- Two sided marketplace extension
- Recommendation system with fairness to suppliers with inverse-popularity score (like in our presentation)
- Feldman or Celis et al. methods of enforcing fairness constraints
- Show that problem is exponential in number of protected attributes
- Use approximation by being fair for each attribute and then performing rank aggregation
- Cant have a ranking that has all the following properties (Arrow 1950):
- Non-dictatorship
- Pareto efficiency (if everyone prefers A to B, then final order should prefer A to B)
- Independence of irrelevant alternatives (Given two inputs in which A and B are ranked identically by everyone, the two outputs should order A and B the same)
- Celis et al. formulate this as an optimization problem with constraints on the number from each type at each position.
- Large number of constraints. Can lead to infeasible problem.
- Need to specify the cutoffs that will guarantee fairness.
- FA*IR uses statistical test for a single protected attribute to generate a table of required protected candidates at each level
- Show that the problem is NP-hard