When training and test data share the same conditioned Gaussian distribution, the paper's theory gives a vanilla-GD test-error plateau of order 1/(eta epsilon) only when both stated conditions hold, and no plateau if either fails; Figure 8 reports empirical agreement with the theoretical curve, while EGD has high test accuracy from the beginning of the plotted run. · CiteArk
When training and test data share the same conditioned Gaussian distribution, the paper's theory gives a vanilla-GD test-error plateau of order 1/(eta epsilon) only when both stated conditions hold, and no plateau if either fails; Figure 8 reports empirical agreement with the theoretical curve, while EGD has high test accuracy from the beginning of the plotted run.
来源：paper:PDF pp. 16-17, Figure 8, equation (19), and Theorem 2
