Ethical Adversaries: Towards Mitigating Unfairness with Adversarial Machine Learning