A Bayesian approach to multi-task learning with network lasso - 42Papers