23 August 2012

= Meeting Notes =


 * New members wanted! Lee already emailed his advisees, but this does not include all people from past GP classes, nor people that changed advisees while Lee was on sabbatical. He is going to do that.
 * NO meeting next week.
 * Following week, meeting on Friday 1-3pm at Hampshire (this will be the regular time henceforth.
 * Zeke has been evolving R programs. Programs seem to be memorizing, not coming up with meaningful data. He would like to see how some "solutions" do on test data to see if they have any predictive power.
 * We all want GP web apps, through Clojure and/or Clojurescript. Zeke and Lee have been working on this, random errors and bugs. Still being worked out--oh wait, we fixed it during meeting, now it works.
 * Omri should really get the Flea Market all set and working (what a thought)!
 * Lee looked at Emma's TSM code, hasn't yet tried out the examples.
 * Emma included input instructions for TSM. Discussed possibility of non-uniform TSM generation, with high amount of pairs and tag instructions.
 * Lee proposed a COSMOS experiment with non-parametric data: COSMOS versus no COSMOS. The idea would be to show that no COMSOS will show "significant" results after some "large" number of runs, BUT after a "larger" number of runs, it will no longer show significance. Ideally COSMOS would not be fooled by this ephemeral "significant result" (stardust?).
 * Tom has been doing Bowling runs. Nothing groundbreaking, progress happening.
 * Lee brought up a very interesting idea about evolving rules for cellular automata given image data.
 * For next meetings' status report, everybody should also include a reading suggestion.
 * We all decided to be BFFs (friendship bracelets pending)!

Next Meeting

CI Home