Cluster expansions with icet
Mattias Ångqvist and Magnus Rahm
Learning objectives
In this tutorial you will learn
- what cluster expansions (CEs) are and what they can be used for
- how to generate training sets for CE construction using icet,
- to apply machine learning algorithms for constructing CEs,
- to sample a CE in various thermodynamic ensembles using Monte Carlo (MC) simulations, and
- to efficiently analyze data from MC simulations.
Overview
This tutorial provides an introduction to icet and cluster expansions. The tutorial comprises four parts, each provided in the form of a jupyter notebook along with supporting data.
- Stage 1: structure generation, enumeration, and mapping.
- Stage 2: cluster expansion construction for the AgPd alloy
- Stage 3: sampling of in the canonical and SGC ensemble for an Ising model
- Stage 4: analyzing MC data using data containers from simulations of the Ising model
During this tutorial developers of icet can also discuss questions concerning own projects with the icet developers.
Prerequisites
You need a laptop with a working Python3 environment with ASE 3.18.0 or newer and icet 1.0 or newer. To install icet run
pip3 install --user icet jupyter
References
- ICET – A Python Library for Constructing and Sampling Alloy Cluster Expansions
M. Ångqvist, W. A. Muñoz, J. M. Rahm, E. Fransson, C. Durniak, P. Rozyczko, T. H. Rod, and P. Erhart
Advanced Theory and Simulation 2, 1900015 (2019)