Data leakage occurs when data that should be outside the training dataset is explicitly or implicitly used to train a model. Data leakage can result in the incorrect estimation of the trained model’s performance.
Training data leakage is when data from the test set or validation set is explicitly or implicitly used to train a model. Future data leakage is when training data is constructed in such a way that training samples include feature values from the future. That is, training leaks knowledge of the future into the model, resulting in a model that performs well on the training set, but poorly on unseen data.
Feature data leakage is when you include a feature when training a model, but that feature is not available at inference time. You should, hopefully, discover feature data leakage when you try to write your inference pipeline and realize the feature is not available. The use of a feature store, that informs you whether your feature will be available for online models or only offline models, should help prevent you from training models with feature data leakage.