Fully-Connected Feedforward Neural Nets

Question 1

Describe what a Fully-Connected Feedforward Neural Net is

Answer 1

Each neuron in the preceding layer connected to every neuron in the subsequent layer without any backward connections or cycles.

Question 2

How do the neurons operate in a Fully-Connected Feedforward Neural Net

Answer 2

The neurons use activation functions to change their output based on the input

Question 3

What activation functions are available in a Fully-Connected Feedforward Neural Net?

Answer 3

linear, sigmoid, hyperbolic tangent, or rectified linear unit (ReLU)

Question 4

What do non-linear activation functions allow for in Fully-Connected Feedforward Neural Nets?

Answer 4

allows for deep neural networks to model complex functions

Question 5

What are some of the restraints of training a Fully-Connected Feedforward Neural Net?

Answer 5

The more neurons that are added means it will require larger computational resources