Towards Understanding Learning in Neural Networks with Linear Teachers
Roei Sarussi 1 Alon Brutzkus 1 Amir Globerson 1
Abstract will separate the data. Which of these will be found by
Can a neural network minimizing cross-entropy gradient descent?
learn linearly separable data? Despite progress
in the theory of deep learning, this question re-
Our work addresses these questions as follows.
...


雷达卡




京公网安备 11010802022788号







