Grey-box techniques for explainble AI make some use of the internal structure of a complex machine learning model, such as deep neural network. They choose an internal layer and then perform some analysis of the layer itself or apply black-box techniques to the parts of the network that lie either side of the layer.
Defined on pages 523, 526
Used on Chap. 21: pages 523, 526, 529
Grey-box techniques prising open the black-box at an internal layer