For a function whose input is a multi-dimensional vector and output is a scalar, a zero-gradient position represents a local maximum if all the eigenvalues of the function's Hessian matrix at that position are negative.