Master Thesis: Machine-Learning potentials for Dislocations

Master Thesis: Active Learning Algorithms for Simulating Dislocations in Metals

Machine-Learning potentials for Dislocations

These topics inspire you or you are already familiar with them? Then you are the right person for this position!

What are we looking for:
A motivated master student from materials science, physics, or computational sciences (e.g., computational mechanics, applied mathematics) with one (or more ☺) of the following skills:
• crystallography/atomic-scale mechanics,
• atomistic simulation,
• linear algebra/machine-Learning methods,
• programming (Python, C++).

Your challenge:
Machine-learning potentials are a relatively new class of interatomic potentials aiming at atomistic simulations with quantum-mechanical accuracy— but a fraction of the cost. In this work, you will extend an active learning algorithm for training such potentials to dislocations in metals. More precisely, your challenge is to
• develop algorithms for constructing training sets for the machine-learning potential to be used in simulations of fcc dislocations,
• testing and validation of the developed algorithms.
The precise tasks may depend on your background and personal preferences.

Our offer:
An employment with immediate start and a gross salary of € 3.000,- for six months.

Please send your complete application documents by email.
We are looking forward to knowing you!
bewerbung@mcl.at