Algorithmic decision making and the cost of fairness

Algorithmic decision making and the cost of fairness

Sam Corbett-Davis, etc

Stanford

Intro

将fairness作为constrained optimization。

Background

$$x_i \in \mathbb{R}^p$$, $$d(x) \in [0, 1]$$.

Def 2.1 (Decision rule) A decision rule is any measurable function $$d: \mathbb{R}^p \to [0, 1]$$, where $$d(x)$$ is the probability that action $$a_1$$ is taken for an individual with feature $$x$$.