Michael Huemer says that Bayesian probability can solve the problem of induction. But is that all that different from justified opinion? That last being the flimsy definition of knowledge. Dr. Huemer' opinion is surely better than that, but still does not bridge the gap.  A priori is different in essence from empirical knowledge as Leonard Nelson pointed out.