Machine Learning Meets Molecular Dynamics

Molecular simulations, often likened to a computational microscope, offer atomic-level resolution of complex chemical behavior and phenomena beginning from the underlying fundamental physics.  Such simulations generally rely on 1) an approximate method for propagating the nuclear coordinates of the system (e.g., traditional molecular dynamics, which treats the nuclei as classical particles governed by Newton's equations of motion), and 2) a sufficiently accurate description of the energy of the system with respect to its atomic coordinates, i.e., a potential energy surface (PES), which is an approximate solution to the Schrödinger equation in the Born-Oppenheimer approximation.  While ab initio electronic structure methods are highly-accurate, their high computational cost and poor scaling with system size routinely necessitates the use of approximate PESs, even for relatively small systems.  

Research projects for the spring semester entail the development of neural network representations of potential energy surfaces.  Two possible avenues of research are currently envisioned.  The first is the construction of surfaces for modeling proton-coupled electron transfer, an important reaction mechanism prevalent in energy applications (e.g., fuel cells, batteries, etc.) and biological processes (e.g., photosynthesis).  The second is the development of PESs for ionic liquids, a unique class of materials with myriad applications, in the interest of undertaking a systematic study of their various structural and thermodynamic properties.  Although the downstream applications and systems of interest are chemical, the actual research to be done is suitable for an array of majors, including computer science, mathematics, physics, and chemistry majors.

 

To apply for this position:

1) Meet with Prof. Brown to talk about the project and your experiences.

2) Do not apply via this website. Instead, apply via the Chemistry Luke Scholars Application. You may find it helpful to look at the form now before you start your application process.

All parts of the application should be submitted by 5pm, December 10th for complete consideration.

 

Name of research group, project, or lab
Peace Frog
Why join this research group or lab?

Learn how to develop a general machine learning model, a highly sought after and transferable skill

Work on an interdisciplinary research project

Enjoy a flexible mode of work

Logistics Information:
Project categories
Chemistry
Computer Science
Mathematics
Physics
Machine Learning
Student ranks applicable
Sophomore
Junior
Senior
Time commitment
Spring - Part Time
Compensation
Academic Credit
Paid Research
Number of openings
1
Contact Information:
Mentor name
Sandra Brown
Mentor email
sabrown@hmc.edu
Mentor position
Visiting Assistant Professor, Chemistry
Name of project director or principal investigator
Sandra E. Brown
Email address of project director or principal investigator
sabrown@g.hmc.edu
1 sp. | 0 appl.
Hours per week
Spring - Part Time
Project categories
Chemistry (+4)
ChemistryComputer ScienceMathematicsPhysicsMachine Learning