Implicit regularisation in stochastic gradient descent: from single-objective to two-player games

Open in new window