Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Locked thread
Mr. Nice!
Oct 13, 2005

c-spam cannot afford



after watching avengers 2 it seems that we just need a space gem to make ai work right.

Adbot
ADBOT LOVES YOU

Bloody
Mar 3, 2013

Romulan Hippie posted:

put it through an lstm net, input sample at time t, predict sample at t+1. use hidden activation at final timestep as features. keras is pretty sweet -- its like they took the learning curve away from theano

the problem is I don't care what individual samples are but instead the shape and magnitude of like 1 to 5 seconds of samples. basically I have several different possible events that last in the neighborhood of a few seconds and I want to differentiate between them. I need to look at more data though it might be super easy. or completely impossible.

Bloody
Mar 3, 2013

although lstms sound dope

  • Locked thread