A drop in model accuracy that happens when real-world inputs look meaningfully different from the data the model was trained on.
Continue to AI University →