README.md

Evolving Neural Networks with Particle Swarm Optimization

How to optimize a neural network without backpropagation

View Demo · Report Bug · Request Feature

About The Project

This project was designed to understand the optimization in the neural networks and how we can use other optimization algorithm rather than backpropagation.

We publish an article on the medium of Semantix about this topic.

Semantix Medium Article

(back to top)

Built With

This project was builded with the next technologies.

• Python
• Jupyter
• Pyswarms
• Tensorflow
• Keras

(back to top)

Getting Started

Prerequisites

You need the next componenets to run this project.

• Docker. To install it follow these steps Click. On Ubuntu, you can run:

sudo apt-get install docker-ce docker-ce-cli containerd.io

• Visual Studio Code. To install it follow these steps Click. On Ubuntu, you can run:

sudo snap install code --classic

• Install the visual studio code extension "Remote - Containers"

Installation

Follow the next steps:

1. Run the visual studio code.
2. Open the folder where you clone the repository.
3. Click on the green button with this symbol in the bottom left of visual studio code "><".
4. Click on reopen in a container.
5. Execute "experiments.ipynb" notebook.

(back to top)

License

Distributed under the GNU v3 License. See LICENSE for more information.

(back to top)

Contact

Esteban Vilca - @ds_estebanvz - esteban.wilfredo.g@gmail.com

Project Link: https://github.com/estebanvz/backpropagation_pso

(back to top)