Efficient anti-symmetrization of a neural network layer by taming the sign problem

Open in new window