Full-Gradient Representation for Neural Network Visualization
This provides a more reliable interpretability tool for neural network users, though it is incremental in improving visualization methods.
The authors tackled the problem of interpreting neural network responses by introducing full-gradients, a representation that decomposes responses into input and per-neuron sensitivities, and proposed FullGrad for convolutional nets, which experimentally explained model behavior more comprehensively and produced sharper saliency maps than other methods.
We introduce a new tool for interpreting neural net responses, namely full-gradients, which decomposes the neural net response into input sensitivity and per-neuron sensitivity components. This is the first proposed representation which satisfies two key properties: completeness and weak dependence, which provably cannot be satisfied by any saliency map-based interpretability method. For convolutional nets, we also propose an approximate saliency map representation, called FullGrad, obtained by aggregating the full-gradient components. We experimentally evaluate the usefulness of FullGrad in explaining model behaviour with two quantitative tests: pixel perturbation and remove-and-retrain. Our experiments reveal that our method explains model behaviour correctly, and more comprehensively than other methods in the literature. Visual inspection also reveals that our saliency maps are sharper and more tightly confined to object regions than other methods.