Simplifying Computational Simulations: Using Large Language Models for Automated Research in Materials Science
Date and time
Thursday, May 9, 2024; 12:00 - 1:00 PM EDT
Presenter
Ethan Holbrook, Materials Engineering PhD candidate at Purdue University
Register
https://purdue.webex.com/weblink/register/rbc8f3b395f5670d2b0d8913417846a94
In this hands-on workshop, we will explore the transformative role of Large Language Models (LLMs) in computational materials science, emphasizing their integration into the LAMMPS molecular dynamics (MD) simulation framework. We introduce a user-friendly Graphical User Interface (GUI) designed to facilitate LAMMPS simulations that is enhanced with LLM-driven script generation and result querying capabilities. This GUI makes complex MD simulations accessible to a broader scientific audience. Additionally, we will discuss the implications of ad hoc languages in scientific research, highlighting how they create access barriers and the potential for LLMs to overcome these challenges. Our work represents a significant step in automating and streamlining research processes, thereby accelerating discoveries in materials science.
Presenter bio
Ethan Holbrook is a Materials Engineering PhD candidate at Purdue University working under the advisement of Dr. Alejandro Strachan. Previously, he completed his Materials Science & Engineering B.S. at Wright State University in 2021. His research focuses on atomistic modeling using LAMMPS. Through simulation, he investigates energetic material interface systems and more broadly seeks to develop FAIR workflows for generating, storing, and exploring scientific data. He plans to continue working with atomistic models and expand methods for doing FAIR science. |