Ali Shakouri, Professor of Electrical and Computer Engineering

Purdue University’s SMART Consortium recently received a $3M National Science Foundation (NSF) Future Manufacturing grant as well as an NSF Engineering Research Center (ERC) Planning grant focusing on Privacy preserving artificial intelligence for agile and resilient manufacturing. SMART (Scalable Manufacturing of Aware & Responsive Thin Films) is a multidisciplinary research center located in the Birck Nanotechnology Center. Led by Ali Shakouri, professor in the Elmore Family School of Electrical and Computer Engineering, the SMART Consortium brings together more than 40 faculty, researchers, and students from a dozen departments in engineering, science and agriculture who work on vertically integrated research ranging from novel materials and devices to manufacturing scale up, device optimization and data analytics. Leveraging industry support, low-cost Internet of Thing (IoT) devices are demonstrated and field-tested for applications in digital agriculture, next generation manufacturing, smart infrastructure, and healthcare.

Pioneering research by Jan Allebach, Hewlett Packard Distinguished Professor of ECE, and Ashraf Alam, Jai N. Gupta Professor of ECE, both at Purdue, has enabled the team to benefit from advances in artificial intelligence and machine learning to create a new paradigm for scale up manufacturing of low-cost devices. Some of the key concepts include certified as built and reliable sensing using unreliable sensors.

Testing the IoT devices in a regional testbed of local manufacturers and farmers, has helped the team to see another opportunity where the latest AI techniques could make US industry more competitive.  The new NSF-sponsored projects are in collaboration with leading groups at Harvard University, Tuskegee University, University of Michigan, UMass Amherst, and Ivy Tech Community College focusing on: Privacy preserving artificial intelligence for agile and resilient manufacturing.

“Artificial Intelligence (AI) has revolutionized many industries and can be an important tool to improve the productivity and agility of manufacturing, but progress has been slow,” says Shakouri. “This project proposes a fundamental reimagination of distributed AI techniques by establishing an AI-Commons that bridges multiple sites and companies using secure, distributed machine learning (ML).”

Shakouri says this addresses a critical shortcoming of the current approach to AI in manufacturing: the limitation of machine learning training data to in-company data.  Since AI algorithms increase in power with more data, secure data sharing and aggregation has the potential to provide vastly better AI solutions for all manufacturers. 

The project team has developed a strong partnership with dozens of small and large manufacturers in ten counties in Indiana through the Purdue-WHIN (Wabash Heartland Innovation Network) project funded by the Lilly Endowment Inc. (LEI).  The LEI project has created a unique regional testbed where companies work closely with Purdue and share their data and know-how. This is a critical element of the NSF Future Manufacturing project; and it will be hopefully the foundation of a future national center of excellence through the NSF Engineering Research Center planning grant.

Engineering Research Centers are the premier large, multi-institutional funded projects for the National Science Foundation. They combine fundamental and applied research efforts, along with substantial industry and community stakeholder involvement and workforce development, to address grand societal challenges.