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

Online food ordering marketplaces are multi-stakeholder systems whererecommendations impact the experience and growth of each participant in thesystem. A recommender system in this setting has to encapsulate the objectivesand constraints of different stakeholders in order to find utility of an itemfor recommendation. Constrained-optimization based approaches to this problemtypically involve complex formulations and have high computational complexityin production settings involving millions of entities. Simplifications andrelaxation techniques (for example, scalarization) help but introducesub-optimality and can be time-consuming due to the amount of tuning needed. Inthis paper, we introduce a method involving multi-goal sampling followed byranking for user-relevance (Sample-Rank), to nudge recommendations towardsmulti-objective (MO) goals of the marketplace. The proposed method's novelty isthat it reduces the MO recommendation problem to sampling from a desiredmulti-goal distribution then using it to build a production-friendlylearning-to-rank (LTR) model. In offline experiments we show that we are ableto bias recommendations towards MO criteria with acceptable trade-offs inmetrics like AUC and NDCG. We also show results from a large-scale online A/Bexperiment where this approach gave a statistically significant lift of 2.64%in average revenue per order (RPO) (objective #1) with no drop in conversionrate (CR) (objective #2) while holding the average last-mile traversed flat(objective #3), vs. the baseline ranking method. This method also significantlyreduces time to model development and deployment in MO settings and allows fortrivial extensions to more objectives and other types of LTR models.

Sample-Rank: Weak Multi-Objective Recommendations Using Rejection Sampling