New Study Examines Fisher Consistency and Calibration in Fairness-Aware Classification Losses
A new arXiv paper analyzes surrogate losses for fairness-aware classification, focusing on their Fisher consistency and calibration properties when balancing accuracy and fairness across demographic groups.
Read the full story at arXiv stat.ML ↗
Timeline · 1 report
- 2026-11-03 05:00 · arXiv stat.ML
Surrogate Losses for Fairness-Aware Classification: Fisher Consistency and Calibration