The Air Force Office of Scientific Research (AFOSR) has named Virginia Smith a recipient of its 2025 Young Investigator Program award, recognizing her research in federated learning and its potential impact on national security

Smith, the Leonardo Associate Professor of Machine Learning in Carnegie Mellon University's School of Computer Science, earned the honor for her proposal, "Secure and Efficient Federated Learning in the Foundation Model Era." The project focuses on developing machine learning methods suited for the communication-limited and adversarial environments commonly encountered in Air Force operations.

The Young Investigator Program supports early career scientists and engineers who show exceptional ability and promise for conducting basic research. The program aims to foster creative inquiry, promote early career development, and strengthen ties between researchers and the Department of the Air Force's mission.

Smith's project centers on federated learning, a technique for training machine learning models across distributed systems without sharing raw data. While widely used in commercial networks, such as mobile phones, existing methods often assume stable, high-bandwidth connectivity — conditions rarely present in military settings.

Her work introduces agent-aware approaches that account for the severe communication constraints and security risks in military networks. It also addresses new challenges posed by foundation models — large, pretrained systems that enhance model accuracy but raise additional efficiency and security concerns. 

Designed to identify and support the next generation of leaders in science and engineering, the Young Investigator Program is among the most competitive research initiatives the Department of the Air Force offers.

For more information about the Young Investigator Program, visit the AFOSR website.