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

Gating modules have been widely explored in dynamic network pruning to reducethe run-time computational cost of deep neural networks while preserving therepresentation of features. Despite the substantial progress, existing methodsremain ignoring the consistency between feature and gate distributions, whichmay lead to distortion of gated features. In this paper, we propose afeature-gate coupling (FGC) approach aiming to align distributions of featuresand gates. FGC is a plug-and-play module, which consists of two steps carriedout in an iterative self-supervised manner. In the first step, FGC utilizes the$k$-Nearest Neighbor method in the feature space to explore instanceneighborhood relationships, which are treated as self-supervisory signals. Inthe second step, FGC exploits contrastive learning to regularize gating moduleswith generated self-supervisory signals, leading to the alignment of instanceneighborhood relationships within the feature and gate spaces. Experimentalresults validate that the proposed FGC method improves the baseline approachwith significant margins, outperforming the state-of-the-arts with betteraccuracy-computation trade-off. Code is publicly available.

Feature-Gate Coupling for Dynamic Network Pruning