Neural network interpretability with layer-wise relevance propagation: novel techniques for neuron selection and visualization

Open in new window