here's a reasonable definition - even if overly generic:

Emergence occurs when complex patterns arise from simple processes combined in a straighforward fashion.

A few examples help illustrate emergence in a more intuitive fashion. In each case, low-level rules cause patterns in the higher-level behavior:

** The movement patterns created by a flock of birds, a school of fish, or a group of cyclists. Locally, the individuals avoid each other while staying close, but the groups coordinate to flow smoothly around obstacles.

** The evolution of intelligence to increase changes of survival is a pattern that arises from the maniupulation of genes.

** Neural networks produce high-level cognition from collections of simple neurons.

** Daily trading drives patterns in the economy.

** The flow of traffic based on the desire of individuals in their automobiles.

Although these examples provide insight into emergence, they also open up gray areas. The next section investigates these issues, and the following subsections present the types of emergence.
(If you want more, buy the book.)

Now... I repost my initial comment here:
Quote:
"Emergent behavior" is really just a fancy way of saying "a coincidence that really looked cool". If you build an agent based model where each agent is thinking for itself using rules to define it's own behavior, sometimes these agents may do things that seem to be working with each other. In fact, there IS no cooperation happening - each agent is only secondarily aware of the other agents (if that), but they have started to do things "near" each other that makes them look like they are cooperating.

With a simple flocking behavior, this can be seen as each "boid" doing it's own thing... (i.e. the rules for moving with the world objects but not coliding with the world objects) and yet the perceived result is that they are moving "together".

In RTS games, there is a high degree of perceived emergent behavior. This happens when simple rules for each agent (unit) is designed in such a way that it tends to compliment other units' simple rules.
Aside from the flipant use of the word "coincidence" - which I already acknowledged and apologized for - what is the problem with what I said? It sounds an awful lot like what ol' A.J. said... and he's a published author ( <-note credential). Add him to the list behind people like Will Wright and Peter Molyneux. (I don't know if they have written books or not... but you may have heard of them anyway.)
Dave Mark - President and Lead Designer of Intrinsic Algorithm LLC

Professional consultant on game AI, mathematical modeling, simulation modeling
Co-advisor of the GDC AI Summit
Co-founder of the AI Game Programmers Guild
Author of the book, Behavioral Mathematics for Game AI

Blogs I write:
IA News - What's happening at IA | IA on AI - AI news and notes | Post-Play'em - Observations on AI of games I play

"Reducing the world to mathematical equations!"

#39 Timkin   Members   -  Reputation: 864

Like
0Likes
Like

Posted 20 August 2004 - 08:26 PM

Quote:
Original post by Ferretman
No one ant knows it's building an anthill...they're just each stacking pebbles on each other.


Why is it that you folk keep suggesting I was saying anything other than this in my original post? Here it is again so there's absolutely NO confusion:

Quote:

...emergent behaviour is (typically) the result of nonlinear feedback mechanisms within sub-units of an entity that aren't specifically encoded to produce these 'emergent' outputs...


This is exactly the definition (albeit with different wording) of the quotes posted by Dave and of the definitions you will find if you look in the literature... even such mundane literature as the dictionary! What I took aim at was the statement that EB was coincidence that looked good. Sure, EB is typically identified by patterns at a system level - which might look good - and these patterns aren't coded explicitly at an entity level, but it is simply NOT coincidence that these patterns are there. To say so is to clearly misunderstand what is happening in such systems. Take the anthill for example. No one ant knows its making an ant hill. However, each ant does know how to climb a pile with a pebble and drop it at the top. The patterns at the system level arise because of the causality of the interactions at the entity level and the design of the entities themselves. If they didn't, then there must be an external hand forcing the system to produce the pattern... and since we're talking about closed systems when discussing EB, then we can reject this. Thus, that behaviour is IN the system because of the entities design or evolution. We call it EB because the pattern is NOT EXPLICITLY coded at the entity level... but you simply cannot suggest that the pattern isn't indirectly coded into the system, no matter what your name is.

(Actually, the ant hill is a bad example for only EB, because it actually shows cooperation and self-organisation as well, when you consider the ant-pebble closed system. The ants cooperate by not all trying to pile their pebbles at the same time and by following each other to the same location... and the ant-pebble system shows more structure and organisation with time) A better example of EB would be a slime (gelatinous organism). The cells in the slime don't explicitly encode the slimes structure or physical properties at the slime level, but its the cells design and interaction with its neighbouring cells that creates the system properties.

Further to this, people should not make the common mistake (even among academics) of mistaking EB for Self-Organisation. They are subtly different beasts, even though they are often found together. Self-organisation is sometimes coincidental, EB is not.


Quote:
Original post by Ferretman
EB is a happy thing that comes from assigning motivations and seeing patterns in what is ultimately random or slightly directed behavior.


'Happy thing'? lol. Now you're saying that EB is a property of the observer!? Are you sure you want to say that? So an ant colony requires someone to watch it for it to be a colony. If we all turn away, they go back to being just a whole lot of ants living close to each other (which is what Dave was saying in his first post about entities acting 'near' each other). Emergent behaviours are system properties, not observer properties. Almost any definition you could find online supports that... let's pull one up...Webster's Online Dictionary

Quote:
Ref: Websters Online Dictionary
An emergent behaviour or emergent property is shown when a number of simple entities (agents) operate in an environment, forming more complex behaviours as a collective. A system made of several things can host properties which the things themselves do not have. For instance, consider two points on a plane. These points will have a distance between them. This distance is not itself a property, but exists in the relation between the points. Emergent properties can arise not only between things in the system, but between other emergent properties. The number and subtlety of these properties can be very much greater than the number of things.


(bolding added by me)

To anyone that suggests that EB is a coincidence of the system noticed by the observer, I challenge you to provide a definition that supports it. So far, no one has done that other than to throw around a few known names in the games industry and suggest that they are right because they had a discussion with these people. It seems quite apparent to me that many arguing for 'coincidence' are simply not seeing what is actually going on in these systems.

Timkin

[Edited by - Timkin on August 21, 2004 8:26:41 AM]

#40 Timkin   Members   -  Reputation: 864

Like
0Likes
Like

Posted 20 August 2004 - 08:26 PM

Quote:
Original post by InnocuousFox
Aside from the flipant use of the word "coincidence" - which I already acknowledged and apologized for


You certainly haven't corrected your definition or apologised (indeed, you certainly don't appear to believe you have anything to apologise for), other than to say that you may have been flippant, terse or over-compensating in your original post (over-compensating for what?).

Quote:
Original post by InnocuousFox
what is the problem with what I said? It sounds an awful lot like what ol' A.J. said...


No. It doesn't. Indeed, what you said was that

Quote:
"Emergent behavior" is really just a fancy way of saying "a coincidence that really looked cool".


This is positing emergence in the observers eyes. Alex certainly doesn't say that anywhere. You're also positing that EB is a coincidental property:

Quote:
do things "near" each other that makes them look like they are cooperating


Again, this is wrong (although I agree there's no explicit cooperation going on here). EB exists because of the system design (the causal rules governing the system); because the behaviour of the individual entities and their interactions has built into it the properties required to produce the observed patterns at the system level.

I'm very well aware of the use of the English language thank you. The use of the word 'our' has another very common usage, one of possesive ownership: 'our thing' for the collective, 'my thing' for the singular. It's used in an exclusionary manner (with reference to the listener/reader) and that's EXACTLY how you came across in your post, whether you intended to or not.

As for me assuming a hell of a lot of motivation on your part, I don't have to assume anything. Your posting history and comments like 'Mr Theory & Research', 'land of academia', 'real world', 'little definitions and theories' clearly ascribe a motivation to you; one of disdain for research and the people that do it. You've made it quite clear, even in your most recent post, that you don't believe there is a place for 'theory' or 'academic AI' in these boards. It's not your place to decide what content does and does not belong here, so get over it and leave people to post what they like. If they make a mistake, feel free to correct them. Feel free to post your own ideas and knowledge... but for heavens sake, stop criticising people. Why do you think they (and me) get pissed off at you?


As for me being 'Mr Theory & Research', you're now making it very clear that you have no idea about me or what I do. As for your comments about this not being an academic AI board, or an AI R&D board, you are right in part. This is a board to discuss all things related to Game AI. Funnily enough, academic AI and AI R&D do have something to say about how to make good game AI. To ignore these very important areas leads to very boring, repetitive and basically crap game AI that has been around far too long in this industry (and continues to get into current games). Attempting to convey complex topics to readers in lay terms is not something I do to try and 'elevate my nasal passages' as you put it, but rather to try and share some knowledge with people so that they too might see the depth of possibilities in their game AI. Sure, some of it is going to go over their heads, but that doesn't mean it shouldn't be posted or discussed. Seeing that there is more to a topic is often enough to spur people to go away and read up on it and learn more about it. To treat everyone like a moron and to give them only the most basic, naive approaches to game AI is not going to advance this industry one iota.

Timkin




Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.



PARTNERS