Machine Learned Potentials for MD Simulations

These are the weekly reports written at the end of my 3rd Year when my previous project was scrapped due to Covid-19. These reports outline my progress from scratch in chronological order. They were written in 3 weeks and total 19,000 words.

The final grade for these reports and viva voce was 88%. Below are the Laboratory Notes documenting the process.

This is a rendering of a graphene lattice Simulated using the Behler-Parrinello Method in LAMMPS. Although there are some imperfections in the final result, the machine learned interatomic potential sucessfully replicated the correct bond angles and atom spacing for a graphene crystal.