Bayesian Active Meta-Learning for Few Pilot Demodulation and Equalization