3 ms·
"A Deep Learning system does not have a temporal component" but isnt that what recursive deep learning is for. And how do they calibrate low/middle/high.
by mooneater 10y ago
"A Deep Learning system does not have a temporal component" but isnt that what recursive deep learning is for.
And how do they calibrate low/middle/high.
- kleebeesh 10y agoRoughly speaking, low/middle/high are just buckets, each of which represents a third of the distribution for position, velocity, or angle. We attempt to illustrate that in this image: http://neuromorphic.eecs.utk.edu/images/graphics/Inv-Examples.jpg http://neuromorphic.eecs.utk.edu/images/graphics/Inv-Example...
- mooneater 10y agoOk so this looks like Fuzzy Logic? Or does the system automatically determine the low/middle/high boundaries?
- throwaway287391 10y ago> "A Deep Learning system does not have a temporal component" but isnt that what recursive deep learning is for. Yes, recurrent nets exist, and have since the 80s, before even convolutional nets. The authors seem a bit out of the loop, considering the (more recent) success of RNNs in speech recognition... Arguably this article itself describes a type of RNN.