EXTENDED ABSTRACT: The success of artiffcial intelligence (AI) in materials research heavily relies on the integrity of structured data and the construction of precise descriptors, placing strong demands on researchers’ expert knowledge. In recent years, large language models have demonstrated their general “intelligence” capabilities via large-scale data, vast neural networks, self-supervised learning and powerful hardware, which show the potential to transform conventional MLdriven materials research by directly harnessing historical text knowledge. In this study, we present an end-to-end pipeline from materials text to properties for steels based on a large language model. The objective is to enable quantitative predictions of properties with high-accuracy and explore new steels. The pipeline includes a materials language encoder, named SteelBERT, and a multimodal deep learning framework that maps the composition and text sequence of complex fabrication processes to mechanical properties. We demonstrate high accuracy on mechanical properties, including yield strength (YS), ultimate tensile strength (UTS), and elongation (EL) by predicting determination coefffcients (R2) reaching 78.17% (±3.40%), 82.56% (±1.96%), and 81.44% (±2.98%) respectively. Further, through an additional ffne-tuning strategy for the design of speciffc steels with small datasets, we show how the performance can be reffned. With only 64 experimental samples of 15Cr austenitic stainless steels, we obtain an optimized model with R2 of 89.85% (±6.17%), 88.34% (±5.95%) and 87.24% (±5.15%) for YS, UTS and EL, that requires the user to input composition and text sequence for processing and which outputs mechanical properties. The model efffciently optimizes the text sequence for the fabrication process by suggesting a secondary round of cold rolling and tempering to yield an exceptional YS of 960 MPa, UTS of 1138 MPa, and EL of 32.5%, exceeding those of reported 15Cr austenitic stainless steels

Keywords: Property prediction; Steel design; Materials language model

REFERENCES:

[1] S. Tian, X. Jiang, W. Wang, et al. Acta Materialia (Under Review), 2024.