| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
Artificial Intelligence Applications in Integrated Computational Materials Engineering
|
| Presentation Title |
High-Throughput and Robust Materials Design Hypothesis Generation via a RAG-Enhanced Large Language Model |
| Author(s) |
Quanliang Liu, Maciej Polak, So yeon Kim, MD Al Amin Shuvo, Hrishikesh Shridhar Deodhar, Jeongsoo Han, Dane Morgan, Hyunseok Oh |
| On-Site Speaker (Planned) |
Quanliang Liu |
| Abstract Scope |
The rapid advancement of Large Language Models (LLMs) offers significant benefits for alloy design by managing and transferring extensive information across various domains. However, applying LLMs faces several challenges, including generating hallucinated information, overly generic responses, and lacking clear sources. In this work, we demonstrate that using a metallurgy literature-based Retrieval-Augmented Generation (RAG) model with GPT-4 can effectively overcome these challenges. Specifically, with the metallurgy database prepared, we can systematically extract Processing-Microstructure-Property (PMP) relationships using prompt engineering. Not only do the Materials System Charts across 70 scientific articles exhibit a high average Human-Machine Index (HMI) score of 0.9 and high mechanism score, but we also compile an alloy design handbook and demonstrate innovative hypothesis generation for materials such as cryogenic high entropy alloys and halide solid electrolytes. This automated approach facilitates groundbreaking research, reduces time and costs, and maintains scientific rigor, showcasing the potential of LLMs in advancing materials science. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Computational Materials Science & Engineering, Machine Learning, Other |