Mittal wins Navy award to make machine learning more secure

Monday, Feb 26, 2018

Prateek Mittal, Assistant Professor of Electrical Engineering, has received a 2018 Young Investigator Award from the United States Office of Naval Research (ONR).
The award supports Mittal’s work in designing machine-learning systems that are more secure in adversarial settings, where attackers may exploit system failures. Previous research has demonstrated such vulnerabilities “across multiple applications, datasets and classifiers, including deep neural networks,” according to the professor.
His winning proposal, entitled “Synergistic Integration of Statistical and Logic-Based Reasoning for Adversarial Learning Mechanisms,” seeks to address those vulnerabilities as part of the ONR’s Cyber Security and Complex Software Systems program.
According to the ONR website, this program “focuses on developing the fundamental principles and models for the design and construction of complex software systems that meet required assurances for security, safety, reliability and performance.”
The 2018 awards, announced on February 21, were given to 31 scientists and totaled $16 million in funds. The number of applicants topped 340, each one assessed on merit, potential, performance and long-term university commitment.
“I am grateful to ONR for supporting my group’s research,” Mittal said. “This award will help advance the science of designing secure machine-learning systems.”