What is Deep SHAP?
Deep explainer (deep SHAP) is an explainability technique that can be used for models with a neural network based architecture. This is the fastest neural network explainability approach and is based on running a SHAP-based version of the original deep lift algorithm.
A popular way to leverage SHAP values to explain predictions of deep learning models (neural networks) is with the method DeepExplainer. DeepExplainer runs on deep learning frameworks to add explainability to neural network models by using DeepLIFT and Shapley values. DeepExplainer is an enhanced version of the DeepLIFT (Deep Learning Important FeaTures), a method for decomposing the output prediction of a neural network on a specific input by backpropagating the contributions of all neurons in the network to every feature of the input.
Lundberg and Lee, NIPS 2017, showed that the per node attribution rules in DeepLIFT can be chosen to approximate Shapley values. By integrating over many background samples, DeepExplainer (deep SHAP) estimates approximate SHAP values such that they sum up to the difference between the expected model output on the passed background samples and the current model output (f(x) – E[f(x)]).