Inductive Bias refers to a set of assumptions (biases) that a model makes (via design choices made by the developer) in order to perform induction (i.e. generalize to unseen data based on a set of examples).