I broadly also agree with Ben: it's good work, but not particularly groundbreaking. The critical insight, to the extent that there is one, seems to be the idea of creating a general-purpose recognition unit, and stacking up multiple instances of those to generate high-level results.
My own work takes very approximately the same approach, although it differs considerably in the details. As for Numenta, the guys I've spoken to there have been very smart, but their tech demos have so far been underwhelming. -mattb ----- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=4007604&user_secret=8eb45b07
