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/

Zendy Plus

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

Zendy Tools

Zendy Open

Compression of convolutional neural network models has recently beendominated by pruning approaches. A class of previous works focuses solely onpruning the unimportant filters to achieve network compression. Anotherimportant direction is the design of sparsity-inducing constraints which hasalso been explored in isolation. This paper presents a novel training schemebased on composite constraints that prune redundant filters and minimize theireffect on overall network learning via sparsity promotion. Also, as opposed toprior works that employ pseudo-norm-based sparsity-inducing constraints, wepropose a sparse scheme based on gradient counting in our framework. Our testson several pixel-wise segmentation benchmarks show that the number of neuronsand the memory footprint of networks in the test phase are significantlyreduced without affecting performance. MobileNetV3 and UNet, two well-knownarchitectures, are used to test the proposed scheme. Our network compressionmethod not only results in reduced parameters but also achieves improvedperformance compared to MobileNetv3, which is an already optimizedarchitecture.

Neural Network Compression by Joint Sparsity Promotion and Redundancy Reduction