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

Machine learning models (e.g., speech recognizers) are usually trained tominimize average loss, which results in representation disparity---minoritygroups (e.g., non-native speakers) contribute less to the training objectiveand thus tend to suffer higher loss. Worse, as model accuracy affects userretention, a minority group can shrink over time. In this paper, we first showthat the status quo of empirical risk minimization (ERM) amplifiesrepresentation disparity over time, which can even make initially fair modelsunfair. To mitigate this, we develop an approach based on distributionallyrobust optimization (DRO), which minimizes the worst case risk over alldistributions close to the empirical distribution. We prove that this approachcontrols the risk of the minority group at each time step, in the spirit ofRawlsian distributive justice, while remaining oblivious to the identity of thegroups. We demonstrate that DRO prevents disparity amplification on exampleswhere ERM fails, and show improvements in minority group user satisfaction in areal-world text autocomplete task.

Fairness Without Demographics in Repeated Loss Minimization