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

Common fairness definitions in machine learning focus on balancing notions ofdisparity and utility. In this work, we study fairness in the context of riskdisparity among sub-populations. We are interested in learning models thatminimize performance discrepancies across sensitive groups without causingunnecessary harm. This is relevant to high-stakes domains such as healthcare,where non-maleficence is a core principle. We formalize this objective usingPareto frontiers, and provide analysis, based on recent works in fairness, toexemplify scenarios were perfect fairness might not be feasible without doingunnecessary harm. We present a methodology for training neural networks thatachieve our goal by dynamically re-balancing subgroups risks. We argue thateven in domains where fairness at cost is required, finding anon-unnecessary-harm fairness model is the optimal initial step. We demonstratethis methodology on real case-studies of predicting ICU patient mortality, andclassifying skin lesions from dermatoscopic images.

Fairness With Minimal Harm: A Pareto-Optimal Approach For Healthcare