Joint Learning of Linear Time-Invariant Dynamical Systems