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

We present a novel evaluation framework for representation bias in latentfactor recommendation (LFR) algorithms. Our framework introduces the concept ofattribute association bias in recommendations allowing practitioners to explorehow recommendation systems can introduce or amplify stakeholder representationharm. Attribute association bias (AAB) occurs when sensitive attributes becomesemantically captured or entangled in the trained recommendation latent space.This bias can result in the recommender reinforcing harmful stereotypes, whichmay result in downstream representation harms to system consumer and providerstakeholders. LFR models are at risk of experiencing AAB due to their abilityto entangle explicit and implicit attributes into the trained latent space.Understanding this phenomenon is essential due to the increasingly common useof entity vectors as attributes in downstream components in hybrid industryrecommendation systems. We provide practitioners with a framework for executingdisaggregated evaluations of AAB within broader algorithmic auditingframeworks. Inspired by research in natural language processing (NLP) observinggender bias in word embeddings, our framework introduces AAB evaluation methodsspecifically for recommendation entity vectors. We present four evaluationstrategies for sensitive AAB in LFR models: attribute bias directions,attribute association bias metrics, classification for explaining bias, andlatent space visualization. We demonstrate the utility of our framework byevaluating user gender AAB regarding podcast genres with an industry case studyof a production-level DNN recommendation model. We uncover significant levelsof user gender AAB when user gender is used and removed as a model featureduring training, pointing to the potential for systematic bias in LFR modeloutputs.

Evaluation Framework for Understanding Sensitive Attribute Association Bias in Latent Factor Recommendation Algorithms