DARPA: Hyper-Dimensional Data Enabled Neural Networks (HyDDENN
| קרן | US Department of Defense |
|---|---|
| מדינה | U.S.A. |
| סוג | Research Grants |
| תאריך אחרון | 11/12/2019 |
| פקולטה | Engineering, Exact Sciences |
| תיאור |
The HyDDENN program seeks new data enabled neural network (NN) architectures to break the reliance on large MAC-based DNNs. The goal of HyDDENN is to significantly reduce AI hardware complexity by reducing parameter count by at least 10x, while maintaining SOA accuracy in comparison with a similar MAC-based DNN solution.
Phase 1: Feasibility Study (base) Phase 2: Proof of Concept (option)
Funding: Phase 1- $300,000 Phase 2- $700,000
Duration: 18 months (Phase 1: 6 months; Phase 2: 12 months)
Researh Authority due date: 4.12.19
|
| קבצים מצורפים | |
| קישור | לאתר |
| עדכון אחרון | עדכון אחרון: 18/11/2019 |
