Nesterov加速梯度（NA, NAG）

Nesterov accelerated gradient (NAG) is an iterative optimization algorithm designed for solving optimization problems. It possesses a faster rate of convergence compared to other optimization methods. NAG achieves acceleration by incorporating an 'acquisition' term into the gradient update process. The approach differs from the standard gradient descent method in that it considers each iteration takes into account the updates from the previous iteration, thereby making the direction of gradient descent more precise and enabling faster convergence toward the minima of the objective function.

以