Verifying Individual Fairness in Machine Learning Models