You are here

Industrial Artificial Intelligence and Smart Manufacturing

Joseph Cohen, University of Michigan
Xun Huan, University of Michigan
 
Data collection is ubiquitous in modern manufacturing production systems. Due to recent advancements motivated by the Fourth Industrial Revolution, data-driven methods have been successfully implemented for quality control, maintenance scheduling, and operations optimization to improve the efficiency and sustainability of complex, high-dimensional manufacturing systems. However, significant challenges remain, particularly for high-throughput multistage processes that lack intermediate observability of key performance indicators. Labeled data are difficult and expensive to acquire, and lack of explainability impedes the adoption and maturity of existing computational methods. New algorithms designed to handle industrial constraints are needed to face the global supply chain crises in the wake of the COVID-19 pandemic. The goal of this minisymposium is to highlight a variety of novel computational approaches in topics including but not limited to: digital twin development, smart additive manufacturing, advanced metrology, human-centered augmented intelligence, quality management, discrete event systems, uncertainty quantification, and prognostics and health management. Relevant techniques include machine learning, deep learning, reinforcement learning, explainable artificial intelligence, dimensionality reduction, stochastic optimization, semi-supervised and unsupervised learning, and Bayesian methodology. We invite contributions focused on computation-oriented methodology with an emphasis on large-scale industrial applications.