Machine Learning for Crystal Structure Prediction (available)

Starting Date: Summer 2020
Duration: 12 weeks
Time commitment: Full time
Prerequisites: None formal, but you will need good programming skills (Python, C++). Some Machine Learning notions might be useful. *NO* prior knowledge of Chemistry is required.

Crystal Structure Prediction (CSP) is one of the major problems in computational chemistry with numerous applications in real life. This is essentially the (global) minimisation of a continuous, high-dimensional, complicated function. Many heuristic methods have been proposed for CSP
and recently new methods based on Machine Learning were introduced.

The goal of this project is to study a simpler-to-understand function and evaluate it for the CSP problem using Machine Learning techniques. The project can be decomposed into three technical parts.

  •  Local minimisation of the simpler function.
  •  Build a classifier for deciding whether to relax a structure or not.
  •  Search for new materials.

I will explain to you the basics of CSP, some heuristic methods for it. In addition, we will see what local minimisation is and we will learn some
methods for solving this problem.
You will be given access to some pre-existing code.
NO prior knowledge of Chemistry is required.

Depending on the results, we can submit our findings to a top ML
Conference (NeurIPS, ICML),  or to a top experimental venue (ALENEX,
ESA track B, SEA).

If you have any questions, feel free to ask me!