Leveraging Posit Arithmetic in Deep Neural Networks

The IEEE 754 Standard for Floating-Point Arithmetic has been for decades imple mented in the vast majority of modern computer systems to manipulate and com pute real numbers. Recently, John L. Gustafson introduced a new data type called positTM to represent real numbers on computers. This emerging f...

Descripción completa

Detalles Bibliográficos
Autor: Murillo Montero, Raúl
Tipo de recurso: tesis de maestría
Fecha de publicación:2021
País:España
Institución:Universidad Complutense de Madrid (UCM)
Repositorio:Docta Complutense
Idioma:español
OAI Identifier:oai:docta.ucm.es:20.500.14352/9180
Acceso en línea:https://hdl.handle.net/20.500.14352/9180
Access Level:acceso abierto
Palabra clave:004(043.3)
Posit arithmetic
Deep neural networks
Training
Inference
Adder
Multiplier
Computer arithmetic.
Aritmética Posit
Redes neuronales profundas
Entrenamiento
Inferencia
Sumador
Multiplicador
Aritmética de computadores.
Informática (Informática)
1203.17 Informática
id ES_e9bb4619deed2f2f01baf0238a0efecb
oai_identifier_str oai:docta.ucm.es:20.500.14352/9180
network_acronym_str ES
network_name_str España
repository_id_str
spelling Leveraging Posit Arithmetic in Deep Neural NetworksAprovechando la Aritmética Posit en las redes neuronales profundasMurillo Montero, Raúl004(043.3)Posit arithmeticDeep neural networksTrainingInferenceAdderMultiplierComputer arithmetic.Aritmética PositRedes neuronales profundasEntrenamientoInferenciaSumadorMultiplicadorAritmética de computadores.Informática (Informática)1203.17 InformáticaThe IEEE 754 Standard for Floating-Point Arithmetic has been for decades imple mented in the vast majority of modern computer systems to manipulate and com pute real numbers. Recently, John L. Gustafson introduced a new data type called positTM to represent real numbers on computers. This emerging format was designed with the aim of replacing IEEE 754 floating-point numbers by providing certain ad vantages over them, such as a larger dynamic range, higher accuracy, bitwise iden tical results across systems, or simpler hardware, among others. The interesting properties of the posit format seem to be really useful under the scenario of deep neural networks. In this Master’s thesis, the properties of posit arithmetic are studied with the aim of leveraging them for the training and inference of deep neural networks. For this purpose, a framework for neural networks based on the posit format is developed. The results show that posits can achieve similar accuracy results as floating-point numbers with half of the bit width without modifications in the training and infer ence flows of deep neural networks. The hardware cost of the posit arithmetic units needed for operating with neural networks (this is, additions and multiplications) is also studied in this work, obtaining great improvements in terms of area and power savings with respect state-of-the-art implementations.Barrio García, Alberto Antonio delBotella Juan, GuillermoUniversidad Complutense de Madrid20212021-01-0120212021-01-01master thesishttp://purl.org/coar/resource_type/c_bdccinfo:eu-repo/semantics/masterThesisapplication/pdfhttps://hdl.handle.net/20.500.14352/9180reponame:Docta Complutenseinstname:Universidad Complutense de Madrid (UCM)Españolspaopen accesshttp://purl.org/coar/access_right/c_abf2Atribución-NoComercial 3.0 Españahttps://creativecommons.org/licenses/by-nc/3.0/es/info:eu-repo/semantics/openAccessoai:docta.ucm.es:20.500.14352/91802026-06-02T12:44:21Z
dc.title.none.fl_str_mv Leveraging Posit Arithmetic in Deep Neural Networks
Aprovechando la Aritmética Posit en las redes neuronales profundas
title Leveraging Posit Arithmetic in Deep Neural Networks
spellingShingle Leveraging Posit Arithmetic in Deep Neural Networks
Murillo Montero, Raúl
004(043.3)
Posit arithmetic
Deep neural networks
Training
Inference
Adder
Multiplier
Computer arithmetic.
Aritmética Posit
Redes neuronales profundas
Entrenamiento
Inferencia
Sumador
Multiplicador
Aritmética de computadores.
Informática (Informática)
1203.17 Informática
title_short Leveraging Posit Arithmetic in Deep Neural Networks
title_full Leveraging Posit Arithmetic in Deep Neural Networks
title_fullStr Leveraging Posit Arithmetic in Deep Neural Networks
title_full_unstemmed Leveraging Posit Arithmetic in Deep Neural Networks
title_sort Leveraging Posit Arithmetic in Deep Neural Networks
dc.creator.none.fl_str_mv Murillo Montero, Raúl
author Murillo Montero, Raúl
author_facet Murillo Montero, Raúl
author_role author
dc.contributor.none.fl_str_mv Barrio García, Alberto Antonio del
Botella Juan, Guillermo
Universidad Complutense de Madrid
dc.subject.none.fl_str_mv 004(043.3)
Posit arithmetic
Deep neural networks
Training
Inference
Adder
Multiplier
Computer arithmetic.
Aritmética Posit
Redes neuronales profundas
Entrenamiento
Inferencia
Sumador
Multiplicador
Aritmética de computadores.
Informática (Informática)
1203.17 Informática
topic 004(043.3)
Posit arithmetic
Deep neural networks
Training
Inference
Adder
Multiplier
Computer arithmetic.
Aritmética Posit
Redes neuronales profundas
Entrenamiento
Inferencia
Sumador
Multiplicador
Aritmética de computadores.
Informática (Informática)
1203.17 Informática
description The IEEE 754 Standard for Floating-Point Arithmetic has been for decades imple mented in the vast majority of modern computer systems to manipulate and com pute real numbers. Recently, John L. Gustafson introduced a new data type called positTM to represent real numbers on computers. This emerging format was designed with the aim of replacing IEEE 754 floating-point numbers by providing certain ad vantages over them, such as a larger dynamic range, higher accuracy, bitwise iden tical results across systems, or simpler hardware, among others. The interesting properties of the posit format seem to be really useful under the scenario of deep neural networks. In this Master’s thesis, the properties of posit arithmetic are studied with the aim of leveraging them for the training and inference of deep neural networks. For this purpose, a framework for neural networks based on the posit format is developed. The results show that posits can achieve similar accuracy results as floating-point numbers with half of the bit width without modifications in the training and infer ence flows of deep neural networks. The hardware cost of the posit arithmetic units needed for operating with neural networks (this is, additions and multiplications) is also studied in this work, obtaining great improvements in terms of area and power savings with respect state-of-the-art implementations.
publishDate 2021
dc.date.none.fl_str_mv 2021
2021-01-01
2021
2021-01-01
dc.type.none.fl_str_mv master thesis
http://purl.org/coar/resource_type/c_bdcc
dc.type.openaire.fl_str_mv info:eu-repo/semantics/masterThesis
format masterThesis
dc.identifier.none.fl_str_mv https://hdl.handle.net/20.500.14352/9180
url https://hdl.handle.net/20.500.14352/9180
dc.language.none.fl_str_mv Español
spa
language_invalid_str_mv Español
language spa
dc.rights.none.fl_str_mv open access
http://purl.org/coar/access_right/c_abf2
Atribución-NoComercial 3.0 España
https://creativecommons.org/licenses/by-nc/3.0/es/
dc.rights.openaire.fl_str_mv info:eu-repo/semantics/openAccess
rights_invalid_str_mv open access
http://purl.org/coar/access_right/c_abf2
Atribución-NoComercial 3.0 España
https://creativecommons.org/licenses/by-nc/3.0/es/
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
dc.source.none.fl_str_mv reponame:Docta Complutense
instname:Universidad Complutense de Madrid (UCM)
instname_str Universidad Complutense de Madrid (UCM)
reponame_str Docta Complutense
collection Docta Complutense
repository.name.fl_str_mv
repository.mail.fl_str_mv
_version_ 1869423077144657920
score 15.300724

Similares en LA Referencia