Basic Artificial Neural Networks in Python

4.1
stars
133 ratings
Offered By
Coursera Project Network
6,820 already enrolled
In this Guided Project, you will:

Generate a sample dataset using Scikit-Learn.

Implement an activation function and feed-forward propagation in a multi-layer ANN in Python code

Utilize gradient descent to adjust the weights of each layer of our ANN through back-propagation implementation in Python code

Clock2 hours
IntermediateIntermediate
CloudNo download needed
VideoSplit-screen video
Comment DotsEnglish
LaptopDesktop only

In this 1-hour long project-based course, you will learn basic principles of how Artificial Neural Networks (ANNs) work, and how this can be implemented in Python. Together, we will explore basic Python implementations of feed-forward propagation, back propagation using gradient descent, sigmoidal activation functions, and epoch training, all in the context of building a basic ANN from scratch. All of this will be done on Ubuntu Linux, but can be accomplished using any Python I.D.E. on any operating system. We will be using the IDLE development environment to write a single script to code our simple ANN. We will avoid using advanced frameworks such as Tensorflow or Pytorch, for educational purposes. Note that the resulting ANN we build will be use-case agnostic and be provided with dummy inputs. Hence, while the ANN we build and train today may appear to be a useless demonstration, it can easily be adapted to any type of use case if given proper, meaningful inputs. I would encourage learners to experiment- How easy is it to add more layers without using frameworks like Tensorflow? What if we add more nodes? What limitations do we come across? The learner is highly encouraged to experiment beyond the scope of the course. Note: This course works best for learners who are based in the North America region. We’re currently working on providing the same experience in other regions.

Skills you will develop

Deep LearningArtificial Neural NetworkPython ProgrammingPropagationTensorflow

Learn step-by-step

In a video that plays in a split-screen with your work area, your instructor will walk you through these steps:

  1. Generate a dataset using Scikit-Learn

  2. Plot generated sample dataset to a graph using pyplot

  3. For each layer, multiply inputs by randomly generated weights

  4. For each layer, calculate the dot products of our two-dimensional sample features

  5. Write a sigmoidal activation function in Python and pass the dot product of our features through it before passing as input to the next layer to accomplish feed-forward propagation

  6. Write a cost function in Python based on the Mean Squared Error method

  7. Utilize gradient descent to adjust the weights of each layer of our ANN through back-propagation implementation in Python code

How Guided Projects work

Your workspace is a cloud desktop right in your browser, no download required

In a split-screen video, your instructor guides you step-by-step

Reviews

TOP REVIEWS FROM BASIC ARTIFICIAL NEURAL NETWORKS IN PYTHON

View all reviews

Frequently asked questions

Frequently Asked Questions

More questions? Visit the Learner Help Center.