DIAL ACCELERATORS FOR DEEP NEURAL NETWORKS: DISTRIBUTED AND NETWORKED, IN-EMBEDDED-HARDWARE, ADVERSARIAL GENERATIVE, LOW-PRECISION ARITHMETIC LEARNING AND INFERENCE
Abstract
The Recipient will seek to perform research that exploits a recent opportunity that special-purpose deep neural networks (DNNs) accelerators for devices, edges, and the cloud can serve as general-purpose computing engines, given that deep learning is widely applicable to many problems. This research will develop solutions to two pressing problems: easing the development and adoption of novel accelerators for Artificial Intelligence (AI) computations and creating trustworthy AI systems guided by physical insights.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Sep 22, 2022
- Source ID
- FA87502210500
Entities
People
- H.t. Kung
Organizations
- President and Fellows of Harvard College
- Rome Laboratory
- United States Air Force