You create a binary classification model and use the Fairlearn package to assess its fairness. You must avoid retraining the model. Which Fairlearn algorithm should you implement?