Skip to yearly menu bar Skip to main content


Gradient Descent Finds Global Minima of Deep Neural Networks

Simon Du · Jason Lee · Haochuan Li · Liwei Wang · Xiyu Zhai

Pacific Ballroom #80

Keywords: [ Non-convex Optimization ] [ Deep Learning Theory ]


Gradient descent finds a global minimum in training deep neural networks despite the objective function being non-convex. The current paper proves gradient descent achieves zero training loss in polynomial time for a deep over-parameterized neural network with residual connections (ResNet). Our analysis relies on the particular structure of the Gram matrix induced by the neural network architecture. This structure allows us to show the Gram matrix is stable throughout the training process and this stability implies the global optimality of the gradient descent algorithm. We further extend our analysis to deep residual convolutional neural networks and obtain a similar convergence result.

Live content is unavailable. Log in and register to view live content