Deep learning

Question 1

What are the step of progressive encoding?

Answer 1

1) Input
2) Segmentation
3) Completion
4)Spatiality
5)Modeling
6)Semantics

Question 2

Basic idea of CNN

Answer 2

local receptive field
shared weights
rectified linear units and pooling

Question 3

2D Convolution formula

Answer 3

O[m,n] = f[m,n]*g[m,n]=
= sum_u(-inf, inf)(sum_v(-inf,inf)(f[u,v]g[m-u,n-v])

Question 4

What are the main task achieved by convolution

Answer 4

Filtering and downsampling

Question 5

Formula of a neuron in a convolutional layer

Answer 5

u_i_m = f(b + sum_l(1,L)_m(1,M)(w[l,m]a[j-1+l,k-1+m]))
f can be tanh for example

Question 6

formula of output size of convolutional layer

Answer 6

sx =(Ix+2zx-fx)/Lx +1

Question 7

Two ways to process RGB images in convolutional layers

Answer 7

1) Using three parallel lines of processing
2)Using a 3D filter to include include all three channel in the computation of each pixel of the feature map

Question 8

Complete series of layer for the processing of an image and what they do

Answer 8

Input
Conv layer: filtering
ReLU layer: makes everything positive
POOL layer: smoothing

Question 9

Main kinds of pooling

Answer 9

Max and average

Question 10

Softmax classifier

Answer 10

It uses normalized exponential activation with many neurons as the number of classes
The cost function is the negative log probability of the correct label (cross-entropy regime)

Question 11

What is the main difference between traditional ANN and deep learning

Answer 11

In traditional ANN we apply learning techniques to well-know quantities. Usually we start from raw data, we extract relevant features by preprocessing and then we train a model
In deep learning we do not know exaxtly what are the features that determine the task. End to end the model execute the task directly on the raw data