
Episode 07: Yujia Huang, Caltech, on neuro-inspired generative models
Generally Intelligent
00:00
The Effects of Priming on Learning
There's probably, I think, lots of different mechanisms. So one is a top down feedback, which is kind of a paper, that sort of an idea. There could be other mechanisms where you pay attention. But even as you process, if you just are staring at the same thing for a little while, or even not very long hundreds of milliseconds, then you'll still continue getting extra information. Maybe what's going on there is I'm treating some kind of prediction internally, based on my previous state and having been told what this is about. In general, it's hard to see some biological optimization method to eat backprop propagation because we don't have loss function in our brain for
Transcript
Play full episode