Injecting fairness into machine-learning models

Paper: https://openreview.net/pdf?id=js62_xuLDDv

I thought it was garbage in, garbage out. Here they seem to be suggesting garbage in, fairness out.