Text

Artificiell intelligens och intelligenta system

Formell modellering och analys av inbäddade system

Heterogena system

Lärande och optimering

Modellbaserad konstruktion av inbäddade system

Statsvetenskap

Product and Production Development

Systemdesign i realtid

Säkerhetskritisk teknik

Simulation and optimisation for future industrial applications (SOFIA)

Sustainable lifestyle and health from a public health perspective

Ubiquitous Computing

Komplexa inbyggda system i realtid

Cyber-Physical Systems Analysis

Tillförlitlig programvaruteknik

Energieffektivisering och minskning av utsläpp

Energy-Efficient Hardware Accelerator for Embedded Deep Learning

In this joint project, we aim at decreasing the power consumption and computation load of the current image processing platform by employing the concept of computation reuse.

Start

2019-01-01

Main financing

STINT - The Swedish Foundation for International Cooperation in Research and Higher Education

Project manager

Senior Lecturer

Masoud Daneshtalab

+4621103111

masoud.daneshtalab@mdh.se

In this joint project, we aim at decreasing the power consumption and computation load of the current image processing platform by employing the concept of computation reuse. Computation reuse suggests temporarily storing and reusing the result of a recent arithmetic operation for anticipated subsequent operations with the same operands. Our proposal is motivated by the high degree of redundancy that we observed in arithmetic operations of neural networks where we show that an approximate computation reuse can eliminate up to 94% of arithmetic operation of simple neural networks. This leads to up to 80% reduction in power consumption, which directly translates to a considerable increase in battery life time. We further presented a mechanism to make a large neural network by connecting basic units in two UT-MDH joint works.