Online learning has emerged as powerful tool in large scale optimization. In this work, we generalize the Douglas-Rachford splitting method for minimizing composite functions to online settings.
