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 prediction is one of the core tasks in commercialrecommender systems. It aims to predict the probability of a user clicking aparticular item given user and item features. As feature interactions bring innon-linearity, they are widely adopted to improve the performance of CTRprediction models. Therefore, effectively modelling feature interactions hasattracted much attention in both the research and industry field. The currentapproaches can generally be categorized into three classes: (1) na\"ivemethods, which do not model feature interactions and only use originalfeatures; (2) memorized methods, which memorize feature interactions byexplicitly viewing them as new features and assigning trainable embeddings; (3)factorized methods, which learn latent vectors for original features andimplicitly model feature interactions through factorization functions. Studieshave shown that modelling feature interactions by one of these methods aloneare suboptimal due to the unique characteristics of different featureinteractions. To address this issue, we first propose a general frameworkcalled OptInter which finds the most suitable modelling method for each featureinteraction. Different state-of-the-art deep CTR models can be viewed asinstances of OptInter. To realize the functionality of OptInter, we alsointroduce a learning algorithm that automatically searches for the optimalmodelling method. We conduct extensive experiments on four large datasets. Ourexperiments show that OptInter improves the best performed state-of-the-artbaseline deep CTR models by up to 2.21%. Compared to the memorized method,which also outperforms baselines, we reduce up to 91% parameters. In addition,we conduct several ablation studies to investigate the influence of differentcomponents of OptInter. Finally, we provide interpretable discussions on theresults of OptInter.

Memorize, Factorize, or be Na\"ive: Learning Optimal Feature Interaction Methods for CTR Prediction