Convolutional Neural Network for the 2D Ising model

This project uses a convolutional neural network (CNN) to determine the most likely temperature for a given Ising model spin instance. From the weights of these CNNs we can approximate the critical temperature for the phase transition in the Ising model.

Written in Python, data wrangling done using the SciPy ecosystem. CNN implemented using Keras. The Ising spin models are generated using a program written in C# (see: ADD LINK TO GITHUB). See the report in the root for more details.

Screenshots:

Visualisation of the transition in the weight matrix of the CNN