Low-Energy architectures of linear classifiers for IoT applications using incremental precision and multi-Level classification

Sandhya Koteshwara, Keshab K Parhi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

This paper presents a novel incremental-precision classification approach that leads to a reduction in energy consumption of linear classifiers for IoT applications. Features are first input to a low-precision classifier. If the classifier successfully classifies the sample, then the process terminates. Otherwise, the classification performance is incrementally improved by using a classifier of higher precision. This process is repeated until the classification is complete. The argument is that many samples can be classified using the low-precision classifier, leading to a reduction in energy. To achieve incremental-precision, a novel data-path decomposition is proposed to design of fixed-width adders and multipliers. These components improve the precision without recalculating the outputs, thus reducing energy. Using a linear classification example, it is shown that the proposed incremental-precision based multi-level classifier approach can reduce energy by about 41% while achieving comparable accuracies as that of a full-precision system.

Original languageEnglish (US)
Title of host publicationGLSVLSI 2018 - Proceedings of the 2018 Great Lakes Symposium on VLSI
PublisherAssociation for Computing Machinery
Pages291-296
Number of pages6
ISBN (Electronic)9781450357241
DOIs
StatePublished - May 30 2018
Event28th Great Lakes Symposium on VLSI, GLSVLSI 2018 - Chicago, United States
Duration: May 23 2018May 25 2018

Publication series

NameProceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI

Other

Other28th Great Lakes Symposium on VLSI, GLSVLSI 2018
CountryUnited States
CityChicago
Period5/23/185/25/18

Fingerprint

Classifiers
Adders
Internet of things
Energy utilization
Decomposition

Cite this

Koteshwara, S., & Parhi, K. K. (2018). Low-Energy architectures of linear classifiers for IoT applications using incremental precision and multi-Level classification. In GLSVLSI 2018 - Proceedings of the 2018 Great Lakes Symposium on VLSI (pp. 291-296). (Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI). Association for Computing Machinery. https://doi.org/10.1145/3194554.3194603

Low-Energy architectures of linear classifiers for IoT applications using incremental precision and multi-Level classification. / Koteshwara, Sandhya; Parhi, Keshab K.

GLSVLSI 2018 - Proceedings of the 2018 Great Lakes Symposium on VLSI. Association for Computing Machinery, 2018. p. 291-296 (Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Koteshwara, S & Parhi, KK 2018, Low-Energy architectures of linear classifiers for IoT applications using incremental precision and multi-Level classification. in GLSVLSI 2018 - Proceedings of the 2018 Great Lakes Symposium on VLSI. Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI, Association for Computing Machinery, pp. 291-296, 28th Great Lakes Symposium on VLSI, GLSVLSI 2018, Chicago, United States, 5/23/18. https://doi.org/10.1145/3194554.3194603
Koteshwara S, Parhi KK. Low-Energy architectures of linear classifiers for IoT applications using incremental precision and multi-Level classification. In GLSVLSI 2018 - Proceedings of the 2018 Great Lakes Symposium on VLSI. Association for Computing Machinery. 2018. p. 291-296. (Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI). https://doi.org/10.1145/3194554.3194603
Koteshwara, Sandhya ; Parhi, Keshab K. / Low-Energy architectures of linear classifiers for IoT applications using incremental precision and multi-Level classification. GLSVLSI 2018 - Proceedings of the 2018 Great Lakes Symposium on VLSI. Association for Computing Machinery, 2018. pp. 291-296 (Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI).
@inproceedings{748a4cbba1b9467b80dc7383efdef14f,
title = "Low-Energy architectures of linear classifiers for IoT applications using incremental precision and multi-Level classification",
abstract = "This paper presents a novel incremental-precision classification approach that leads to a reduction in energy consumption of linear classifiers for IoT applications. Features are first input to a low-precision classifier. If the classifier successfully classifies the sample, then the process terminates. Otherwise, the classification performance is incrementally improved by using a classifier of higher precision. This process is repeated until the classification is complete. The argument is that many samples can be classified using the low-precision classifier, leading to a reduction in energy. To achieve incremental-precision, a novel data-path decomposition is proposed to design of fixed-width adders and multipliers. These components improve the precision without recalculating the outputs, thus reducing energy. Using a linear classification example, it is shown that the proposed incremental-precision based multi-level classifier approach can reduce energy by about 41{\%} while achieving comparable accuracies as that of a full-precision system.",
author = "Sandhya Koteshwara and Parhi, {Keshab K}",
year = "2018",
month = "5",
day = "30",
doi = "10.1145/3194554.3194603",
language = "English (US)",
series = "Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI",
publisher = "Association for Computing Machinery",
pages = "291--296",
booktitle = "GLSVLSI 2018 - Proceedings of the 2018 Great Lakes Symposium on VLSI",

}

TY - GEN

T1 - Low-Energy architectures of linear classifiers for IoT applications using incremental precision and multi-Level classification

AU - Koteshwara, Sandhya

AU - Parhi, Keshab K

PY - 2018/5/30

Y1 - 2018/5/30

N2 - This paper presents a novel incremental-precision classification approach that leads to a reduction in energy consumption of linear classifiers for IoT applications. Features are first input to a low-precision classifier. If the classifier successfully classifies the sample, then the process terminates. Otherwise, the classification performance is incrementally improved by using a classifier of higher precision. This process is repeated until the classification is complete. The argument is that many samples can be classified using the low-precision classifier, leading to a reduction in energy. To achieve incremental-precision, a novel data-path decomposition is proposed to design of fixed-width adders and multipliers. These components improve the precision without recalculating the outputs, thus reducing energy. Using a linear classification example, it is shown that the proposed incremental-precision based multi-level classifier approach can reduce energy by about 41% while achieving comparable accuracies as that of a full-precision system.

AB - This paper presents a novel incremental-precision classification approach that leads to a reduction in energy consumption of linear classifiers for IoT applications. Features are first input to a low-precision classifier. If the classifier successfully classifies the sample, then the process terminates. Otherwise, the classification performance is incrementally improved by using a classifier of higher precision. This process is repeated until the classification is complete. The argument is that many samples can be classified using the low-precision classifier, leading to a reduction in energy. To achieve incremental-precision, a novel data-path decomposition is proposed to design of fixed-width adders and multipliers. These components improve the precision without recalculating the outputs, thus reducing energy. Using a linear classification example, it is shown that the proposed incremental-precision based multi-level classifier approach can reduce energy by about 41% while achieving comparable accuracies as that of a full-precision system.

UR - http://www.scopus.com/inward/record.url?scp=85049486203&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85049486203&partnerID=8YFLogxK

U2 - 10.1145/3194554.3194603

DO - 10.1145/3194554.3194603

M3 - Conference contribution

AN - SCOPUS:85049486203

T3 - Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI

SP - 291

EP - 296

BT - GLSVLSI 2018 - Proceedings of the 2018 Great Lakes Symposium on VLSI

PB - Association for Computing Machinery

ER -