Toward breaking the curse of dimensionality in (ro)vibrational computations of molecular systems with multiple large-amplitude motions.