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

The strong correlation between neurons or filters can significantly weakenthe generalization ability of neural networks. Inspired by the well-knownTammes problem, we propose a novel diversity regularization method to addressthis issue, which makes the normalized weight vectors of neurons or filtersdistributed on a hypersphere as uniformly as possible, through maximizing theminimal pairwise angles (MMA). This method can easily exert its effect byplugging the MMA regularization term into the loss function with negligiblecomputational overhead. The MMA regularization is simple, efficient, andeffective. Therefore, it can be used as a basic regularization method in neuralnetwork training. Extensive experiments demonstrate that MMA regularization isable to enhance the generalization ability of various modern models andachieves considerable performance improvements on CIFAR100 and TinyImageNetdatasets. In addition, experiments on face verification show that MMAregularization is also effective for feature learning. Code is available at:https://github.com/wznpub/MMA_Regularization.

MMA Regularization: Decorrelating Weights of Neural Networks by Maximizing the Minimal Angles