Neural Network

Question 1

What is a neuron?
What is an activation function?
What are the most common?
NN formalism
What is the hypothesis class of a NN?

Answer 1

…

Question 2

What type of functions can be implemented using a neural network?
Sample complexity.
What is the runtime of learning a nn?
What is the solution?

Answer 2

…

Question 3

Matrix notation. Forward propagation algorithm.

Answer 3

…

Question 4

General construction of NN for a given Boolean formula

Answer 4

…

Question 5

How do we learn the nn parameters?
What is a sensitivity vector?
How to compute sensitivities for a given datapoint (x,y)?

Answer 5

…

Question 6

Backpropagation algorithm

Answer 6

…

Question 7

How we can regularize NN?

Answer 7

…

Question 8

What are the issues of traditional NN?
What are CNN?
How the Convolutional product works? What are its properties?

Answer 8

…

Question 9

What is padding? Relu? Pooling? SUbsampling and its properties

Answer 9

….

Question 10

What is ADAM? How it works?

Answer 10

…

Question 11

What are the techniques to avoid overfitting? And how they works?

Answer 11

…

Question 12

What is the Loss Function for CNN?

Answer 12

…