硅酸盐学报

AI+材料科学：大模型在新材料领域的研究进展与展望

PDF下载

张林桐 ¹ , 刘传宝 ² , 白洋 ²

硅酸盐学报 | 综合评述 2026,54(7): 2363-2379

收起

硅酸盐学报 |综合评述 2026 , 54 (7) : 2363 -2379

AI+材料科学：大模型在新材料领域的研究进展与展望

全屏

[Author(id=1320107546338243571, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=ltZhang@ustb.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1320107546413741045, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, authorId=1320107546338243571, language=EN, stringName=Lintong ZHANG, firstName=Lintong, middleName=null, lastName=ZHANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=¹, address=^1.Basic Experimental Center for Natural Science, University of Science and Technology Beijing, Beijing 100083, China, bio={"content":"

ZHANG Lintong (1996-), female, Master, Engineer. E-mail: ltZhang@ustb.edu.cn

"}, bioImg=null, bioContent=

ZHANG Lintong (1996-), female, Master, Engineer. E-mail: ltZhang@ustb.edu.cn

, aboutCorrespAuthor=null), CN=AuthorExt(id=1320107546480849910, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, authorId=1320107546338243571, language=CN, stringName=张林桐, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=¹, address=^1．北京科技大学自然科学基础实验中心，北京 100083, bio={"content":"

张林桐（1996—），女，硕士，工程师。

"}, bioImg=null, bioContent=

张林桐（1996—），女，硕士，工程师。

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1320107546166277100, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, xref=1., ext=[AuthorCompanyExt(id=1320107546174665709, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, companyId=1320107546166277100, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=^1.Basic Experimental Center for Natural Science, University of Science and Technology Beijing, Beijing 100083, China), AuthorCompanyExt(id=1320107546183054318, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, companyId=1320107546166277100, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=^1．北京科技大学自然科学基础实验中心，北京 100083)])]), Author(id=1320107546556347384, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1320107546707342330, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, authorId=1320107546556347384, language=EN, stringName=Chuanbao LIU, firstName=Chuanbao, middleName=null, lastName=LIU, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=², address=^2.Institute for Advanced Materials and Technology, University of Science and Technology Beijing, Beijing 100083, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1320107546770256891, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, authorId=1320107546556347384, language=CN, stringName=刘传宝, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=², address=^2．北京科技大学新材料技术研究院，北京 100083, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1320107546262746095, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, xref=2., ext=[AuthorCompanyExt(id=1320107546271134704, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, companyId=1320107546262746095, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=^2.Institute for Advanced Materials and Technology, University of Science and Technology Beijing, Beijing 100083, China), AuthorCompanyExt(id=1320107546279523313, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, companyId=1320107546262746095, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=^2．北京科技大学新材料技术研究院，北京 100083)])]), Author(id=1320107546837365757, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=baiy@mater.ustb.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=1, authorType=1, ext={EN=AuthorExt(id=1320107547114189823, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, authorId=1320107546837365757, language=EN, stringName=Yang BAI, firstName=Yang, middleName=null, lastName=BAI, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=², address=^2.Institute for Advanced Materials and Technology, University of Science and Technology Beijing, Beijing 100083, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1320107547189687296, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, authorId=1320107546837365757, language=CN, stringName=白洋, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=², address=^2．北京科技大学新材料技术研究院，北京 100083, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1320107546262746095, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, xref=2., ext=[AuthorCompanyExt(id=1320107546271134704, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, companyId=1320107546262746095, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=^2.Institute for Advanced Materials and Technology, University of Science and Technology Beijing, Beijing 100083, China), AuthorCompanyExt(id=1320107546279523313, tenantId=1146029695717560320, journalId=1263187385517883426, articleId=1320107545495188457, companyId=1320107546262746095, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=^2．北京科技大学新材料技术研究院，北京 100083)])])]

张林桐¹, 刘传宝², 白洋²

作者信息

^1．北京科技大学自然科学基础实验中心，北京 100083

^2．北京科技大学新材料技术研究院，北京 100083

通讯作者:

白洋（1979—），男，博士，研究员。

作者简介:

张林桐（1996—），女，硕士，工程师。

AI + Materials Science: Research Progress and Perspectives on Large Models in New Materials

Lintong ZHANG¹, Chuanbao LIU², Yang BAI²

Affiliations

^1.Basic Experimental Center for Natural Science, University of Science and Technology Beijing, Beijing 100083, China

^2.Institute for Advanced Materials and Technology, University of Science and Technology Beijing, Beijing 100083, China

Corresponding Author:

{BAI} Yang (1979-), male, Ph.D., Research Fellow. E-mail: baiy@mater.ustb.edu.cn

ZHANG Lintong (1996-), female, Master, Engineer. E-mail: ltZhang@ustb.edu.cn

出版时间: 2026-04-23 doi: 10.14062/j.issn.0454-5648.20250943

文章导航

摘要

收起

随着数据驱动科学研究第四范式的深入发展，以大模型为代表的人工智能技术正深度重构材料研发模式。本文系统综述了材料大模型在晶体结构设计与发现、性能预测与基础知识挖掘、具体材料垂类大模型、冶金工艺优化及多智能体协同框架的应用场景。在晶体结构设计与发现方面，GNoME、CrystaLLM分别通过主动学习与自回归建模，显著提升了稳定新材料的发现速率；在性能预测与基础知识挖掘领域，MatChat与Galactica等利用检索增强生成与大规模科学语料预训练，实现了从海量非结构化文献中精准提取规律并预测材料属性；针对具体材料垂类大模型，PolyLLMem、MatMind及ChipNeMo等通过整合高分子、无机非金属及半导体设计的专业先验知识，有效弥合了通用大模型在特定学科中的知识断层；在冶金工艺优化层面，SteelBERT及PDGPT等通过关联工艺参数与显微组织，为工业级黑盒问题的协同优化提供了高保真度决策支持；在多智能体协同框架维度，AtomAgents和AgentMat等系统通过层次化架构调度多种专业工具，驱动了全流程闭环的智能化演进。最后，针对当前材料大模型面临的数据稀缺性、物理一致性约束及黑盒可解释性等核心挑战，提出了未来在多模态融合、物理信息引导以及自主实验室方向的发展展望，旨在为材料基因工程与智能化研发提供理论参考与技术指南。

关键词

材料科学 / 人工智能 / 材料大模型 / 智能体

Abstract

收起

The field of materials science as a fundamental cornerstone for technological advancements and industrial transformation in society currently undergoes a profound paradigm shift propelled by artificial intelligence (AI). Conventional research and development (R&D) of materials, heavily reliant on empirical "trial-and-error" approaches and expert intuition, are inherently constrained by protracted R&D cycles and substantial financial costs. The advancement of the Materials Genome Initiative (MGI) has accumulated various large-scale scientific databases covering structure, property and process, but it remains a challenge that efficiently mines inherent law from these complex and multimodal massive data and integrates interdisciplinary information. The recent emergence and rapid evolution of large-scale foundation models offer a transformative solution to this critical bottleneck. Pre-trained on massive and diverse scientific corpora that include research literature, structural databases, and computational outputs, these models are evolving beyond mere pattern-recognition tools. They advance into sophisticated cognitive engines capable of decoding specialized material languages, integrate physical laws across multiple scales, and ultimately restructure the entire material discovery and optimization workflow. This review systematically examines a pivotal role of Materials Foundation Models (MFMs) in this AI-driven revolution. It delves into their core methodological foundations, showcases landmark applications across key sub-disciplines, and outlines the trajectory leading toward a future of fully intelligent and autonomous materials discovery.

The significance of MFMs is fundamentally rooted in their capacity to establish a unified and comprehensive evaluative framework, which is structured around their core underlying training paradigms. This framework is primarily built on three synergistic pillars, including large-scale pre-training on structural and textual corpora, supervised fine-tuning for specific property regressions and reinforcement learning from physical feedback, which aligns model outputs with inviolable thermodynamic and kinetic constraints. This review deconstructs their capabilities and limitations via analyzing models through this tripartite lens. A detailed architectural analysis of pioneering vertical domain models, such as MatMind for inorganic non-metallics and SteelBERT for metallurgy is given. We assess the specific mechanisms through that these models encode a deep domain-specific knowledge, including crystallographic symmetry rules, polymer chain topology dynamics, and solid-state reaction kinetics, into their latent representation spaces, thereby achieving a superior performance over general-purpose models. Furthermore, this review critically analyzes the sophisticated bridging mechanisms enabled by multi-agent frameworks. We examine how multi-agent systems like AtomAgents function as cognitive hubs, connecting and orchestrating high-throughput computational pipelines with experimental databases through unified natural language interfaces. These frameworks effectively manage the complex orchestration of diverse simulation tools and characterization equipment, thereby automating and accelerating the entire "design-simulate-characterize-optimize" cycle.

The research advances demonstrate a disruptive impact of these models across the materials R&D chain. MFMs significantly enhance design efficiency and prediction accuracy via suppressing "scientific hallucinations" through the integration of explicit physical rules and attention visualization techniques. In the realm of crystal discovery, the GNoME model effectively expands the library of stable inorganic crystals by ten times through an active learning closed-loop. For structure generation, models like CrystaLLM demonstrate an ability to capture atomic-scale symmetry breaking during phase stability predictions via deconstructing attention distributions over crystal descriptors. In specialized sectors, vertical models such as PolyLLM and ChipNeMo effectively bridge knowledge gaps via embedding domain-specific heuristics, while in the metallurgical sector, SteelBERT and PDGPT achieve a high accuracy in predicting steel properties via correlating process parameters with microstructural evolution. Moreover, multimodal frameworks such as MatterChat bridges a gap between natural language and atomic-scale features, where attention weights are automatically assigned to active-site orbitals, which is highly consistent with coordination field theory. The integration of agentic workflows further demonstrates that these foundation models can manage the entire cycle with an unprecedented precision.

Summary and Prospects

MFMs decisively evolve from auxiliary computational tools into core engines for end-to-end intelligent R&D. They offer a unifying framework for the longstanding inverse design problem, enabling a systematic search of the MGI guided by target performance. Nevertheless, several challenges must be overcome to realize their full potential. Key limitations include the data scarcity for novel or extreme-condition materials, a need for even stricter integration of physical laws to ensure causal reasoning beyond correlation, and a "black-box" interpretability issue that can hinder trust in critical applications. Future development must focus on creating models with intrinsic causal reasoning capabilities, moving from statistical prediction to understanding the underlying mechanism of material features. This will be coupled with the integration of MFMs into fully autonomous laboratory platforms, where real-time perception from in-situ characterization and robotic control will enable closed-loop, self-driving experimentation. Ultimately, the convergence of AI and materials science steers a field from task automation toward high-level cognitive intelligence. This transition promises to fundamentally disrupt a conventional design logic, accelerating the discovery and deployment of advanced materials and ushering in a novel era of intelligent materials genome engineering.

Key words

materials science / artificial intelligence / materials foundation models / agent

引用本文

张林桐, 刘传宝, 白洋. AI+材料科学：大模型在新材料领域的研究进展与展望. 硅酸盐学报, 2026 , 54 (7) : 2363 -2379 . DOI: 10.14062/j.issn.0454-5648.20250943

Lintong ZHANG, Chuanbao LIU, Yang BAI. AI + Materials Science: Research Progress and Perspectives on Large Models in New Materials[J]. Journal of the Chinese Ceramic Society, 2026 , 54 (7) : 2363 -2379 . DOI: 10.14062/j.issn.0454-5648.20250943

基金

收起

国家自然科学基金(92463311)

参考文献

收起

文献

收起

参考文献引证文献

排序方式：

2026年第54卷第7期

PDF下载

引用本文

BibTeX

文章信息

doi: 10.14062/j.issn.0454-5648.20250943

接收时间：2025-12-30
首发时间：2026-10-22
出版时间：2026-04-23

补充材料

相关文章

文章信息

作者

出版历史

收稿日期：2025-12-30
修回日期：2026-01-30

基金

国家自然科学基金(92463311)

作者信息

^1．北京科技大学自然科学基础实验中心，北京 100083

^2．北京科技大学新材料技术研究院，北京 100083

通讯作者:

白洋（1979—），男，博士，研究员。

参考文献

分享链接

https://castjournals.cast.org.cn/joweb/gsyxb/CN/10.14062/j.issn.0454-5648.20250943

分享至

全文二维码

扫描看全文

引用本文

BibTeX

本文的引用情况

2种不同金属材料的力学参数

科 Family	属数 Number of genus	种数 Number of species	占总种数比例 Percentage of total species (%)	属 Genus	种数 Number of species	占总种数比例 Percentage of total species (%)
鹅膏菌科Amanitaceae	2	11	5.26	鹅膏菌属 Amanita	10	4.78
小菇科 Mycenaceae	2	12	5.74	丝盖伞属 Inocybe	5	2.39
多孔菌科 Polyporaceae	8	14	6.70	蜡蘑属 Laccaria	5	2.39
红菇科 Russulaceae	3	23	11.00	小皮伞属 Marasmius	6	2.87
				小菇属 Mycena	11	5.26
				光柄菇属 Pluteus	5	2.39
				红菇属 Russula	17	8.13
				栓菌属 Trametes	5	2.39

关闭全屏

BibTeX
EndNote
RefWorks
TxT