Skip to main content
Lockheed Martin Awards Research Grant to Rickard Ewetz

Lockheed Martin Corporation has awarded a $150K research grant to UCF Department of Electrical and Computer Engineering associate professor Rickard Ewetz, Ph.D., to investigate robust intelligent systems based on explainability. The one-year project will involve two Ph.D. students and several undergraduate researchers.

Neural network models have surpassed human-level capabilities for many computer vision applications. Unfortunately, it has been shown that an adversary can manipulate the model input data to cause model misclassifications, which can result in catastrophic outcomes for safety-critical systems.

The first part of Ewetz’ project focuses on understanding the inner reasoning and decision making of artificial intelligence (AI) systems. His team will develop attribution methods that identify the contribution of each input feature to the output prediction of a model. The second part of the project is centered on utilizing these explainable AI algorithms to develop neural networks that are robust against attacks. The researchers will investigate defenses against adversarial attacks and patch attacks using human perception metrics, entropy analysis and modular neural networks.

“This project is an exciting opportunity to transfer academic contributions into practical knowledge for industry partners,” Ewetz said.

Ewetz directs the AI and Emerging Computing laboratory, which is focused of the broad areas of artificial intelligence, emerging technology, design automation and security. He has published more than 50 papers and received best paper nominations from ASP-DAC 2019 and DATE 2021. His research is supported by DARPA, NSF, DOE, Lockheed Martin Corporation and Cyber-Florida.