Industrial foundation model: architecture, key technologies, and typical applications

With the development of the new generation of artificial intelligence, foundation models (such as GPT-4o), leveraging extensive training data, parameters, and computational power, have emerged unique generative capabilities, generalization ability, and natural interaction skills, showcasing tremendous potential to transform industrial production processes. However, despite significant advancements in various fields, such as natural language processing, the exploration of foundation models in industrial applications remains in its infancy. There is a lack of systematic research on industrial foundation models. Industrial applications have unique characteristics, such as heterogeneous cross-modal data, complex and diverse specialized scenarios, long-process integrative decision-making, and high demands for reliability and real-time performance. Generalized foundation models are not directly applicable to solving complex industrial problems. Therefore, there is an urgent need for new research on the basic theories and key technologies of industrial foundation models. This paper systematically discusses the challenging issues, concepts, architecture, construction methods, key technologies, and typical applications of industrial foundation models. It analyzes five key challenges and proposes a new definition and system architecture for industrial foundation models. Additionally, it outlines a four-stage construction method and elaborates on the key technologies of the industrial foundation model. Then, based on the six core application capabilities of the proposed industrial foundation model, the paper explores typical application scenarios of the industrial foundation model for the entire life cycle and provides examples of the application of the industrial foundation model prototype system in generative artificial intelligence. Finally, the paper discusses future research directions and openness issues of the industrial foundation model, aiming to provide comprehensive guidance on the basic theories, key technologies, and industrial applications for the development of this new research direction.

Authors

Publication Details

Journal
Scientia Sinica Informationis
Published
2024-10-08
DOI
https://doi.org/10.1360/ssi-2024-0185
Citations
6
Primary Topic
Probabilistic and Robust Engineering Design
Type
article
Field-Weighted Citation Impact
2.77
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Industrial foundation model: architecture, key technologies, and typical applications

Liyuanjun Lai, Di Huang, Haiteng Wang, Bohu Li et al.
6 citations
Scientia Sinica Informationis
Probabilistic and Robust Engineering Design
2.77
article

Industrial foundation model: architecture, key technologies, and typical applications

Liyuanjun Lai, Di Huang, Haiteng Wang, Bohu Li, Shixiang Li, Zhang Lin, Yuqing Wang, Wenjun Wu, Zhaoxuan Jia, Jiabao Dong, Lei Ren
article en
6 citations

Abstract

With the development of the new generation of artificial intelligence, foundation models (such as GPT-4o), leveraging extensive training data, parameters, and computational power, have emerged unique generative capabilities, generalization ability, and natural interaction skills, showcasing tremendous potential to transform industrial production processes. However, despite significant advancements in various fields, such as natural language processing, the exploration of foundation models in industrial applications remains in its infancy. There is a lack of systematic research on industrial foundation models. Industrial applications have unique characteristics, such as heterogeneous cross-modal data, complex and diverse specialized scenarios, long-process integrative decision-making, and high demands for reliability and real-time performance. Generalized foundation models are not directly applicable to solving complex industrial problems. Therefore, there is an urgent need for new research on the basic theories and key technologies of industrial foundation models. This paper systematically discusses the challenging issues, concepts, architecture, construction methods, key technologies, and typical applications of industrial foundation models. It analyzes five key challenges and proposes a new definition and system architecture for industrial foundation models. Additionally, it outlines a four-stage construction method and elaborates on the key technologies of the industrial foundation model. Then, based on the six core application capabilities of the proposed industrial foundation model, the paper explores typical application scenarios of the industrial foundation model for the entire life cycle and provides examples of the application of the industrial foundation model prototype system in generative artificial intelligence. Finally, the paper discusses future research directions and openness issues of the industrial foundation model, aiming to provide comprehensive guidance on the basic theories, key technologies, and industrial applications for the development of this new research direction.

Scientia Sinica InformationisVol. 54(11)
Openalex Percentile: Top 9%
Probabilistic and Robust Engineering Design
2.77
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.