Deep learning models excel at learning from a large number of labeled examples, but typically do not generalize to conditions not seen during training. This post gives an overview of transfer learning, motivates why it warrants our application, and discusses practical applications and methods.