Fair Representation Learning with Controllable High Confidence Guarantees via Adversarial Inference | Read Paper on Bytez