This repository contains the figures and the code used for the development of the project "Machine Learning Implementation on the Prediction of the Thermodynamic Arrow of Time Direction in Brownian Motion". This project was based on the work made by Seif et al. (2021) available in https://doi.org/10.1038/s41567-020-1018-2.

Contents

1. Figuras Proyecto: Folder containing the final figures of the project.
2. LogisticRegressionModel.ipynb: Jupyter Notebook containing the data generation process along with the model setup, training and testing.
3. modeloProyecto12000Tr.pth: Trained model parameters.
4. TrajectoryGenerator.py: Python file containing the code used to generate all the trajectories used for training/testing the classification model.

Final document (in spanish, with abstract in english) available in https://fisstat.uniandes.edu.co/wp-content/uploads/Documento-Final-Proyecto-Teorico-Juan-Sandoval.pdf)