Deep Learning - Deep Neural Network for Beginners Using Python - Other Activation Functions

The video tutorial discusses various activation functions used in neural networks, including sigmoid, tanh, and ReLU. It explains the properties and formulas of these functions, highlighting their differences and applications. Sigmoid and tanh are compared, noting their output ranges and impact on the vanishing gradient problem. ReLU is introduced as a simple yet widely used function, with examples illustrating its behavior. The tutorial concludes with a summary of these activation functions and mentions other available options in the data science community.

What is the main difference between the output ranges of the sigmoid and tanh activation functions?

• Sigmoid ranges from -1 to 1, tanh from 0 to 1

• Sigmoid ranges from 0 to 1, tanh from -1 to 1

• Both range from 0 to 1

• Both range from -1 to 1

How does the ReLU activation function handle negative input values?

• It returns zero

• It returns the negative value

• It returns one

• It returns the absolute value

Which activation function is known for its simplicity and is widely used in the data science community?

• Softmax

• ReLU

• Sigmoid

• Tanh

What is the output range of the ReLU activation function?

• 0 to 1

• -1 to 1

• 0 to positive infinity

• -infinity to infinity

Which of the following is NOT a characteristic of the tanh activation function?

• Output range from -1 to 1

• Helps mitigate vanishing gradient problem

• Output range from 0 to 1

• Similar curve to sigmoid