Alter-ego - A Reactive AI Library

http://nakkaya.com/2010/06/29/alter-ego-a-reactive-ai-library/

by joubert • 16 years ago

13 7 16 years ago

Is it fair to call this an AI library? It looks like a nice, graphical way to put together control flow statements. But this description doesn't mention features you might find in other cognitive architectures, like keeping a working memory or any kind of learning.

_delirium2• 16 years ago

There's kind of a trend towards reactive-AI architectures in some areas, especially in game-AI and robotics, with a focus on concise human authoring of situationally reactive plans, as opposed to autonomous learning/etc. The intelligence comes in the ability for a human to specify a fairly concise, high-level plan, which will then "do the right thing" in various situations, without having to explicitly specify all the variations of what that might be. That admittedly gives it a strong PLs focus rather than being a pure AI problem.

Three papers on different approaches, getting sort of progressively further from classical planning as time goes on:

http://www.ai.sri.com/pubs/files/1364.pdf

http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.106....

http://www.gamasutra.com/gdc2005/features/20050311/isla_01.s...

The last one is the direct inspiration for this library I assume, since it's the source of the term "behavior trees". I agree that many do have somewhat more complexity; for example, the 2nd link above is a reactive system with working memory, resolution of joint and dependent behaviors, etc. But game developers have tended to want to strip things down as bare-bones as possible, due to wanting: 1) minimal CPU usage; 2) a simple enough model that not-very-technical gameplay designers can specify a lot of the behaviors; and 3) something simple and deterministic enough that the engineers can predict and debug its behavior.

pvg• 16 years ago

I'm not sure you answered the poster's question, though. The criteria you describe are equally applicable to a bunch of if statements.

nakkaya• 16 years ago

In addition to the links _delirium provided,

http://chrishecker.com/My_Liner_Notes_for_Spore/Spore_Behavi...

http://aigamedev.com/insider/presentations/behavior-trees/#r...

http://books.google.ru/books?id=1OJ8EhvuPXAC&lpg=PA370&#...

and for people who prefer XML there is a simular architecture based on FSMs called XABSL,

http://www.xabsl.de/

which is used heavily among Robocup teams.

mrfabbri• 16 years ago

It's a "weak AI" approach http://en.wikipedia.org/wiki/Weak_AI (i.e. demonstrate intelligence) but it's definitely fair to call this an AI library. Subsumption architecture http://en.wikipedia.org/wiki/Subsumption_architecture is the cognitive architecture underlying reactive AI.

growingconcern• 16 years ago

Regarding the question about whether it's fair to call this an AI library...well it's kind of a valid question. But really this is game AI which is isn't trying to be Artificial Intelligence, it's just the subject of making game agents perform behaviours and look intelligent - or even just functional. But is your question really even an important question? Much of classical AI can just be reduced to search (which is why prolog is so popular in university AI courses). Is search AI? Is it any more AI than putting together control flow statements? There is a lot of stuff that can happen in those condition and behaviour nodes. And yes in trivial examples the control flow is fixed, but the "learning" is done by hand by the coder/designer (is it any more "AI" to have it done by a evolutionary algorithms?). And more advanced versions of behaviour trees can incorporate much more sophisticated selectors than just "pick the first that works".

Alter-ego - A Reactive AI Library

Related Stories