Moreover, for gradient descent yt+1 =y t −η∇E(y t), there exists a stepsize η >0 and an open neighborhood U⊂ Y of ¯ysuch that every initialization y0 ∈U converges to ¯y

Therefore,¯yis a strict local minimum ofE, but not a global minimizer: E(¯y)> E(y⋆)