| About this Abstract |
| Meeting |
MS&T24: Materials Science & Technology
|
| Symposium
|
High Entropy Materials: Concentrated Solid Solutions, Intermetallics, Ceramics, Functional Materials and Beyond V
|
| Presentation Title |
Large Language Model-assisted Intuitive Materials Design |
| Author(s) |
Quanliang Liu, Maciej Polak, So Yeon Kim, Md Al Amin Shuvo, Hrishikesh Deodhar, Jeongsoo Han, Dane Morgan, Hyunseok Oh |
| On-Site Speaker (Planned) |
Hyunseok Oh |
| Abstract Scope |
Materials design has traditionally relied on the designer’s intuition, drawn from empirical knowledge, and quantitative predictions grounded in scientific principles. Although these predictions are significantly enhanced by computational methods, utilizing computation to support human intuition has been challenging due to the difficulty of reducing complex scientific principles into a manageable set of computable variables. In this study, we harness the advanced pattern recognition and generative capabilities of the recent large language model, GPT-4, to explore an extensive repository of materials science literature. We developed prompts to synthesize materials system charts—a structured framework that comprehends the dynamic interplay among the Processing, Structure, Property, and Performance aspects of materials—from individual research papers. The LLM can then integrate these system charts from multiple sources (e.g., 50 papers) to compile a comprehensive handbook or review paper and generate innovative scientific hypotheses for novel material designs, including cryogenic high entropy alloys. |