ProjectIII

Chapter 2: Neural Networks

2.5: example_building.py

Chapter 3: Solving Differential Equations

3.4.2: example_dNdX.py

3.5.5: example_dJdT.py

Chapter 4: Deep Learning

4.1.1: learning_rate2.py

4.1.5: TDataGen

4.2.4: nesterov_momentum.py

4.4.1: deep_learning.py

4.4.2: DLDataGen

Chapter 5: Activation Functions and Random Initialisation

5.3: AFDataGen

Chapter 6: Solving Laplace Equations

6.1: Dirichlet_Training, fixed.gif, pre.gif, free.gif

6.2: Dirichlet2_Training

6.3: Neumann_Training

Extras

These are code and animations that are not covered in the report, but are still related the project:

ode_numpy.py This is a numpy only implementation of the neural network (chapters 3 & 4).

ode_jax.py This is the JAX implementatio of the ODE solving neural network, this code is the basis for the comparisons (chapters 4 & 5).

runge_kutta_comparison.py This is a comparison of the neural network ODE solver versus runge-kutta integration methods (poster).

*.gif Animations for various learning rates (chapter 4) & Laplace equations (chapter 6).