# Derivatives Of Activation Functions (C1W3L08) | Summary and Q&A

60.6K views
โข
August 25, 2017
by
DeepLearningAI
Derivatives Of Activation Functions (C1W3L08)

## TL;DR

This content discusses different activation functions used in neural networks and how to compute their derivatives.

## Key Insights

• ๐๏ธ Activation functions play a crucial role in neural networks as they introduce non-linearity and enable modeling complex relationships.
• ๐งก The sigmoid activation function has a range between 0 and 1, and its derivative can be simplified to G(Z) * (1 - G(Z)).
• ๐งก The hyperbolic tangent activation function has a range between -1 and 1, and its derivative is calculated as 1 - (G(Z) * G(Z)).
• ๐ค The Leaky ReLU activation function is defined as 0 for Z < 0 and Z for Z >= 0, with a derivative of 0 for Z < 0 and 1 for Z > 0.

## Transcript

when you implement back-propagation for your neural network you need to really compute the slope or the derivative of the activation functions so let's take a look at our choices of activation functions and how you can compute the slope of these functions can see familiar sigmoid activation function and so for any given value of Z maybe this value ... Read More

### Q: What is the purpose of activation functions in neural networks?

Activation functions introduce non-linearity to the neural network, allowing it to learn complex patterns and make predictions. They determine the output of a neuron or node.

### Q: How do you compute the derivative of the sigmoid activation function?

The derivative of the sigmoid function is calculated as G(Z) * (1 - G(Z)), where G(Z) is the output of the sigmoid function.

### Q: What is the derivative of the hyperbolic tangent activation function?

The derivative of the hyperbolic tangent function is computed as 1 - (G(Z) * G(Z)), where G(Z) is the output of the hyperbolic tangent function.

### Q: How is the derivative of the Leaky ReLU activation function defined?

The derivative of the Leaky ReLU function is 0 for Z < 0 and 1 for Z > 0. When Z = 0, the derivative is undefined, but it is commonly set to either 0 or 1 in practice.

## Summary & Key Takeaways

• The content provides an overview of activation functions and their derivatives, focusing on the sigmoid, hyperbolic tangent, Leaky ReLU, and ReLU functions.

• It explains how to compute the derivatives of each activation function and provides examples to demonstrate their behavior.

• The content concludes by mentioning the importance of computing the derivatives for implementing gradient descent in neural networks.