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/

Zendy Plus

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

Zendy Tools

Zendy Open

We study the problem of deep recall model in industrial web search, which is,given a user query, retrieve hundreds of most relevance documents from billionsof candidates. The common framework is to train two encoding models based onneural embedding which learn the distributed representations of queries anddocuments separately and match them in the latent semantic space. However, allthe exiting encoding models only leverage the information of the documentitself, which is often not sufficient in practice when matching with queryterms, especially for the hard tail queries. In this work we aim to leveragethe additional information for each document from its co-click neighbour tohelp document retrieval. The challenges include how to effectively extractinformation and eliminate noise when involving co-click information in deepmodel while meet the demands of billion-scale data size for real time onlineinference.  To handle the noise in co-click relations, we firstly propose a web-scaleMulti-Intention Co-click document Graph(MICG) which builds the co-clickconnections between documents on click intention level but not on documentlevel. Then we present an encoding framework MIRA based on Bert and graphattention networks which leverages a two-factor attention mechanism toaggregate neighbours. To meet the online latency requirements, we only involveneighbour information in document side, which can save the time-consuming queryneighbor search in real time serving. We conduct extensive offline experimentson both public dataset and private web-scale dataset from two major commercialsearch engines demonstrating the effectiveness and scalability of the proposedmethod compared with several baselines. And a further case study reveals thatco-click relations mainly help improve web search quality from two aspects: keyconcept enhancing and query term complementary.

MIRA: Leveraging Multi-Intention Co-click Information in Web-scale Document Retrieval using Deep Neural Networks