Momentum

/moʊˈmɛntəm/

An optimization technique that accelerates gradient descent by accumulating a velocity vector in directions of persistent reduction in the loss.

In italiano: Momentum

Momentum helps navigate ravines and accelerates convergence. It adds a fraction of the previous update vector to the current one, smoothing optimization trajectory and reducing oscillations.

Examples

  • SGD with momentum 0.9
  • Nesterov momentum
  • Accelerating convergence