Hyperparameter Optimization and Benchmarking of Machine Learning Potentials for Molecular Dynamics Simulations

Institut

Professur für Multiscale Modeling of Fluid Materials

Typ

Bachelorarbeit / Semesterarbeit / Masterarbeit /

Inhalt

theoretisch /  

Beschreibung

Machine learning potentials (MLPs) have revolutionized molecular dynamics (MD) simulations by

providing a bridge between the accuracy of quantum-mechanical methods and the efficiency of

classical force fields. However, the performance of MLPs heavily depends on the choice of

hyperparameters, such as network architecture, learning rate, and training data composition.

Systematic hyperparameter optimization and benchmarking are essential to ensure that MLPs are

accurate, transferable, and computationally efficient.

This thesis focuses on the hyperparameter optimization and benchmarking of developed MLPs,

particularly those based on Graph Neural Networks (GNNs). The project will involve designing and

implementing a framework for hyperparameter tuning, evaluating the performance of optimized MLPs

on diverse datasets, and benchmarking their accuracy and efficiency against traditional force fields

and ab initio methods.

Voraussetzungen

We are looking for a highly motivated student with the following qualifications:

● Machine Learning Practice: Experience with training and evaluating machine learning

models, particularly neural networks.

● Programming Expertise: Strong proficiency in Python. Familiarity with machine learning

frameworks (e.g., PyTorch, JAX)

● Molecular Dynamics Knowledge: Basic understanding of MD simulations and force fields is

advantageous but not mandatory.

● Motivation and Curiosity: A strong interest in machine learning, computational chemistry,

and materials modeling.

Möglicher Beginn

sofort

Kontakt

Michał Sanocki
Tel.: 783880014
m.sanockitum.de

Ausschreibung