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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Click-through rate (CTR) prediction is a crucial task in online displayadvertising. The embedding-based neural networks have been proposed to learnboth explicit feature interactions through a shallow component and deep featureinteractions using a deep neural network (DNN) component. These sophisticatedmodels, however, slow down the prediction inference by at least hundreds oftimes. To address the issue of significantly increased serving delay and highmemory usage for ad serving in production, this paper presents\emph{DeepLight}: a framework to accelerate the CTR predictions in threeaspects: 1) accelerate the model inference via explicitly searching informativefeature interactions in the shallow component; 2) prune redundant layers andparameters at intra-layer and inter-layer level in the DNN component; 3)promote the sparsity of the embedding layer to preserve the most discriminantsignals. By combining the above efforts, the proposed approach accelerates themodel inference by 46X on Criteo dataset and 27X on Avazu dataset without anyloss on the prediction accuracy. This paves the way for successfully deployingcomplicated embedding-based neural networks in production for ad serving.

DeepLight: Deep Lightweight Feature Interactions for Accelerating CTR Predictions in Ad Serving