Publication Details

Approximation of Hardware Accelerators driven by Machine-Learning Models

MRÁZEK, V. Approximation of Hardware Accelerators driven by Machine-Learning Models. In Proceedings of International Symposium on Design and Diagnostics of Electronic Circuits and Systems (DDECS '23). Tallinn: Institute of Electrical and Electronics Engineers, 2023. p. 91-92. ISBN: 979-8-3503-3277-3.

Czech title

Aproximace hardwarových akcelerátorů řízená modely strojového učení (Embedded Tutorial)

Type

conference paper

Language

English

Authors

Mrázek Vojtěch, Ing., Ph.D. (DCSY)

Keywords

approximate computing, machine learning, hardware accelerators

Abstract

The goal of this tutorial is to introduce functional hardware approximation techniques employing machine learning methods. Functional approximation changes the function of a circuit slightly in order to reduce its power consumption. Machine learning models can help to estimate the error and the resulting circuit power consumption. The use of these techniques will be presented at multiple levels - at the individual component level and the higher level of HW accelerator synthesis.

Published

2023

Pages

91–92

Proceedings

Proceedings of International Symposium on Design and Diagnostics of Electronic Circuits and Systems (DDECS '23)

ISBN

979-8-3503-3277-3

Publisher

Institute of Electrical and Electronics Engineers

Place

Tallinn

DOI

10.1109/DDECS57882.2023.10139484

UT WoS

001012062000018

EID Scopus

2-s2.0-85162246870

BibTeX

@inproceedings{BUT183763,
  author="Vojtěch {Mrázek}",
  title="Approximation of Hardware Accelerators driven by Machine-Learning Models",
  booktitle="Proceedings of International Symposium on Design and Diagnostics of Electronic Circuits and Systems (DDECS '23)",
  year="2023",
  pages="91--92",
  publisher="Institute of Electrical and Electronics Engineers",
  address="Tallinn",
  doi="10.1109/DDECS57882.2023.10139484",
  isbn="979-8-3503-3277-3"
}