Resilient AI
Architectures and learning methods that adapt under pressure, reason with missing evidence, and keep working when conditions shift.
Resilient AI and Grounded Sensing Lab
AI for chaos.
The RAGS Lab builds artificial intelligence that actually works in moments of crisis: chaotic, unpredictable, and unforgiving environments where information is incomplete and lives depend on systems that cannot afford to fail.
Problems we work on
Grounded systems for unforgiving conditions.
Our research brings together machine learning, sensing, policy, and end-user input to build AI that sees through noise, reasons under uncertainty, and endures outside ideal conditions.
Grounded sensing
AI grounded in the physics and limits of sensors, including radar, low-quality audio, imagery, and other difficult signals.
Crisis response
Systems for unforgiving settings where information is incomplete, time is scarce, and people need reliable support.
End-user alignment
Research shaped by responders, operators, and policy constraints so prototypes fit real workflows and real risk.
Featured projects
From sensor physics to field use.
Current work turns imperfect field signals into AI systems that can reason with uncertainty, operational context, and real constraints.
Ongoing
Perception for Active Imaging
We are creating architectures and learning methods for active sensing modalities like radar and sonar which contemporary approaches fail at.
Ongoing
Multimodal Inverse Problems
The world is sensed through multiple means; how do we transfer representations across different types of sensors by viewing the world as an inverse problem?
Partner with the lab
Bring hard problems from the field.
We collaborate with first responders, emergency managers, public safety teams, humanitarian operators, students, and researchers who want technology to work in the environments where it will be used.