This is the result of my internship at the LCPMR from May to Jully 2022.
The goal is to predict the energy of TWo ELectronic INtegrals using Marchine Learning more particularly Neural Network.
Due to the nature of two electronic integrals, we need to develop methods to face two majors issue :
- the size of the training set
- the over representation of near zeros values in the dataset
This project contains :
-
DATA :
This library is a set of tools that deals with the dataset and the training.
-
docs :
The reports and the website documenting this project.
-
examples :
Some commented and working examples.
-
test :
A series of test to assert the basic functions works.