Linear interpolation gives better gradients than Gaussian smoothing in derivative-free optimization

Open in new window