Training Strategies for Coarse-grained Machine Learning Potentials

Institut

Professur für Multiscale Modeling of Fluid Materials

Typ

Bachelorarbeit / Semesterarbeit / Masterarbeit /

Inhalt

theoretisch /  

Beschreibung

Project Description

Machine learning potentials (MLPs) are an emerging approach that

bridges the gap between quantum-mechanical accuracy and classical

molecular dynamics. Coarse-graining, which reduces the system’s

degrees of freedom, is essential for accessing larger length scales and

longer timescales. The application of MLPs to coarse-grained systems

has only recently begun, and in this project, we aim to explore and

optimise different training strategies.

Project Aims

• Explore different training strategies for coarse-grained machine

learning potentials

• Augment data pipelines and machine learning model architecture

• Perform ablation studies to assess model performance

Voraussetzungen

• Programming experience: proficiency in Python

• Experience working with molecular data

Möglicher Beginn

sofort

Kontakt

Franz Görlich
f.goerlichtum.de

Ausschreibung