Frugality in second-order optimization: floating-point approximations for Newton's method

Open in new window