Accession Number:

AD1041633

Title:

CMOS-based Stochastically Spiking Neural Network for Optimization under Uncertainties

Descriptive Note:

Conference Paper

Corporate Author:

University of Illinois at Chicago Chicago United States

Personal Author(s):

Report Date:

2017-03-01

Pagination or Media Count:

4.0

Abstract:

We present CMOS-based stochastically spiking neural network for optimization under uncertainties. We discuss a scenario generation circuit to non-parametrically estimateemulate statistics of uncertain costconstraints variables in an optimization problem. We also present a spiking neural network for linearquadratic programming. Scenario generation block stochastically controls spiking neural network to extract optimal solution of an optimization problem minimizing its expected cost.

Subject Categories:

  • Solid State Physics

Distribution Statement:

APPROVED FOR PUBLIC RELEASE