Current studentsExisting staffContact us
 

Secure Quantized Training for Deep Learning

Simple item page
dc.contributor.authorKeller, M.
dc.contributor.authorSun, K.
dc.contributor.conference39th International Conference on Machine Learning, ICML 2022 (17 Jul 2022 - 23 Jul 2022 : Baltimore, Maryland, USA, PMLR 162)
dc.contributor.editorK, K.
dc.contributor.editorJegelka, S.
dc.contributor.editorSong, L.
dc.contributor.editorSzepesvari, C.
dc.contributor.editorNiu, G.
dc.contributor.editorSabato, S.
dc.date.issued2022
dc.description.abstractWe implement training of neural networks in secure multi-party computation (MPC) using quantization commonly used in said setting. We are the first to present an MNIST classifier purely trained in MPC that comes within 0.2 percent of the accuracy of the same convolutional neural network trained via plaintext computation. More concretely, we have trained a network with two convolutional and two dense layers to 99.2% accuracy in 3.5 hours (under one hour for 99% accuracy). We have also implemented AlexNet for CIFAR-10, which converges in a few hours. We develop novel protocols for exponentiation and inverse square root. Finally, we present experiments in a range of MPC security models for up to ten parties, both with honest and dishonest majority as well as semi-honest and malicious security.
dc.identifier.citationProceedings of Machine Learning Research, 2022 / K, K., Jegelka, S., Song, L., Szepesvari, C., Niu, G., Sabato, S. (ed./s), pp.10912-10938
dc.identifier.issn2640-3498
dc.identifier.urihttps://hdl.handle.net/11541.2/38666
dc.language.isoen
dc.publisherJMLR
dc.publisher.placeUS
dc.relation.ispartofseries162, 2640-3498
dc.rightsCopyright 2022 by the author(s)
dc.source.urihttps://proceedings.mlr.press/v162/
dc.subjectconvolutional neural networks
dc.subjectdeep learning
dc.subjectmultilayer neural networks
dc.subjectnetwork security
dc.titleSecure Quantized Training for Deep Learning
dc.typeConference paper
pubs.publication-statusPublished
ror.fileinfo12285850200001831 13285850190001831 keller22a
ror.mmsid9916856324301831

Files

Original bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
keller22a.pdf
Size:
484.66 KB
Format:
Adobe Portable Document Format
Description:
Published version
Download

Collections