English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Deep neural networks (DNNs) have made great progress in various fields. Inparticular, the quantized neural network is a promising technique making DNNscompatible on resource-limited devices for memory and computation saving. Inthis paper, we mainly consider a non-convex minimization model with threeblocks to train quantized DNNs and propose a new stochastic three-blockalternating minimization (STAM) algorithm to solve it. We develop a convergencetheory for the STAM algorithm and obtain an $\epsilon$-stationary point withoptimal convergence rate $\mathcal{O}(\epsilon^{-4})$. Furthermore, we applyour STAM algorithm to train DNNs with relaxed binary weights. The experimentsare carried out on three different network structures, namely VGG-11, VGG-16and ResNet-18. These DNNs are trained using two different data sets, CIFAR-10and CIFAR-100, respectively. We compare our STAM algorithm with some classicalefficient algorithms for training quantized neural networks. The test accuracyindicates the effectiveness of STAM algorithm for training relaxed binaryquantization DNNs.

A stochastic three-block splitting algorithm and its application to quantized deep neural networks