SDL Materials Design presents
Machine Learning for Materials Science
Date: October 21 - 23, 2024, 9.00 am - 01.00 pm; Format: online
Short abstract:
Data Science and Machine Learning are seen as the “Forth Paradigm” in Materials Science and are reshaping the research direction in many areas. In this training, the students/participants will gain an overview and obtain hand-on experience on the most relevant machine learning algorithms for theoretical simulations, experimental characterization, and in general statistical analysis in materials science. The participants will work with established packages within Python to develop their own simple machine learning based programs, and are going to tackle a challenging project. Though exemplary datasets, the participants will practice to apply appropriate methods to basic materials science problems, in particular Machine Learning assisted image segmentation, Machine-learning interatomic potentials, microstructure-property correlation analysis, and data-driven multiscale modeling.
Language: English
Capacity: 300
Requirement: On day 2 and 3, own laptop with working Google Collab account/ Jupyternotebook installed. A python installation with tensorflow, scikit-learn, ase, dscribe, and pacemaker
Further information: This training is organized in cooperation with the CRC FLAIR 1548 FLAIR – Start – FLAIR – TU Darmstadt and GRK 2561 MatCom-ComMat IAM – WerkstoffkundeKooperationen – GRK 2561.
Registration will follow soon
Agenda
October 21, 2024, Day 1, 9.00 am – 1.00 pm : “Introduction to Machine Learning (ML) in Material Science (MS)
In the first day, the participants will learn basic concepts of ML, followed by more in-depth content of shallow learning, deep learning and Bayesian optimization. The lecturer will contextualize the techniques on MS-related topics such as microstructure characterization, microstructure-property relations and further.
October 22, 2024, Day 2, 9.00 am – 1.00 pm: “Short refresher on Python and Introduction to PyTorch. Hands-On exercise with focus on microstructures”
In the second day, an introduction on google Colab and some basics of Python and the deep learning in
PyTorch framework will be introduced. The participants will learn how to program a simple regression task using gradient descent method followed by more tasks on microstructure characterization of 3D voxel-based microstructures, as well as training a surrogate model using simple descriptors for microstructure-property relations. In addition, participants will learn how to use convolutional neural networks to correlate the microstructure to the desired property. Further hands-on examples of Bayesian optimization on the same dataset will be provided as homework.
October 23, 2024, Day 3, 9.00 am – 1.00 pm “Introduction to local descriptors of atomic environments and how to connect them to atomic properties through machine-learning. hands-on session: How to train simple models for atomic properties (e.g., NMR shifts)”
The participants will learn how we can describe local environments of atoms (number of neighbours, bond angles, bond distances) by different descriptors. In a hands-on session we will show, how we can relate the descriptors to atomic properties, e.g., NMR shifts, atomic energies, through machine-learning. Building on that we will introduce pacemaker, a machine-learning interatomic potential fitting framework, and provide a database for an exemplary fitting hands-on session.
A more comprehensive schedule and access to all learning materials and code will be provided to registered participants a few days prior to the workshop.
Contact person
Aditi Mohadarkar
TU Darmstadt
Vasilios Karanikolas
TU Darmstadt