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

Binary Neural Networks (BNNs) are neural networks which use binary weightsand activations instead of the typical 32-bit floating point values. They havereduced model sizes and allow for efficient inference on mobile or embeddeddevices with limited power and computational resources. However, thebinarization of weights and activations leads to feature maps of lower qualityand lower capacity and thus a drop in accuracy compared to traditionalnetworks. Previous work has increased the number of channels or used multiplebinary bases to alleviate these problems. In this paper, we instead present anarchitectural approach: MeliusNet. It consists of alternating a DenseBlock,which increases the feature capacity, and our proposed ImprovementBlock, whichincreases the feature quality. Experiments on the ImageNet dataset demonstratethe superior performance of our MeliusNet over a variety of popular binaryarchitectures with regards to both computation savings and accuracy.Furthermore, with our method we trained BNN models, which for the first timecan match the accuracy of the popular compact network MobileNet-v1 in terms ofmodel size, number of operations and accuracy. Our code is published online athttps://github.com/hpi-xnor/BMXNet-v2

MeliusNet: Can Binary Neural Networks Achieve MobileNet-level Accuracy?