A new fairness audit metric, Predictive Representativity, reveals that five common skin-cancer classifiers trained on HAM10000 have much lower precision for darker Fitzpatrick skin types when tested on an independent Colombian dataset.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Predictive Representativity: Uncovering Racial Bias in AI-based Skin Cancer Detection
A new fairness audit metric, Predictive Representativity, reveals that five common skin-cancer classifiers trained on HAM10000 have much lower precision for darker Fitzpatrick skin types when tested on an independent Colombian dataset.