January 24
@
10:00 AM
–
11:00 AM
The recent developments in artificial intelligence (AI) and machine learning (ML), together with advancements in sensor technologies and computing power, have ushered in an era of data-driven solutions across diverse advanced manufacturing applications. This talk presents some new AI/ML methodologies for quality and productivity improvement in advanced manufacturing.
The first half of the talk focuses on the developed robust learning methods for label-efficient monitoring of high-dimensional data. Conventional robust learning methods adhering to low-rank or smooth assumptions, often fall short when dealing with complex signals, like defect detection in products with intricate surface patterns. I will introduce the developed Robust GAN-inversion (RGI) method that generalizes the robust learning method for unsupervised anomaly detection in complex signals in a distributional-assumption-free manner, and discuss its applications in advanced manufacturing systems.
The second half of the talk extends the scope of the research to AI/ML enabled control and design optimization of advanced manufacturing systems, through the incorporation of domain-specific engineering knowledge, and discusses its application in dimensional variability reduction in the composite fuselage assembly process. A summary of the current challenges and future research will also be discussed.
Bio: Shancong Mou is a final year PhD student at H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology, advised by Prof. Jianjun Shi. His research interest is in the area of System Informatics and Control, with an emphasis on AI/ML-enabled data fusion for quality and productivity improvement in advanced manufacturing systems, bridging high-dimensional statistics, optimization research, machine learning, and computational science. Research outcomes have been published in leading journals, including ASME Transactions, IEEE Transactions, IISE Transactions, INFORMS J. Data Science, Technometrics; and top machine learning conference.