I haven't been blogging here much because I've been spending time on a research project. It requires a lot of programming and machine-time to actually run the programs. I have it down to a science now, so at this point it's mostly tweaking parameters to see what happens. Basically, I created an artificial life workbench to test out some ideas about evolution and survival in a computational context. You can read background
here if you're interested. The "artificial life" consists of little computer programs that have to solve certain tasks to survive. They reproduce and mutate if successful enough. The special programming language is very concise, and the critters look like >-]- and +#}O{]<>}. I chose the symbols for their (mostly) bilateral symmetry, so if you squint they actually do look like weird life forms. The first one, when run in the environment does this:
- > Look at the old environment (looks like feelers, no?)
- - Subtract one from it
- ] Skip to the end unless the result is zero
- - Subtract one more if it was a zero
- Output the result
The environment is only ever 0,1, or 2, so the critter maps 0 or 1 to 9 (zero minus one is nine) and 2 to 1. This is exactly a recipe for survival. It took 109 generations to evolve, and only happened once in a 100 trials.
In other news I came across a nice list of education blogs and news sites. I haven't gone through them all yet, but many I haven't seen before. They are not all higher education. The list is
here.
I also discovered
Crossing the Finish Line, described in one review on the Amazon page as "The most comprehensive look yet possible at the determinants of graduation rates--and what might be done to improve them." It looks fascinating, but I've only been able to take a peek so far.
Also of note: NPR's piece "
Who Needs College, and Who Shouldn't Go?"
No comments:
Post a Comment