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

In an era of information explosion, recommendation systems play an importantrole in people's daily life by facilitating content exploration. It is knownthat user activeness, i.e., number of behaviors, tends to follow a long-taildistribution, where the majority of users are with low activeness. In practice,we observe that tail users suffer from significantly lower-qualityrecommendation than the head users after joint training. We further identifythat a model trained on tail users separately still achieve inferior resultsdue to limited data. Though long-tail distributions are ubiquitous inrecommendation systems, improving the recommendation performance on the tailusers still remains challenge in both research and industry. Directly applyingrelated methods on long-tail distribution might be at risk of hurting theexperience of head users, which is less affordable since a small portion ofhead users with high activeness contribute a considerate portion of platformrevenue. In this paper, we propose a novel approach that significantly improvesthe recommendation performance of the tail users while achieving at leastcomparable performance for the head users over the base model. The essence ofthis approach is a novel Gradient Aggregation technique that learns commonknowledge shared by all users into a backbone model, followed by separateplugin prediction networks for the head users and the tail userspersonalization. As for common knowledge learning, we leverage the backwardadjustment from the causality theory for deconfounding the gradient estimationand thus shielding off the backbone training from the confounder, i.e., useractiveness. We conduct extensive experiments on two public recommendationbenchmark datasets and a large-scale industrial datasets collected from theAlipay platform. Empirical studies validate the rationality and effectivenessof our approach.

Personalizing Intervened Network for Long-tailed Sequential User Behavior Modeling