On Wed, May 5, 2010 at 6:10 PM, Neal Alexander
<relapse.dev@gmx.com> wrote:
Yea, I'm interested. Over the last several months I've been reading a few books on AI and have been trying to distill a Haskell library out of them:
The library is pretty primitive so far, but this is what i have laid out:
- Neural Networks (usable)
- Blackboard Architecture (work in progress)
- FSM (usable)
- Genetic Algorithms (usable)
- Goal Oriented Behaviors (work in progress)
- Goal Oriented Planning (work in progress)
- Markov Chains (work in progress)
- Steering (usable)
- Fuzzy Logic (usable)
- Decision Tree (work in progress)
At the moment I'm working on some constrained delaunay triangulation algorithms to use for spatial reasoning / path planning.
That looks awesome!
Would you have some time to drop by the #haskell-math IRC channel ?
I'm planning to ask for a mailing list and a wiki for the project because a single mailing list thread won't be of help here. But for the moment most discussions happen on #haskell-math.
Anyway, would you be willing to integrate your library in that project ?
Side note : we're currently discussing the minimal algebra framework we will have to ship with the library, letting us implement general algorithms instead of specific ones (why being restricted to integers when any group / ring / field can be used ? etc).