Final Project for EC523: Deep Learning
The file 'dataset_generator.ipynb' contains code to generate trajectories of the three-body problem that can be used to train the neural network models. This code was authored by Billy Harrington.
The file 'baseCase.ipynb' contains code for the base case, vanilla neural ODE. This code was primarily authored by Mark Tracy, with contributions from Parth Ghayal. Parth also worked on a separate version of the base case model, which was ultimately abandoned in favor of the present version.
The file 'LTC.ipynb' contains code for the liquid time constant (LTC) network case. This code was co-authored by Mark Tracy and Parth Ghayal.
The files 'plotter_LTC.ipynb,' 'plotter_baseCase_Ghalf.ipynb,' 'basecase_sample_efficiency.py,' and 'ltc_sample_efficiency.py' contain code for post-processing of the model runs. This code was primarily authored by Evan Donovan.
The folder 'Instances' contains instances of the files 'baseCase.ipynb' and 'LTC.ipynb' that were used to train the six runs that are included in our report.