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

We propose WaveMix -- a novel neural architecture for computer vision that isresource-efficient yet generalizable and scalable. WaveMix networks achievecomparable or better accuracy than the state-of-the-art convolutional neuralnetworks, vision transformers, and token mixers for several tasks, establishingnew benchmarks for segmentation on Cityscapes; and for classification onPlaces-365, five EMNIST datasets, and iNAT-mini. Remarkably, WaveMixarchitectures require fewer parameters to achieve these benchmarks compared tothe previous state-of-the-art. Moreover, when controlled for the number ofparameters, WaveMix requires lesser GPU RAM, which translates to savings intime, cost, and energy. To achieve these gains we used multi-leveltwo-dimensional discrete wavelet transform (2D-DWT) in WaveMix blocks, whichhas the following advantages: (1) It reorganizes spatial information based onthree strong image priors -- scale-invariance, shift-invariance, and sparsenessof edges, (2) in a lossless manner without adding parameters, (3) while alsoreducing the spatial sizes of feature maps, which reduces the memory and timerequired for forward and backward passes, and (4) expanding the receptive fieldfaster than convolutions do. The whole architecture is a stack of self-similarand resolution-preserving WaveMix blocks, which allows architecturalflexibility for various tasks and levels of resource availability. Our code andtrained models are publicly available.

WaveMix: A Resource-efficient Neural Network for Image Analysis