Challenges and Future Developments

View at publisher Summarise with VIRAT
Authors: Xue Jiang, Yanjing Su
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This chapter examines the transformative impact of natural language processing (NLP) and large language models (LLMs) on materials science, focusing on challenges and future advancements. Key obstacles include enhancing numerical understanding, quantitative prediction, and scientific reasoning. Numerical proficiency is improved through high-quality datasets and domain-specific fine-tuning. Quantitative prediction is advanced by integrating language encoders with property-prediction networks and deploying AI agents for computational resource integration. Efficiency is optimized via knowledge distillation, balancing model size and performance. Scientific reasoning is augmented through retrieval-augmented generation and advanced training protocols like reinforcement learning, as demonstrated by the DeepSeek-R1 model. Current applications in materials design have shown promise in predicting properties and suggesting compositions, particularly in alloy and polymer engineering. Future advancements are expected to refine these capabilities and integrate language models with computational and experimental tools for real-time insights. The evolution of AI is set to revolutionize materials science by enabling autonomous experimentation, cross-disciplinary synergy, and the development of a hybrid workforce skilled in both AI and materials engineering, ultimately driving sustainable and intelligent manufacturing.

Fields of Study & Subjects

Machine Learning in Materials Science Topic Modeling Artificial Intelligence Applications

Publication details

Journal / source
Wiley eBooks
Pages
551–558
Publication date
17 March 2026
Language
EN
Document type
Book chapter