We consider the problem of learning a one-hidden-layer neural network with non-overlapping convolutional layer and ReLU activation function, i.e., $f(\mathbf{Z}; \mathbf{w}, \mathbf{a}) = \sum_j a_j\sigma(\mathbf{w}^\top\mathbf{Z}_j)$, in which both the convolutional weights $\mathbf{w}$ and the output weights $\mathbf{a}$ are parameters to be learned. We prove that with Gaussian input $\mathbf{Z}$, there is a spurious local minimum that is not a global mininum. Surprisingly, in the presence of local minimum, starting from randomly initialized weights, gradient descent with weight normalization can still be proven to recover the true parameters with constant probability (which can be boosted to arbitrarily high accuracy with multiple restarts). We also show that with constant probability, the same procedure could also converge to the spurious local minimum, showing that the local minimum plays a non-trivial role in the dynamics of gradient descent. Furthermore, a quantitative analysis shows that the gradient descent dynamics has two phases: it starts off slow, but converges much faster after several iterations.

研究了在卷积层和ReLU激活下的一层神经网络的学习问题，证明了随机初始化并使用归一化权重的梯度下降可以恢复真实参数，但存在虚假局部最小值，且该局部最小值在梯度下降的动力学中起到了重要作用。

梯度下降学习一层卷积神经网络：不必担心虚假局部极小值