Advances in Neuromorphic Computing to Support USAF Mission Effectiveness
The Air Force Research Laboratory (AFRL) is advancing high-performance embedded computing (HPEC) by pioneering energy-efficient information exploitation techniques using neuromorphic computing. Leveraging neuroscience principles, our research, which includes applications of Intel's Loihi 2 neuromorphic processor, addresses critical processing challenges for the U.S. Air Force and Space Force. We are developing low-power methods for autonomous information exploitation at the edge, aiming to compress timelines for find, fix, track, and assess. These technologies are being integrated into HPEC systems like AFRL/SRC's Agile Condor (US Patent 10,915,152), enabling enhanced multi-function processing capabilities for Air Force assets. We will share our progress and discuss potential collaboration opportunities in this research area.
IMPACT
Increased model performance with Neuromorphic Hardware: 4x throughput & +3% Average Precision - Provides efficient compute options for ground, air, high altitude and space missions.
PRESENTER
Barnell, Mark
mark.barnell.1@us.af.mil
315-330-3273
AFRL/RI
CO-AUTHOR(S)
Isereau, DJ
isereau@srcinc.com
Vidal, Francesca
fvidal@srcinc.com
Loomis, Lisa
lisa.loomis.3@us.af.mil
Wise, David
david.wise.13@us.af.mil
CATEGORY
Other
SECONDARY CATEGORY
AI/ML for HPC
SYSTEM(S) USED
n/a