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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Considering the impact of recommendations on item providers is one of theduties of multi-sided recommender systems. Item providers are key stakeholdersin online platforms, and their earnings and plans are influenced by theexposure their items receive in recommended lists. Prior work showed thatcertain minority groups of providers, characterized by a common sensitiveattribute (e.g., gender or race), are being disproportionately affected byindirect and unintentional discrimination. Our study in this paper handles asituation where ($i$) the same provider is associated with multiple items of alist suggested to a user, ($ii$) an item is created by more than one providerjointly, and ($iii$) predicted user-item relevance scores are biasedlyestimated for items of provider groups. Under this scenario, we assessdisparities in relevance, visibility, and exposure, by simulating diverserepresentations of the minority group in the catalog and the interactions.Based on emerged unfair outcomes, we devise a treatment that combinesobservation upsampling and loss regularization, while learning user-itemrelevance scores. Experiments on real-world data demonstrate that our treatmentleads to lower disparate relevance. The resulting recommended lists show fairervisibility and exposure, higher minority item coverage, and negligible loss inrecommendation utility.

Interplay between Upsampling and Regularization for Provider Fairness in Recommender Systems