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

Position bias is a critical problem in information retrieval when dealingwith implicit yet biased user feedback data. Unbiased ranking methods typicallyrely on causality models and debias the user feedback through inversepropensity weighting. While practical, these methods still suffer from twomajor problems. First, when inferring a user click, the impact of thecontextual information, such as documents that have been examined, is oftenignored. Second, only the position bias is considered but other issues resultedfrom user browsing behaviors are overlooked. In this paper, we propose anend-to-end Deep Recurrent Survival Ranking (DRSR), a unified framework tojointly model user's various behaviors, to (i) consider the rich contextualinformation in the ranking list; and (ii) address the hidden issues underlyinguser behaviors, i.e., to mine observe pattern in queries without any click(non-click queries), and to model tracking logs which cannot truly reflect theuser browsing intents (untrusted observation). Specifically, we adopt arecurrent neural network to model the contextual information and estimates theconditional likelihood of user feedback at each position. We then incorporatesurvival analysis techniques with the probability chain rule to mathematicallyrecover the unbiased joint probability of one user's various behaviors. DRSRcan be easily incorporated with both point-wise and pair-wise learningobjectives. The extensive experiments over two large-scale industrial datasetsdemonstrate the significant performance gains of our model comparing with thestate-of-the-arts.

A Deep Recurrent Survival Model for Unbiased Ranking