Fetching the paper…

Efficient Per-Example Gradient Computations in Convolutional Neural Networks · Around