The Fairness Indicator?
For Who, Why & What !!
A Fairness Indicator is a phrase you have to know and ask yourself the question who, why and what are these built for?
Let’s take a Dive:
Fairness indicators enable easy computation of commonly identified fairness metrics for binary and multiclass classifiers.
Many existing tools for evaluating fairness concerns don’t work well on large scale datasets and models.
Fairness Indicators includes the ability to:
1.Evaluate the distribution of datasets.
2.Evaluate model performance, sliced across defined groups of users.
3.Feel confident about your results with confidence intervals and evals at multiple thresholds.
4.Dive deep into individual slices to explore root causes and opportunities for improvement.
Fairness indicators are useful tools for evaluating binary and multi class classifiers for fairness.
But just keep in mind that quantitative evaluation is only one part of evaluating a broader user experience.
Start by thinking about the different contexts through which a user may experience your product.
