Why not combine Nesterov Accelerated GD and RMSProp? - Optimization Algorithms

In optimization Algorithms, we have discussed Adam optimizer as a combination of features of Momentum Based GD and RMSProp. Is there a mathematical reason for not combining NAG with RMSProp or is it that no researchers have not yet published such a combination?

I think We do have one, called NAdam Optimizer.
Paper | Explanation | Keras API

1 Like