RNN Training along Locally Optimal Trajectories via Frank-Wolfe Algorithm

Re-assign community