Faculty Directory

Dr. Jay Lee is the Clark Distinguished Chair Professor and Director of the Industrial AI Center in the Department of Mechanical Engineering at the University of Maryland, College Park. An internationally recognized authority in Industrial AI and smart manufacturing, his research focuses on non-traditional machine learning methodologies—including transfer learning, domain adaptation, similarity-based learning, Stream-of-X AI, and Industrial Large Knowledge Models (ILKM), etc. In addition, he emphasizes “Industrial Physical AI” over purely digital AI or "physical AI, focusing on how AI can interact with and improve physical assets (machines, turbines, complex engineering systems, etc) to achieve resilient and sustainable performance.  He also leads the Data Foundry, a unique platform comprising more than 100 diverse industrial datasets spanning semiconductor manufacturing, jet engines, wind turbines, electric vehicles, high-speed rail, robotics, machine tools, and medical applications. This infrastructure supports both rapid validation of industrial AI technologies and the development of a new generation of data-centric engineering talent. Complementing this effort, he is spearheading the AI Factory initiative, an open laboratory environment designed to cultivate industry-ready Industrial AI engineers.

Previously, Dr. Lee served as Ohio Eminent Scholar, L.W. Scott Alter Chair, and University Distinguished Professor at the University of Cincinnati. He was also Founding Director of the NSF Industry/University Cooperative Research Center on Intelligent Maintenance Systems (IMS) from 2001 to 2019, which grew into a global consortium of more than 130 corporate partners and was recognized in the NSF Economic Impact Study as the most economically impactful I/UCRC. His mentorship has led to multiple technology startups, including Predictronics, launched through the NSF I-Corps program.

From 2019 to 2021, he took academic leave to serve as Vice Chairman and Board Member of Foxconn Technology Group, where he led Foxconn Industrial Internet and advanced Industrial AI–enabled manufacturing. Under his leadership, Foxconn’s smart "Lights-out" factory initiatives were recognized by the World Economic Forum as part of its Lighthouse Network in 2019, and he further advised business units that collectively earned six Lighthouse designations.  In 2007, Jay Lee donated a bearing run-to-failure dataset to the NASA Prognostics Center of Excellence. The dataset has since become a foundational benchmark for research in prognostics and health management (PHM), widely used (over 100,000 downloads) for developing and validating predictive maintenance algorithms. Earlier in his career, he held leadership roles in both government and industry, including Director of Product Development and Manufacturing at United Technologies Research Center and Program Director at the U.S. National Science Foundation, overseeing major national research initiatives such as ERCs and I/UCRCs. He also contributed to early machine learning history as part of the team at the U.S. Postal Service that developed the automated mail processing with handwritten ZIP code dataset used in the backpropagation neural networks research community.

Dr. Lee currently serves on World Economic Forum’s Global Future Council in Advanced Manufacturing and Value Chains, the Board of Governors of the Manufacturing Executive Leadership Council of the National Association of Manufacturers, and as Senior Advisor to McKinsey. He also serve as Editor-in-Chief of the IOP Journal on Machine Learning: Engineering and Chair of ASTM F50 Standards Committee on AI in Manufacturing. He is a Fellow of ASME, SME, ISEAM, and the PHM Society, and has received numerous honors, including SME’s 30 Visionaries in Smart Manufacturing (2016), 20 Most Influential Professors in Smart Manufacturing (2020), the SME Eli Whitney Productivity Award, and the SME/NAMRC S.M. Wu Research Implementation Award (2022). His book on Industrial AI was published by Springer in 2020, and he contributed to the NSF ERVA AI Engineering report in 2024.

Dr. Lee received his Doctor of Science in Mechanical Engineering from George Washington University, following doctoral studies at Columbia University. He holds master’s degrees in Industrial Management from Stony Brook University and in Mechanical Engineering from the University of Wisconsin–Madison, and a bachelor’s degree in Mechanical Engineering from Tamkang University, Taiwan.

Prof. Lee's Research Journey from Sensor Encoding to Industrial AI Foundation Model:

Professor Jay Lee’s early work in the 1990s pioneered the use of the Cerebellar Model Articulation Controller (CMAC), originally developed by James Albus, for real-time machine fault detection and predictive maintenance. He applied this neural network to transition manufacturing from reactive maintenance to quantitative, proactive degradation assessment. Prof. Jay Lee’s CMAC-based prognostics research demonstrated how machine sensor data could be discretized, encoded, and mapped into learnable representations for degradation assessment and predictive maintenance. The pioneered work established an early computational framework for transforming continuous physical sensor streams into structured machine-understandable representations, a concept that parallels modern tokenization and embedding processes.

For publication citations, highlights, and recent news:

• Google Scholar https://scholar.google.com/citations?user=g9GtqgQAAAAJ&hl=en&oi=ao 
• ResearchID:  https://researchid.co/jay.lee   
• Scopus ID: 7601455731  https://www.scopus.com/authid/detail.uri?authorId=7601455731 
• ResearchGate https://www.researchgate.net/profile/Jay_Lee10
• ScholarGPS: https://scholargps.com/scholars/25430747289210/jay-lee 
• Research.com: https://research.com/u/jay-lee 
• World Best Mechanical and Aerospace Professors (national and world  ranking) https://research.com/scientists-rankings/mechanical-and-aerospace-engineering)
• Top 2% of World Scientists by Stanford Univ. 2020-2026  (ranked 44 out of 131,085 in the field of Industrial Engineering and Automation https://topscinet.com/scientist_profile/Lee,%20Jay%20K./1988/?stype=sing... )
• Editor-in-Chief, Machine Learning: Engineering,  IOP Science, https://iopscience.iop.org/journal/3049-4761
• Chair, AI in Manufacturing Committee (F50), ASTM, https://www.digitalengineering247.com/article/astm-forms-ai-committee-fo...
• WEF Global Future Council in Advanced Manufacturing and Value Chains.  https://initiatives.weforum.org/global-future-council-on-advanced-manufa...
• Board of Governors, Manufacturing Leadership Council, https://manufacturingleadershipcouncil.com/about-mlc/board-of-governors/
• Advisory Board, Association of Industrial AI https://www.industrial-ai-network.com/about 
• Board Member, Monte Jade West Science and Technolog Association  https://www.montejade.org/2026-board 
• WEF Davos 2023 Featured video on AI for Smart Manufacturing. https://app.frame.io/presentations/c6b734dd-9db2-4900-bab2-903ace86979f
• WEF interview article https://theinnovator.news/interview-of-the-week-jay-lee-industrial-ai-expert/
• AI Engineering Report, NSF ERVA Report, 2024. https://www.ervacommunity.org/visioning-report/report-ai-engineering/ 
• Industrial AI Executive Program, EPFL Switzerland, https://www.linkedin.com/feed/update/urn:li:activity:7130849048877522947...

Industrial AI (Course Number ENME 485/691) 

Term: Spring/2026

Professor: Jay Lee

Pronouns: He/His/Him

Office Phone: 301-405-5205

Email: leejay@umd.edu 

Office Hours: Monday 3:30 pm- 4:30 pm, Jeong H. Kim Engineering Building 3226

Teaching Assistant: Hanqi Su, Meredith Wang, 

Pronouns: He/His/Him, She/Her/Her

Email: hanqisu@umd.edu & rwang135@umd.edu 

Office Hours: Monday 3:30 pm- 4:30 pm, Jeong H. Kim Engineering Building, 1210

Credits: 3

Course Dates: January 26, 2026 – May 22, 2026  

Course Times: Monday 12:30 pm - 3:15 pm

Classroom: CHM 1228

Course Description



With the proliferation of connected devices, industrial systems generate massive volumes of data at every level—from individual machines to 4F (factory, facility, field, and fleet) enterprises. Yet much of this data remains underutilized. This course introduces students to the principles and practices of Industrial Artificial Intelligence (Industrial AI), with a focus on data-centric engineering and applied machine learning for prognostics and health management (PHM) of complex engineering systems. 



Students will learn to harness industrial big data to enhance system reliability, monitor machine conditions, assess performance, and predict and prevent potential failures. Through case studies, hands-on coding, and real-world datasets, students will gain practical experience in developing machine learning from a system engineering perspective. 



Learning Outcomes



By the end of the course, students will be able to:



•    Understand core concepts of Industrial AI and its role in digital transformation and Industry 4.0.



•    Evaluate and manage data quality for industrial machine learning applications.



•    Apply machine learning techniques for machine condition monitoring, diagnostics, and prognostics.



•    Develop, validate, and deploy machine learning and AI models using real-world industrial datasets.



•    Collaborate on team projects using integrated knowledge of domain, data, and algorithms.



•    (Graduate Students only) Perform advanced data-model analysis using PHM Data Challenge Competition datasets (https://data.phmsociety.org/). 



Required Resources 



●    Course Website: elms.umd.edu There will be no formal textbook for this class, but all lecture notes will be posted online. The reference below is strongly recommended to use on a regular basis throughout the semester.



●    Text Book: 



1. Jay Lee, Industrial AI: Applications with Sustainable Performance, Springer, 2020.



●    Reference Books:



1. Bishop, Christopher M. Pattern recognition and machine learning. springer, 2006.



2. Gelman, Andrew, et al. Bayesian data analysis. Vol. 2. London: Chapman & Hall/CRC, 2014.



3. Randall, Robert Bond. Vibration-based condition monitoring: industrial, aerospace and automotive applications. John Wiley & Sons, 2011.

●    Software: Python or MATLAB



Prerequisites



•    Proficiency in Python or MATLAB



•    Or permission of the instructor



Course Structure



This is a 15-week in-person course. Attendance and participation are expected. Students are responsible for any content missed during absences.

Major Assignments

Homework Assignments

• HW1    Literature review on Industrial AI in Industry 4.0 and digital transformation. Students write a short report and give a 5-minute class presentation.
• HW2    Vibration-based fault detection using FFT and logistic regression. Group work with provided sample code. 10-minute presentation required.
• HW3    SVM-based fault classification on bearing data. Group analysis with in-class presentati
• HW4    Use of Self-Organizing Maps (SOM and SOM-MQE) for bearing fault detection. Sample code provided. Group presentation required.(Advanced Data-Model Analysis (ADDA) (Graduate students only)

Graduate students are required to complete 2 of 7 advanced analyses based on PHM Data Challenge competition datasets. Each analysis includes:



•    Paper Review (1 page): Summary of challenge task, dataset, and winning methods.



•    Exploratory Data Analysis (1–2 pages): Visualization, trends, data quality, and preprocessing.



•    Model Development (2–3 pages): At least two models, performance comparison, and critical reflection.



Possible datasets include:



•    2018 PHM NA – Ion Mill Etching System



•    2019 PHM NA – Fatigue Crack



•    2020 PHM EU – Filtration System



•    2021 PHM EU – Production Line



•    2021 PHM NA – Turbofan Engine



•    2022 PHM EU – PCBs



•    2022 PHM NA – Spacecraft Propulsion

A suggested reference paper: “Machine learning approaches for diagnostics and prognostics of industrial systems using open-source data from PHM data challenges: a review.”



Final Group Projects



Each group will select one of the following real-world datasets:



1.    Semiconductor Etching Fault Detection



2.    Spacecraft Propulsion System Monitoring



3.    Aero-engine Remaining Useful Life Prediction



4.    Gearbox Fault Diagnosis and Prognosis



5.    Bearing Fault Diagnosis and Prognostics

•    Each group will work independently with TA support.



•    A 15-minute presentation and a detailed project report are due during finals week.



•    Promising projects may be encouraged to submit for conference or journal publication.

Grading Structure

Assignment    ENME485 Percentage %    ENME691 Percentage %



Homework & Presentation    20%    15%



Midterm Test    40%    35%



Final Term-Project Report    40%    30%



Advanced Data-Model Analysis (ADDA)     0%    20%



Total    100%    100%

Grades



Final Letter Grades will be determined based on the table below. See the grading structure section for the breakdown of all assessments. 

All assessment scores will be posted on the course ELMS page.  If you would like to review any of your grades (including the exams), or have questions about how something was scored, please email me to schedule a time for us to meet and discuss.

Late work will not be accepted for course credit so please plan to have it submitted well before the scheduled deadline.  I am happy to discuss any of your grades with you, and if I have made a mistake, I will immediately correct it.  Any formal grade disputes must be submitted in writing and within one week of receiving the grade.



Final letter grades are assigned based on the percentage of total assessment points earned.  To be fair to everyone, I have to establish clear standards and apply them consistently, so please understand that being close to a cutoff is not the same as making the cut (89.99 ≠ 90.00).  It would be unethical to make exceptions for some and not others.

Final Grade Cutoffs



+    97.00%    +    87.00%    +    77.00%    +    67.00%    +    



A    94.00%    B    84.00%    C    74.00%    D    64.00%    F    <60.0%



-    90.00%    -    80.00%    -    70.00%    -    60.00%    -    

• Grading Structure

Course Outline