Commentary activity: similarity-based learning. Similarity-based machine learning works by applying an identified inductive bias to training data, developing a model that can then be used to make decisions about new cases. But there's a catch. Can you identify any potential problems with using similarity-based learning to inform decisions about, e.g.: 1) whether to offer someone a loan to buy a home? 2) whether to offer someone a job? 3) whether to release someone on bail pending trial, or to remand them in custody (until they may be tried)? Hint: consider the stationarity' assumption