Abstract: We consider the problem of learning graphs from smooth signals in an online fashion. A major challenge of this task is that the underlying graph may be time-varying. Existing works typically ...