A technique in ML where a pre-trained model is reused on a new but related task. This can significantly reduce the amount of data and computation needed for the new task. Examples include using a model trained on ImageNet for other Image Classification ta
