Position innerhalb des Seitenbaumes

Institut für Automatisierungstechnik und Softwaresysteme
Institut
Team
Xia, Yuchen

Yuchen Xia

M.Sc.

Akademischer Mitarbeiter
Institut für Automatisierungstechnik und Softwaresysteme

Kontakt

+49 711 685 67307
+49 711 685 67302
E-Mail
Google Scholar

Pfaffenwaldring 47
70550 Stuttgart
Raum: 1.138

Publikationen

Zeitschriften und Konferenzen:

2024
1. A. Shoshi, Y. Xia, A. Fieschi, T. Ackermann, P. Reimann, M. Weyrich, B. Mitschang, T. Bauernhansl, and R. Miehe, “A Flexible Digital Twin Framework for ATMP Production – Towards an efficient CAR T Cell Manufacturing,” Procedia CIRP, vol. 125, pp. 124–129, Jan. 2024.
  - BibTeX
  BibTeX
  @article{Shoshi2024, author = {Shoshi, Arber and Xia, Yuchen and Fieschi, Andrea and Ackermann, Thomas and Reimann, Peter and Weyrich, Michael and Mitschang, Bernhard and Bauernhansl, Thomas and Miehe, Robert}, doi = {10.1016/J.PROCIR.2024.08.022}, issn = {2212-8271}, journal = {Procedia CIRP}, month = {01}, pages = {124-129}, publisher = {Elsevier}, title = {A Flexible Digital Twin Framework for ATMP Production – Towards an efficient CAR T Cell Manufacturing}, volume = 125, year = 2024 }
2. Y. Xia, N. Jazdi, and M. Weyrich, “Applying Large Language Models for Intelligent Industrial Automation: From Theory to Application: Towards Autonomous Systems with Large Language Models,” atp magazin, vol. 66, pp. 62–71, Jul. 2024.
  Zusammenfassung
  Dieser Artikel erforscht das transformative Potenzial großer Sprachmodelle (Large Language Models, LLMs) in der industriellen Automatisierung und stellt umfassende Ansätze für ihre Integration in komplexe Industriesysteme vor. Wir beginnen mit einem theoretischen Überblick über LLMs und erläutern ihre entscheidenden Fähigkeiten wie Interpretation, Aufgabenautomatisierung und autonome Agentenfunktionalität. Eine generische Methodik zur Integration von LLMs in industrielle Anwendungen wird dargelegt, die erläutert, wie LLMs für spezifische Aufgaben angewendet werden können. Vier Fallstudien demonstrieren den praktischen Einsatz von LLMs in verschiedenen industriellen Umgebungen: die Umwandlung unstrukturierter Daten in strukturierte Daten als Verwaltungsschalen-Modell, die Verbesserung des Wissensmanagements mit Dokumentendatenbanken durch ein Chat-system, die autonome Produktionsplanung und -steuerung für Automatisierungssysteme und die Interaktion mit Simulationsmodellen zur Bestimmung der Parametrisierung des Prozesses. Die Studien veranschaulichen die Fähigkeit von LLMs, diverse Aufgaben zu bewältigen und mit digitalen Zwillingen und Automatisierungssystemen zu interagieren, was nahlegt, dass Effizienz- und Produktivitätssteigerungen durch den strategischen Einsatz von LLM-Technologien in industriellen Umgebungen erreicht werden können.
  BibTeX
  @article{Xia2024, abstract = {Dieser Artikel erforscht das transformative Potenzial großer Sprachmodelle (Large Language Models, LLMs) in der industriellen Automatisierung und stellt umfassende Ansätze für ihre Integration in komplexe Industriesysteme vor. Wir beginnen mit einem theoretischen Überblick über LLMs und erläutern ihre entscheidenden Fähigkeiten wie Interpretation, Aufgabenautomatisierung und autonome Agentenfunktionalität. Eine generische Methodik zur Integration von LLMs in industrielle Anwendungen wird dargelegt, die erläutert, wie LLMs für spezifische Aufgaben angewendet werden können. Vier Fallstudien demonstrieren den praktischen Einsatz von LLMs in verschiedenen industriellen Umgebungen: die Umwandlung unstrukturierter Daten in strukturierte Daten als Verwaltungsschalen-Modell, die Verbesserung des Wissensmanagements mit Dokumentendatenbanken durch ein Chat-system, die autonome Produktionsplanung und -steuerung für Automatisierungssysteme und die Interaktion mit Simulationsmodellen zur Bestimmung der Parametrisierung des Prozesses. Die Studien veranschaulichen die Fähigkeit von LLMs, diverse Aufgaben zu bewältigen und mit digitalen Zwillingen und Automatisierungssystemen zu interagieren, was nahlegt, dass Effizienz- und Produktivitätssteigerungen durch den strategischen Einsatz von LLM-Technologien in industriellen Umgebungen erreicht werden können.}, author = {Xia, Yuchen and Jazdi, Nasser and Weyrich, Michael}, doi = {10.17560/ATP.V66I6-7.2739}, issn = {2364-3137}, journal = {atp magazin}, month = {07}, number = {6-7}, pages = {62-71}, publisher = {Vulkan-Verlag GmbH}, title = {Applying Large Language Models for Intelligent Industrial Automation: From Theory to Application: Towards Autonomous Systems with Large Language Models}, url = {https://ojs.di-verlag.de/index.php/atp_edition/article/view/2739}, volume = 66, year = 2024 }
  Link
  https://ojs.di-verlag.de/index.php/atp_edition/article/view/2739
3. Y. Xia, N. Jazdi, J. Zhang, C. Shah, and M. Weyrich, “Control Industrial Automation System with Large Language Models,” Sep. 2024.
  Zusammenfassung
  Traditional industrial automation systems require specialized expertise to operate and complex reprogramming to adapt to new processes. Large language models offer the intelligence to make them more flexible and easier to use. However, LLMs' application in industrial settings is underexplored. This paper introduces a framework for integrating LLMs to achieve end-to-end control of industrial automation systems. At the core of the framework are an agent system designed for industrial tasks, a structured prompting method, and an event-driven information modeling mechanism that provides real-time data for LLM inference. The framework supplies LLMs with real-time events on different context semantic levels, allowing them to interpret the information, generate production plans, and control operations on the automation system. It also supports structured dataset creation for fine-tuning on this downstream application of LLMs. Our contribution includes a formal system design, proof-of-concept implementation, and a method for generating task-specific datasets for LLM fine-tuning and testing. This approach enables a more adaptive automation system that can respond to spontaneous events, while allowing easier operation and configuration through natural language for more intuitive human-machine interaction. We provide demo videos and detailed data on GitHub: https://github.com/YuchenXia/LLM4IAS
  BibTeX
  @article{Xia2024, abstract = {Traditional industrial automation systems require specialized expertise to operate and complex reprogramming to adapt to new processes. Large language models offer the intelligence to make them more flexible and easier to use. However, LLMs' application in industrial settings is underexplored. This paper introduces a framework for integrating LLMs to achieve end-to-end control of industrial automation systems. At the core of the framework are an agent system designed for industrial tasks, a structured prompting method, and an event-driven information modeling mechanism that provides real-time data for LLM inference. The framework supplies LLMs with real-time events on different context semantic levels, allowing them to interpret the information, generate production plans, and control operations on the automation system. It also supports structured dataset creation for fine-tuning on this downstream application of LLMs. Our contribution includes a formal system design, proof-of-concept implementation, and a method for generating task-specific datasets for LLM fine-tuning and testing. This approach enables a more adaptive automation system that can respond to spontaneous events, while allowing easier operation and configuration through natural language for more intuitive human-machine interaction. We provide demo videos and detailed data on GitHub: https://github.com/YuchenXia/LLM4IAS}, author = {Xia, Yuchen and Jazdi, Nasser and Zhang, Jize and Shah, Chaitanya and Weyrich, Michael}, month = {09}, title = {Control Industrial Automation System with Large Language Models}, url = {https://arxiv.org/abs/2409.18009v1}, year = 2024 }
  Link
  https://arxiv.org/abs/2409.18009v1
4. Y. Xia, N. Jazdi, and M. Weyrich, “Enhance FMEA with Large Language Models for Assisted Risk Management in Technical Processes and Products,” 2024 IEEE 29th International Conference on Emerging Technologies and Factory Automation (ETFA), pp. 1–4, Sep. 2024.
  - BibTeX
  - Link
  BibTeX
  @article{xia2024enhance, author = {Xia, Yuchen and Jazdi, Nasser and Weyrich, Michael}, doi = {10.1109/ETFA61755.2024.10710996}, editor = {IEEE}, isbn = {979-8-3503-6123-0}, journal = {2024 IEEE 29th International Conference on Emerging Technologies and Factory Automation (ETFA)}, month = {09}, organization = {IEEE}, pages = {1--4}, title = {Enhance FMEA with Large Language Models for Assisted Risk Management in Technical Processes and Products}, url = {https://ieeexplore.ieee.org/document/10710996/}, year = 2024 }
  Link
  https://ieeexplore.ieee.org/document/10710996/
5. Y. Xia, Z. Xiao, N. Jazdi, and M. Weyrich, “Generation of Asset Administration Shell With Large Language Model Agents: Toward Semantic Interoperability in Digital Twins in the Context of Industry 4.0,” IEEE Access, vol. 12, pp. 84863–84877, 2024.
  - Zusammenfassung
  - BibTeX
  Zusammenfassung
  This research introduces a novel approach for achieving semantic interoperability in digital twins and assisting the creation of Asset Administration Shell (AAS) as digital twin model within the context of Industry 4.0. The foundational idea of our research is that the communication based on semantics and the generation of meaningful textual data are directly linked, and we posit that these processes are equivalent if the exchanged information can be serialized in text form. Based on this, we construct a 'semantic node' data structure in our research to capture the semantic essence of textual data. Then, a system powered by large language models is designed and implemented to process the 'semantic node' and generate standardized digital twin models (AAS instance models in the context of Industry 4.0) from raw textual data collected from datasheets describing technical assets. Our evaluation demonstrates an effective generation rate of 62-79%, indicating a substantial proportion of the information from the source text can be translated error-free to the target digital twin instance model with the generative capability of large language models. This result has a direct application in the context of Industry 4.0, and the designed system is implemented as a data model generation tool for reducing the manual effort in creating AAS model by automatically translating unstructured textual data into a standardized AAS model. The generated AAS model can be integrated into AAS-compliant digital twin software for seamless information exchange and communication. In our evaluation, a comparative analysis of different LLMs and an in-depth ablation study of Retrieval-Augmented Generation (RAG) mechanisms provide insights into the effectiveness of LLM systems for interpreting technical concepts and translating data. Our findings emphasize LLMs' capability to automate AAS instance creation and contribute to the broader field of semantic interoperability for digital twins in industrial applications. The prototype implementation and evaluation results are presented on our GitHub Repository: https://github.com/YuchenXia/AASbyLLM.
  BibTeX
  @article{Xia2024, abstract = {This research introduces a novel approach for achieving semantic interoperability in digital twins and assisting the creation of Asset Administration Shell (AAS) as digital twin model within the context of Industry 4.0. The foundational idea of our research is that the communication based on semantics and the generation of meaningful textual data are directly linked, and we posit that these processes are equivalent if the exchanged information can be serialized in text form. Based on this, we construct a 'semantic node' data structure in our research to capture the semantic essence of textual data. Then, a system powered by large language models is designed and implemented to process the 'semantic node' and generate standardized digital twin models (AAS instance models in the context of Industry 4.0) from raw textual data collected from datasheets describing technical assets. Our evaluation demonstrates an effective generation rate of 62-79%, indicating a substantial proportion of the information from the source text can be translated error-free to the target digital twin instance model with the generative capability of large language models. This result has a direct application in the context of Industry 4.0, and the designed system is implemented as a data model generation tool for reducing the manual effort in creating AAS model by automatically translating unstructured textual data into a standardized AAS model. The generated AAS model can be integrated into AAS-compliant digital twin software for seamless information exchange and communication. In our evaluation, a comparative analysis of different LLMs and an in-depth ablation study of Retrieval-Augmented Generation (RAG) mechanisms provide insights into the effectiveness of LLM systems for interpreting technical concepts and translating data. Our findings emphasize LLMs' capability to automate AAS instance creation and contribute to the broader field of semantic interoperability for digital twins in industrial applications. The prototype implementation and evaluation results are presented on our GitHub Repository: https://github.com/YuchenXia/AASbyLLM.}, author = {Xia, Yuchen and Xiao, Zhewen and Jazdi, Nasser and Weyrich, Michael}, doi = {10.1109/ACCESS.2024.3415470}, editor = {of Electrical, Institute and Inc., Electronics Engineers}, issn = {21693536}, journal = {IEEE Access}, pages = {84863-84877}, publisher = {Institute of Electrical and Electronics Engineers Inc.}, title = {Generation of Asset Administration Shell With Large Language Model Agents: Toward Semantic Interoperability in Digital Twins in the Context of Industry 4.0}, volume = 12, year = 2024 }
6. Y. Xia, J. Zhang, N. Jazdi, and M. Weyrich, “Incorporating Large Language Models into Production Systems for Enhanced Task Automation and Flexibility,” Automation 2024, Jul. 2024.
  Zusammenfassung
  This paper introduces a novel approach to integrating large language model (LLM) agents into automated production systems, aimed at enhancing task automation and flexibility. We organize production operations within a hierarchical framework based on the automation pyramid. Atomic operation functionalities are modeled as microservices, which are executed through interface invocation within a dedicated digital twin system. This allows for a scalable and flexible foundation for orchestrating production processes. In this digital twin system, low-level, hardware-specific data is semantically enriched and made interpretable for LLMs for production planning and control tasks. Large language model agents are systematically prompted to interpret these production-specific data and knowledge. Upon receiving a user request or identifying a triggering event, the LLM agents generate a process plan. This plan is then decomposed into a series of atomic operations, executed as microservices within the real-world automation system. We implement this overall approach on an automated modular production facility at our laboratory, demonstrating how the LLMs can handle production planning and control tasks through a concrete case study. This results in an intuitive production facility with higher levels of task automation and flexibility. Finally, we reveal the several limitations in realizing the full potential of the large language models in autonomous systems and point out promising benefits. Demos of this series of ongoing research series can be accessed at: https://github.com/YuchenXia/GPT4IndustrialAutomation
  BibTeX
  @article{Xia2024, abstract = {This paper introduces a novel approach to integrating large language model (LLM) agents into automated production systems, aimed at enhancing task automation and flexibility. We organize production operations within a hierarchical framework based on the automation pyramid. Atomic operation functionalities are modeled as microservices, which are executed through interface invocation within a dedicated digital twin system. This allows for a scalable and flexible foundation for orchestrating production processes. In this digital twin system, low-level, hardware-specific data is semantically enriched and made interpretable for LLMs for production planning and control tasks. Large language model agents are systematically prompted to interpret these production-specific data and knowledge. Upon receiving a user request or identifying a triggering event, the LLM agents generate a process plan. This plan is then decomposed into a series of atomic operations, executed as microservices within the real-world automation system. We implement this overall approach on an automated modular production facility at our laboratory, demonstrating how the LLMs can handle production planning and control tasks through a concrete case study. This results in an intuitive production facility with higher levels of task automation and flexibility. Finally, we reveal the several limitations in realizing the full potential of the large language models in autonomous systems and point out promising benefits. Demos of this series of ongoing research series can be accessed at: https://github.com/YuchenXia/GPT4IndustrialAutomation}, author = {Xia, Yuchen and Zhang, Jize and Jazdi, Nasser and Weyrich, Michael}, doi = {10.51202/9783181024379}, journal = {Automation 2024}, month = {07}, publisher = {VDI Verlag}, title = {Incorporating Large Language Models into Production Systems for Enhanced Task Automation and Flexibility}, url = {http://arxiv.org/abs/2407.08550 http://dx.doi.org/10.51202/9783181024379}, year = 2024 }
  Link
  http://arxiv.org/abs/2407.08550 http://dx.doi.org/10.51202/9783181024379
7. Y. Xia, D. Dittler, N. Jazdi, H. Chen, and M. Weyrich, “LLM experiments with simulation: Large Language Model Multi-Agent System for Simulation Model Parametrization in Digital Twins,” 2024 IEEE 29th International Conference on Emerging Technologies and Factory Automation (ETFA), pp. 1–4, Sep. 2024.
  - BibTeX
  - Link
  BibTeX
  @article{Xia2024, author = {Xia, Yuchen and Dittler, Daniel and Jazdi, Nasser and Chen, Haonan and Weyrich, Michael}, doi = {10.1109/ETFA61755.2024.10710900}, isbn = {979-8-3503-6123-0}, journal = {2024 IEEE 29th International Conference on Emerging Technologies and Factory Automation (ETFA)}, month = {09}, pages = {1-4}, publisher = {IEEE}, title = {LLM experiments with simulation: Large Language Model Multi-Agent System for Simulation Model Parametrization in Digital Twins}, url = {https://ieeexplore.ieee.org/document/10710900/}, year = 2024 }
  Link
  https://ieeexplore.ieee.org/document/10710900/
2023
1. Y. Xia, M. Shenoy, N. Jazdi, and M. Weyrich, “Towards autonomous system: flexible modular production system enhanced with large language model agents,” Apr. 2023.
  - BibTeX
  BibTeX
  @article{unknown, author = {Xia, Yuchen and Shenoy, Manthan and Jazdi, Nasser and Weyrich, Michael}, month = {04}, title = {Towards autonomous system: flexible modular production system enhanced with large language model agents}, year = 2023 }

Forschung

Forschungsschwerpunkt (Research Focus):

Large Language Models for Specialized Tasks in Industrial Automation
Intelligent Autonomous Systems based on Large Language Model Agents and Digital Twins.
Semantics in Digital Twins

Beschreibung (Description):

LLMs exhibit remarkable capabilities in interpreting complex texts and utilizing collective human knowledge. These models exhibit impressive proficiency in comprehending and analyzing semantics in natural language, while also displaying advanced abilities in reasoning and problem-solving.

By combining automation systems, digital twins and large language models, my research seeks to develop more intelligent and efficient systems. On one hand, automation systems are enhanced with digital twins and LLMs to unlock the potential of data- and AI-driven smart factories. On the other hand, LLMs can interact with physical environments by having an embodiment in reality through the established infrastructure that consists of automation systems and digital twins. This approach equips an artificial “brain” with mechatronic “hands” and “eyes” for more intelligent interaction. By combining these interdisciplinary technologies, my objective is to develop innovative technology stacks that enhance the efficiency and intelligence of smart factories.

Forschungsportal (Research Portal):

Google Scholar: https://scholar.google.de/citations?user=hi1srxkAAAAJ

Research Gate: https://www.researchgate.net/profile/Yuchen-Xia-4

LinkedIn: https://www.linkedin.com/in/xiayuchen/

My academic voyage started at the esteemed Wuhan University in China (2011), where I earned my first Bachelor’s Degree in Mechanical Design and Automation (2017). During this period, I also participated in a dual-degree program that took me to the University of Stuttgart. Here, I completed the full study program in Automotive and Engine Technology, earning my second Bachelor’s Degree in Germany (2017).

For my master’s studies, I delved into the specialized fields of Mechatronics and Autonomous Driving. I earned my master’s degree from the University of Stuttgart in 2019 and felt compelled to continue my academic journey, leading me to pursue a doctoral degree.

In 2020, I proposed my doctoral research, which received financial backing from the Stiftung der Deutschen Wirtschaft, which aims to supporting entrepreneurship. This project was conducted at the Institute of Industrial Automation and Software Engineering (IAS) and under the affiliation of the Graduate School of Excellence in Advanced Manufacturing Engineering (GSaME) at the University of Stuttgart. Since then, my focus has shifted to the semantic analysis of information models. My exploratory research journey has led me to the fields of Neural Language Models and Deep Learning. My expertise lies in Large Language Models and their applications in industrial automation. Driven by an entrepreneurial spirit, I am focused on developing research products that offer both economic and social value.

Supervised Students’ Research Projects: