Python for Deep Learning - Build Neural Networks in Python - Leaky Rectified Linear Unit function

The video tutorial introduces the leaky version of the rectified linear unit (ReLU) function, highlighting its modification to address the negative impact of the standard ReLU. It explains the typical value for the parameter 'a' in the function's diagram and discusses the range of the leaky ReLU, which extends from negative to positive infinity. The tutorial also covers the differentiable and monotonic nature of the leaky ReLU function.

What is the primary purpose of the leaky version of the rectified linear unit function?

• To increase the positive output range

• To solve the negative impact of the ReLU function

• To make the function non-differentiable

• To limit the output to a fixed range

What is the typical value for 'a' in the leaky ReLU diagram?

• 0.01

• 0.1

• 1.0

• 10

What is the range of the leaky ReLU function?

• 0 to 1

• Negative infinity to positive infinity

• Negative infinity to 0

• 0 to positive infinity

Which of the following is true about the leaky ReLU function?

• It has a limited range

• It is a monotonic function

• It is not differentiable

• It is non-linear

What property does the leaky ReLU function share with the standard ReLU function?

• It has a fixed output range

• It is monotonic

• It is differentiable

• It is non-monotonic