Adversarial Sampling for Fairness Testing in Deep Neural Network