Please enable JavaScript.
Coggle requires JavaScript to display documents.
About neural network ? I know nothing (MNIST <<Neural Networks and…
About neural network ? I know nothing
MNIST
<<Neural Networks and Deep Learning>>
deep learning
convolutional networks
Local receptive fields
each neuron in the first hidden layer will be connected to a small region of the input neurons
shared weights and bias for each feature and you might need more than one features which will require more parameters for
feature map
pooling layer used immediately after converlutional layer
max pooling
L2 pooling
single layer
the cross-entropy cost function
universality
gradient descent by backpropagation
vanishing gradient problem
exploding gradient problem
backpropagation
auto-encoder
https://www.youtube.com/playlist?list=PL6Xpj9I5qXYEcOhn7TqghAJ6NAPrNmUBH
denoising auto-encode is that some kind of dropout ????
https://www.youtube.com/watch?v=t2NQ_c5BFOc&list=PL6Xpj9I5qXYEcOhn7TqghAJ6NAPrNmUBH&index=48#t=142.81507
https://www.youtube.com/watch?v=vfnxKO2rMq4
text