this is for holding javascript data
Deepak Menghani edited sectionFormatting_yo.tex
about 8 years ago
Commit id: 53341dece09b2e072ee8ebcfdc89107c8e048d08
deletions | additions
diff --git a/sectionFormatting_yo.tex b/sectionFormatting_yo.tex
index c4bd34f..9344fcf 100644
--- a/sectionFormatting_yo.tex
+++ b/sectionFormatting_yo.tex
...
We have been able to construct simple gradients of a neuron with respect to the input image. Gradients of neurons at different layers of VGG net are shown below in Fig[].
[Insert figure1]
[Insert figure2]
[Insert figure3]
[Insert figure4] \label{fig:g1} Gradients of activation of Neuron at Layer 1 of VGG Net
\label{fig:g5} Gradients of activation of Neuron at Layer 5 of VGG Net
\label{fig:g10} Gradients of activation of Neuron at Layer 10 of VGG Net
\label{fig:g16} Gradients of activation of Neuron at Layer 16 of VGG Net
From these results it seems, that the neurons in the early layers, do not contain information rich enough for us to sample and work backwards from. Neurons further down the layers represent more abstract and macro feature from the image eg. a neuron which fires up when the input image contains an ear. We think it would be ideal for us to re-construct the image using sampled activations from some layer in the middle as we hypothesize that the information contained is maximal somewhere in between the first and last layer. We aim to do experiments to figure out which layer would be the best to work with using our reconstruction algorithm.