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 network architectures have attained remarkable improvements inscene understanding tasks. Utilizing an efficient model is one of the mostimportant constraints for limited-resource devices. Recently, severalcompression methods have been proposed to diminish the heavy computationalburden and memory consumption. Among them, the pruning and quantizing methodsexhibit a critical drop in performances by compressing the model parameters.While the knowledge distillation methods improve the performance of compactmodels by focusing on training lightweight networks with the supervision ofcumbersome networks. In the proposed method, the knowledge distillation hasbeen performed within the network by constructing multiple branches over theprimary stream of the model, known as the self-distillation method. Therefore,the ensemble of sub-neural network models has been proposed to transfer theknowledge among themselves with the knowledge distillation policies as well asan adversarial learning strategy. Hence, The proposed ensemble of sub-models istrained against a discriminator model adversarially. Besides, their knowledgeis transferred within the ensemble by four different loss functions. Theproposed method has been devoted to both lightweight image classification andencoder-decoder architectures to boost the performance of small and compactmodels without incurring extra computational overhead at the inference process.Extensive experimental results on the main challenging datasets show that theproposed network outperforms the primary model in terms of accuracy at the samenumber of parameters and computational cost. The obtained results show that theproposed model has achieved significant improvement over earlier ideas ofself-distillation methods. The effectiveness of the proposed models has alsobeen illustrated in the encoder-decoder model.

Be Your Own Best Competitor! Multi-Branched Adversarial Knowledge Transfer