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

This paper shows how to train binary networks to within a few percent points($\sim 3-5 \%$) of the full precision counterpart. We first show how to build astrong baseline, which already achieves state-of-the-art accuracy, by combiningrecently proposed advances and carefully adjusting the optimization procedure.Secondly, we show that by attempting to minimize the discrepancy between theoutput of the binary and the corresponding real-valued convolution, additionalsignificant accuracy gains can be obtained. We materialize this idea in twocomplementary ways: (1) with a loss function, during training, by matching thespatial attention maps computed at the output of the binary and real-valuedconvolutions, and (2) in a data-driven manner, by using the real-valuedactivations, available during inference prior to the binarization process, forre-scaling the activations right after the binary convolution. Finally, we showthat, when putting all of our improvements together, the proposed model beatsthe current state of the art by more than 5% top-1 accuracy on ImageNet andreduces the gap to its real-valued counterpart to less than 3% and 5% top-1accuracy on CIFAR-100 and ImageNet respectively when using a ResNet-18architecture. Code available at https://github.com/brais-martinez/real2binary.

Training Binary Neural Networks with Real-to-Binary Convolutions