In function estimation (as opposed to classification), finiteness of which of the following is sufficient for PAC learning? VC-dimension; pseudo-dimension;

Popper dimension; dimension-X; fractal dimension; feature space dimension;

Hausdorff dimension.