Published January 1, 2018 | Version v1
Conference paper Open

APPROXIMATE FULLY CONNECTED NEURAL NETWORK GENERATION

  • 1. Istanbul Tech Univ, Fac Elect & Elect Engn, Elect & Commun Engn Dept, TR-34469 Istanbul, Turkey

Description

Approximate computing is exploited in implementation of fully connected networks for classification problems. A multiplier structure whose area is scalable over accuracy through approximate computing is proposed. In order to employ the multipliers in a network, an area reduction algorithm is formed. It can adjust the approximation level of multipliers while still maintaining the target classification performance, without prior information on the value of network weights. Implementing on a Spartan6 FPGA, up to 79% area saving is recorded for various performance targets.

Files

bib-e951d009-08ba-4a0c-a90c-897ab874919b.txt

Files (217 Bytes)

Name Size Download all
md5:e59313c7869dc900fc9716706483f47a
217 Bytes Preview Download