Teams first define relevant harms and measure them on representative data, then adjust data, objectives, thresholds, interfaces, or policies. Improvements on one fairness measure can conflict with another, making the chosen criteria important.
Bias mitigation aims to reduce systematic and unwanted performance or treatment differences across groups or contexts.
Teams first define relevant harms and measure them on representative data, then adjust data, objectives, thresholds, interfaces, or policies. Improvements on one fairness measure can conflict with another, making the chosen criteria important.