Show simple item record

dc.contributor.authorPerić, Zoran
dc.contributor.authorSavić, Milan
dc.contributor.authorSimić, Nikola
dc.contributor.authorDenić, Bojan
dc.contributor.authorDespotović, Vladimir
dc.date.accessioned2023-04-11T10:48:52Z
dc.date.available2023-04-11T10:48:52Z
dc.date.issued2021-07-22
dc.identifier.citationIII44006en_US
dc.identifier.urihttps://platon.pr.ac.rs/handle/123456789/1191
dc.description.abstractAchieving real-time inference is one of the major issues in contemporary neural network applications, as complex algorithms are frequently being deployed to mobile devices that have constrained storage and computing power. Moving from a full-precision neural network model to a lower representation by applying quantization techniques is a popular approach to facilitate this issue. Here, we analyze in detail and design a 2-bit uniform quantization model for Laplacian source due to its significance in terms of implementation simplicity, which further leads to a shorter processing time and faster inference. The results show that it is possible to achieve high classification accuracy (more than 96% in the case of MLP and more than 98% in the case of CNN) by implementing the proposed model, which is competitive to the performance of the other quantization solutions with almost optimal precision.en_US
dc.language.isoen_USen_US
dc.publisherMolecular Diversity Preservation Internationalen_US
dc.titleDesign of a 2-Bit Neural Network Quantizer for Laplacian Sourceen_US
dc.title.alternativeEntropyen_US
dc.typeclanak-u-casopisuen_US
dc.description.versionpublishedVersionen_US
dc.identifier.doihttps://doi.org/10.3390/e23080933
dc.citation.volume23
dc.citation.issue8
dc.subject.keywordsimage classificationen_US
dc.subject.keywordsLaplacian sourceen_US
dc.subject.keywordsneural networken_US
dc.subject.keywordsquantizationen_US
dc.type.mCategoryM22en_US
dc.type.mCategoryopenAccessen_US
dc.type.mCategoryM22en_US
dc.type.mCategoryopenAccessen_US


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record