PeterOMalley's blog

Alternate Neural Network Training Method: Real Life!

OK, so my title was a bit provocative, but here's what I'm going to do for my project (and hopefully it will work). (When we went around on Wednesday and explained our projects, I said basically this, but now I'd like to elaborate it more.) Training Neural Networks to do ANDs and ORs is all fine and good, but I feel that it misses the point, at least in terms of emergence. Neural Networks show great potential in terms of solving computing and AI problems, but I'd like to go somewhere different. I want to write a simulated world where the creatures are run by neural networks. The inputs to the neural networks will be the "sense": vision, for example, could be represented by two parameters: one for the distance of the nearest object in the line of sight, and another one for the "color", where food would have one color, other creatures another color, and obstacles a third. (The distance and color would have to be normalized so as to be a number between zero and one, of course.) The outputs, then, could be actions: one output could be whether to move forward or not, another to be whether to turn left, right, or not at all, and maybe another could be to change the creature's own color.

Slight Modification of Paul's Multi-Leveled Ant

I modified Paul's last NetLogo program so that the ant doesn't base his moves purely on how successful a bold or safe move has been in the past, but rather on how successful each type of move has been from his current color patch. That is, there is a success ratio for a bold move from a yellow patch, a bold move from a black patch, a safe move from a yellow patch, and a safe move from a black patch. In the clumpier setups, he learns pretty quickly to move safely when on yellow and boldly when on black. Java Applet <


I've heard the term "emergence" applied to astronomy, and since I seem to be the only person studying astronomy in the class (though if I'm not, please say so!), I feel like I should say something about it. I admittedly don't know to much about it, but my understanding is that the concept of emergence is applied to the origin of large-scale structure in the universe; that is, galaxies, clusters of galaxies, and superclusters, the biggest structures after the universe itself. A Haverford alum, Ravi Sheth, came to give a talk about this last semester, but I don't believe he used the term "emergence." He works with computer models of galaxies and such, and linked it to, among other things, the stock market and traffic.