Level: advanced
English language
Format: online
Course duration: 6 hours lectures/15 hours (0.5 ECTS)
Start: 06/17/2024 (access to video lectures and other materials on the platform eduportal.kau.org.ua)
Target Audience: This course is designed for researchers, engineers, undergraduates and graduate students working in the field of materials science, computational materials science and related fields and have a good knowledge of atomistic simulations and are also interested in applying machine learning in their practice.
Teacher: Olexander Vasiliev, Ph.D., Associate Professor, Head of the Department of Applied Mathematics and Computational Experiments in Materials Science, Institute of Materials Science Problems named after. THEM. Frantsevich NAS of Ukraine, Associate Professor of the Department of Applied Physics of Materials Science KAU.
The course explores the complexities of creating machine learning models that complement and extend the capabilities of quantum mechanics methods in atomistic materials simulations. The training materials cover topics from the specific architecture of atomistic machine learning models through the details of data collection and preparation to training, validating and applying models. The course provides practical guidance on the steps you need to take to create models and overcome some of the challenges along the way.
Required knowledge
average level of machine learning knowledge
good knowledge of atomistic material modeling using density functional theory
intermediate level of Python programming
Acquired skills
Upon completion of the course, participants will be able to
understand the specifics of machine learning models for atomistic simulations
choose a base model for your task
collect and prepare calculation data for training machine learning models
understand how to train, validate and apply atomistic machine learning models
The course is free of charge. Upon completion of the course, all students will receive certificates indicating the number of ECTS credits.