You don't have to "believe" the axioms. You can test them, by testing whether the experimental predictions derived from them are confirmed or not.