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

In this paper, we propose to train a network with binary weights andlow-bitwidth activations, designed especially for mobile devices with limitedpower consumption. Most previous works on quantizing CNNs uncritically assumethe same architecture, though with reduced precision. However, we take the viewthat for best performance it is possible (and even likely) that a differentarchitecture may be better suited to dealing with low precision weights andactivations.  Specifically, we propose a "network expansion" strategy in which we aggregatea set of homogeneous low-precision branches to implicitly reconstruct thefull-precision intermediate feature maps. Moreover, we also propose agroup-wise feature approximation strategy which is very flexible and highlyaccurate. Experiments on ImageNet classification tasks demonstrate the superiorperformance of the proposed model, named Group-Net, over various populararchitectures. In particular, with binary weights and activations, weoutperform the previous best binary neural network in terms of accuracy as wellas saving more than 5 times computational complexity on ImageNet with ResNet-18and ResNet-50.

Training Compact Neural Networks with Binary Weights and Low Precision Activations