How do different optimization algorithms such as SGD, Adam, and RMSprop navigate the loss landscape? Describe how momentum helps accelerate training.
