rorg314 / ml-physicists-examples

This repository contains worked examples that I found useful when following the Max Planck Institute course 'Machine Learning for Physicists' https://machine-learning-for-physicists.org/

ML-Physicists-Examples

This repository contains worked examples that I found useful when following the Max Planck Institute course 'Machine Learning for Physicists' https://machine-learning-for-physicists.org/

Examples

The ExampleNotebooks folder contains several files (.py converted from .ipynb files) that illustrate several examples of convolutional neural networks.

• The CNNTraining file contains a simple example for training a convolutional network to recognise shapes (circles and squares)

• The DenoiseCNN file contains a denoising auto-encoder, using the same shape recognition network from CNNTraining

• The DigitRecognition contains an example network that recognises digits using the MNIST handwritten digits database.

Written code

The WrittenExamples folder contains python files that I wrote to practice implementing simple neural networks from scratch (no libraries).

• The IntroNetwork file contains a manual implementation of a simple sequential network with three hidden layers.

• The BackpropNetwork file implements backpropagation to the IntroNetwork.