Category: Optimization Algorithms

• Stanford's New Theory Unravels the Mystery of Neural Network Generalization, Adding One Line of Code to Adam Yields 2.4x Speedup