Algorithm 1 Gradient Ascent for solving the optimization.

1:$k=0,{\mathbf{f}}^0={\mathbf{f}}^{initial}$ 2:repeat 3:    Compute the gradient: $\nabla \nu \left({\mathbf{f}}^k\right)$ 4:    Choose a step size: $t^k>0$ 5:    Update ${\mathbf{f}}^{k+1}:={\mathbf{f}}^k+t^k\nabla \nu ({\mathbf{f}}^k$) 6:$k\leftarrow k+1$ 7:until$max\left(\right|\nabla \nu \left({\mathbf{f}}^k\right)\left|\right)\le \eta$ 8:return${\mathbf{f}}^k$.