Pete Warden

“When a model is overfit, it has learned its training data too well. The model is able to exactly predict the minutiae of its training data, but it is not able to generalize its learning to data it has not previously seen. Often this happens because the model has managed to entirely memorize the training data, or it has learned to rely on a shortcut present in the training data but not in the real world.”
― Pete Warden, TinyML: Machine Learning with TensorFlow Lite on Arduino and Ultra-Low-Power Microcontrollers