Machine learning approaches for CO2 geological storage monitoring: autoencoders for solving the problems of acquisitions repeatability