Counterfactual Fairness by Combining Factual and Counterfactual Predictions

Counterfactual Fairness (CF), which posits that an ML model's outcome on any individual should remain unchanged if they had belonged to a different demographic