1 paper · 1 filter
Miao Lu, Beining Wu, Xiaodong Yang +1
In this work, we theoretically investigate the generalization properties of neural networks (NN) trained by stochastic gradient descent (SGD) algorithm with large learning rates. U…