What are the three stages at which bias mitigation techniques can be applied in the machine learning pipeline according to the chapter?

Bias mitigation techniques can be applied at three stages in the machine learning pipeline: pre-processing, in-processing, and post-processing.

In the pre-processing stage, data can be adjusted to remove biased features before the training of the model begins. During the in-processing phase, model parameters are adjusted to minimize bias while training. Finally, in the post-processing stage, methods are employed to address any discriminatory patterns that the model has learned during inference.