检索结果-内蒙古大学图书馆

Large Pre-Trained Models and Few-Shot FineTuning for Virtual Metrology: A Framework for Uncertainty-Driven Adaptive Process Control in Semiconductor manufacturing

引用

IEEE Transactions on Automation Science and engineering 2025年 22卷 15351-15370页

作者： Lin, Chin-Yi Tseng, Tzu-Liang Emon, Solayman Hossain Tsai, Tsung-Han University of Texas at EL Paso Department of Industrial Manufacturing and Systems Engineering Texas El PasoTX79968 United States Computational Science Program University of Texas at EL Paso Texas El PasoTX79968 United States National Taipei University of Business Institute of Information and Decision Sciences Taipei10051 Taiwan

High-precision wafer metrology poses significant cost and throughput challenges in modern semiconductor manufacturing, where frequent process changes and recipe variations demand highly adaptive and scalable solutions. In this paper, we present a Generative-FewShot-Active Virtual Metrology (GFA-VM) framework that unifies large-scale generative modeling, few-shot fine-tuning, and uncertainty-driven active sampling into a single, data-centric system. A foundational generative model, built on a hybrid architecture of Transformer networks and Variational Autoencoders (VAEs), learns diverse sensor characteristics in an offline stage without relying on extensive labeled data. During online inference, the model produces both wafer quality predictions and predictive uncertainties;samples exceeding a dynamic uncertainty threshold are selected for physical measurement and few-shot model recalibration. This selective sampling both reduces measurement costs and adapts rapidly to new process conditions (e.g., novel recipes or equipment upgrades), requiring only a handful of freshly labeled wafers. The paper further addresses the long-term stability of the system through a self-updating mechanism that adjusts the uncertainty threshold when distributional shifts occur. Empirical evaluations confirm that our GFA-VM approach achieves state-of-the-art accuracy while significantly reducing metrology overhead compared to conventional virtual metrology methods. Additionally, rigorous theoretical analyses—including proofs of convergence and label cost bounds—demonstrate the reliability of using a generative foundation plus meta-learning technique. By fostering on-demand adaptation within a closedloop framework, GFA-VM offers a comprehensive, scalable strategy for next-generation semiconductor process control. © 2004-2012 IEEE.

关键词： Semiconductor device measurement Metrology Semiconductor device modeling Semiconductor process modeling Uncertainty Adaptation models Accuracy Data models Predictive models manufacturing

来源：评论

学校读者我要写书评

暂无评论

Generative AI-Driven Data Augmentation for Robust Virtual Metrology: GANs, VAEs, and Diffusion Models

引用

IEEE Transactions on Semiconductor manufacturing 2025年

作者： Lin, Chin-Yi Tseng, Tzu-Liang Emon, Solayman Hossain Tsai, Tsung-Han Department of Industrial Manufacturing and Systems Engineering University of Texas at El Paso El Paso 79968 TX United States Computational Science Program University of Texas at El Paso El Paso 79968 TX United States Institute of Information and Decision Sciences National Taipei University of Business Taipei 10051 Taiwan

Advanced semiconductor manufacturing increasingly depends on Virtual Metrology (VM) for real-time quality monitoring, yet conventional data-driven models rarely capture the scarce or extreme process conditions critical for robust predictions. We propose a Multi-Stage Constrained Data Generative Augmentation (MSC-DGA) framework that integrates Variational Autoencoders (VAE), Normalizing Flows, and Constrained Diffusion to systematically expand coverage of seldom-seen regimes. By embedding strict engineering constraints during generation and applying a two-stage quality filter, MSCDGA ensures physically plausible synthetic samples. We further present theoretical proofs showing that multi-stage generation can approximate complex sensor distributions while enforcing domain validity, thereby improving coverage and preserving essential process physics. Empirically, we demonstrate the approach on real WBG SiC data, incorporating these curated samples into a Generative Foundation Model (GFA-VM) with few-shot finetuning and uncertainty-based active sampling, yielding significant accuracy gains for rarely observed conditions. Experiments confirm notable performance improvements over single-stage augmentation and naive oversampling. By rigorously balancing distribution realism with engineering feasibility, MSC-DGA offers a practical and theoretically grounded advancement for VM, enhancing adaptive control and product quality in next-generation power semiconductor manufacturing. © 2024 IEEE.

关键词： Advanced Process Control (APC) Data augmentation Diffusion models Generative models Physics-informed constraints Rare event coverage Semiconductor manufacturing Variational Autoencoders (VAE) Virtual Metrology (VM)

来源：评论

学校读者我要写书评

暂无评论

FLEXIBLE AUTOMATED MATERIAL HANDLING-SYSTEM

引用

JOURNAL OF MECHANICAL WORKING TECHNOLOGY 1989年第C期20卷 433-440页

作者： MAZOUZ, AK HAN, CP Assistant Professors Manufacturing Systems Engineering Program Department of Mechanical Engineering Florida Atlantic University Boca Raton FL 33431 USA

This paper discusses the development of an expert system based on axis orientation of the parts in a three dimensional system. For any part fed to a machine cell, the material handling planning system presents the part to the assembly robot in the desired orientation. Group Technology classification and coding method is used for part identification, as well as for the feeding mechanism design.

关键词： Materials Handling

来源：评论

学校读者我要写书评

暂无评论

Enterprise information system modeling for continuous improvement

引用

COMPUTERS & INDUSTRIAL engineering 1996年第1-2期31卷 273-276页

作者： Manley, JH Director Manufacturing Systems Engineering Program Professor of Industrial Engineering 1044 Benedum Hall University of Pittsburgh Pittsburgh Pennsylvania 15261 USA

A three-phase information system analysis and design methodology is being used to continuously improve enterprise information systems as part of a six-step annual business improvement process. Following senior management's strategic decisions on next year's product and/or service portfolio content, interacting financial, management, engineering, and quality improvement processes are analyzed to determine their output product and/or service quality and timeliness. Concurrently, facilities, equipment, and personnel resources required for individual processes are inspected for possible immediate or future improvement. Throughout these analyses minimum essential information (MEI) requirements are derived using the Object Transformation Process Model (OTPM). Individual OTPM models are linked to help identify all pertinent data sources, information destinations, and timing requirements. The linked OTPM models are mapped onto an Embedded Computer System (ECS) model that defines a physical architecture for improving telecommunication paths between all humans, machines and embedded computers that are component parts of the integrated processes. This approach yields comprehensive information system logical and physical architectural models that can recursively guide high-leverage enterprise-wide improvement projects over succeeding fiscal years.

关键词： information systems process reengineering continuous improvement enterprise modeling

来源：评论

学校读者我要写书评

暂无评论

manufacturing information systems reengineering

Manufacturing information systems reengineering

引用

Portland International Conference on Management of engineering and Technology (PICMET 97)

作者： Manley, JH Manufacturing Systems Engineering Program University of Pittsburgh 1044 Benedum Hall Pittsburgh 15261 PA United States

ISBN: (纸本)0780335740

manufacturing process planning, analysis and control information systems must be periodically reengineered as an integral part of any enterprise-wide continuous improvement strategy. An annual business planning approach and associated object transformation process macro-modeling methodology is recommended to satisfy this manufacturing enterprise improvement requirement.

关键词： Information systems

来源：评论

学校读者我要写书评

暂无评论

INFORMATION-PROCESS FLOW-ANALYSIS (IPFA) FOR REengineering manufacturing systems

引用

COMPUTERS & INDUSTRIAL engineering 1993年第1-4期25卷 275-278页

作者： MANLEY, JH Director Manufacturing Systems Engineering Program and Professor of Industrial Engineering 1044 Benedum Hall University of Pittsburgh Pittsburgh Pennsylvania USA (412) 624-9846

Automated manufacturing lines, aircraft, nuclear power plants, and similar real-time intelligent or embedded systems contain three control elements: (1) embedded computers, (2) humans-in-the-loop, and (3) certain machines, i.e., mechanically, electrically, or electronically operated devices. Embedded system control elements are interconnected through accurate, precisely-timed, closed-loop information systems to provide high quality products and/or services from their parent systems, and in some cases to ensure human safety. Highly-reliable controlling computers and associated software embedded in time- and safety-critical systems that compensate for possible human and/or machine error are especially difficult to design and develop, or ''forward engineer.'' Reengineering embedded control elements in old systems is even more troublesome, especially when they require re-manufacturing, e.g., nuclear power plant refueling, aircraft avionics system upgrading, or manufacturing plant retooling. This paper describes how industrial engineers can assist in reengineering worn out, error prone, or obsolescent real-time manufacturing systems (embedded systems) by helping computer systems and communication engineers ensure that critical information control loops, both feed forward and feed back, are complete and efficient. Two conceptual models, the Embedded Computer System (ECS) physical model and the Object Transformation Process Model (OTPM) are used to guide a modified process flow analysis (PFA) of existing large-scale, complex embedded systems that takes into account process-supporting information. This modified PFA is called an Information Process Flow Analysis (IPFA).

关键词：

来源：评论

学校读者我要写书评

暂无评论

Correction: Materials laboratories of the future for alloys, amorphous, and composite materials (MRS Bulletin, (2025), 50, 2, (190-207), 10.1557/s43577-024-00846-y)

引用

MRS Bulletin 2025年 1-2页

作者： Banerjee, Sarbajit Meng, Y. Shirley Minor, Andrew M. Zhang, Minghao Zaluzec, Nestor J. Chan, Maria K.Y. Seidler, Gerald McComb, David W. Agar, Joshua Mukherjee, Partha P. Melot, Brent Chapman, Karena Guiton, Beth S. Klie, Robert F. McCue, Ian D. Voyles, Paul M. Robertson, Ian Li, Ling Chi, Miaofang Destino, Joel F. Devaraj, Arun Marquis, Emmanuelle A. Segre, Carlo U. Liu, Huinan H. Yang, Judith C. Momeni, Kasra Misra, Amit Abdolrahim, Niaz Medvedeva, Julia E. Cai, Wenjun Sehirlioglu, Alp Dizbay-Onat, Melike Mehta, Apurva Graham-Brady, Lori Maruyama, Benji Rajan, Krishna Warner, Jamie H. Taheri, Mitra L. Kalinin, Sergei V. Reeja-Jayan, B. Schwarz, Udo D. Simon, Sindee L. Brown, Craig M. Department of Chemistry and Department of Materials Science & Engineering Texas A & M University College Station United States Pritzker School of Molecular Engineering The University of Chicago Chicago United States Department of Materials Science & Engineering University of California Berkeley Berkeley United States National Center for Electron Microscopy Molecular Foundry Lawrence Berkeley National Laboratory Berkeley United States Center for Nanoscale Materials Argonne National Laboratory Lemont United States Department of Physics University of Washington Seattle United States Department of Materials Science & Engineering The Ohio State University Columbus United States Department of Mechanical Engineering and Mechanics Drexel University Philadelphia United States School of Mechanical Engineering Purdue University West Lafayette United States Department of Chemistry University of Southern California Los Angeles United States Department of Chemistry Stony Brook University The State University of New York Stony Brook United States Department of Chemistry University of Kentucky Lexington United States Department of Physics University of Illinois at Chicago Chicago United States Department of Materials Science & Engineering Northwestern University Evanston Bangladesh Department of Materials Science & Engineering University of Wisconsin–Madison Madison United States Department of Materials Science and Engineering University of Pennsylvania Philadelphia United States Department of Mechanical Engineering and Materials Science Duke University Durham United States Department of Chemistry Creighton University Omaha United States Physical and Computational Sciences Directorate Pacific Northwest National Laboratory Richland United States Department of Materials Science & Engineering University of Michigan–Ann Arbor Ann Arbor United States Department of Physics Illinois Institute of Technology Chicago United States Department of Bioengineering and the Mater

This article was updated to correct Minghao Zhang’s affiliation from "Pritzker School of Molecular engineering, The University of Chicago, Chicago, USA" to "Pritzker School of Molecular engineering, The University of Chicago, Chicago, USA;Argonne National Laboratory, Lemont, USA." Argonne National Laboratory was left off the original publication. © The Author(s) 2025.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Cooperation in aircraft design

引用

Research in engineering Design 1992年第2期4卷 115-130页

作者： Bond, Alan H. Ricci, Richard J. Manufacturing Engineering Program University of California Los Angeles California United States Automation Systems Lockheed Aeronautical Systems Company Burbank California United States

We describe how aircraft are designed in a large organization. We discuss the different phases of design and interaction with the customer. We then describe the models used by each specialist department and the interactions among departments during the design process. We observe that the main design choices are refinement operations on the design. We then briefly describe how the negotiation process is controlled by an organizationally agreed sequence of commitment steps. We then describe negotiation at higher levels in the organization. What decisions are made, the compromises worked out, and the effect of these higher-level commitments on the design process. We conclude that: (1) aircraft design proceeds by the cooperation of specialists (specialist teams or departments);(2) each specialist has its own model of the design, and may use several different models or partial models for different purposes;(3) specialists have limited ability to understand each other's models. They communicate using a shared vocabulary, but not necessarily shared technical knowledge;(4) design proceeds by successive refinement of the models, which are coordinated and updated together;(5) the design decisions, which are acts of commitment and model refinement, are negotiated by the specialists among themselves;(6) one way this negotiation process is organized and controlled is by the use of commitment steps;(7) negotiations occur at higher levels in the organization, resulting in commitments which greatly influence and constrain the design process and its organization, and which have the greatest effect on the cost of the product. © 1992 Springer-Verlag New York Inc.

关键词：

来源：评论

学校读者我要写书评

暂无评论

USE OF NEURAL NETWORKS TO IDENTIFY AND COMPENSATE FOR FRICTION IN PRECISION, POSITION CONTROLLED MECHANISMS

USE OF NEURAL NETWORKS TO IDENTIFY AND COMPENSATE FOR FRICTI...

引用

27TH ANNUAL MEETING OF THE INDUSTRY APPLICATIONS SOC OF IEEE

作者： SEIDL, DR REINEKING, TL LORENZ, RD Department of Electrical and Computer Engineering Manufacturing Systems Engineering Program University of Wisconsin - Madison 1415 Johnson Drive Madison 53706 WI United States

ISBN: (纸本)0780306341

This paper demonstrates that artificial neural networks can identify and compensate for friction in precision, position controlled mechanisms improving performance substantially. A major contribution of this research is that knowledge of the friction's form is used to determine the neural network's structure. This unique approach solves network sizing and weight initializing problems. The friction model is used for feed-forward decoupling of friction-induced torque. The neural network also explicitly incorporates inertia compensation and linear feedback control. It can thus be readily initialized to perform identically to an existing controller. Adaptation to enhance performance can then proceed from that point. This solves the problem of providing good control in the early stages of adaptating the network's weights. Another contribution of this research is a demonstration of the trajectory dependence of static friction compensation with a discrete time controller. A hypothesis to explain this trajectory-dependent behavior is proposed. The paper includes both theoretical formulation and practical implementation results applied to control of a commercial d.c. motor having a significant amount of static friction. © IEEE.

关键词： Neural networks

来源：评论

学校读者我要写书评

暂无评论

An examination of ME449 redesign and prototype fabrication: A new senior/grad design and fabrication course at the university of Wisconsin-madison 122

An examination of ME449 redesign and prototype fabrication: ...

引用

2015 122nd ASEE Annual Conference and Exposition

作者： Manner, Kim J. Pfefferkorm, Frank E. University of Wisconsin Department of Mechanical Engineering Madison United States University of Wisconsin Manufacturing Systems Engineering Program United States

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：