Are deep ResNets provably better than linear predictors?
arXiv:1907.03922
Abstract
Recent results in the literature indicate that a residual network (ResNet) composed of a single residual block outperforms linear predictors, in the sense that all local minima in its optimization landscape are at least as good as the best linear predictor. However, these results are limited to a single residual block (i.e., shallow ResNets), instead of the deep ResNets composed of multiple residual blocks. We take a step towards extending this result to deep ResNets. We start by two motivating examples. First, we show that there exist datasets for which all local minima of a fully-connected ReLU network are no better than the best linear predictor, whereas a ResNet has strictly better local minima. Second, we show that even at the global minimum, the representation obtained from the residual block outputs of a 2-block ResNet do not necessarily improve monotonically over subsequent blocks, which highlights a fundamental difficulty in analyzing deep ResNets. Our main theorem on deep ResNets shows under simple geometric conditions that, any critical point in the optimization landscape is either (i) at least as good as the best linear predictor; or (ii) the Hessian at this critical point has a strictly negative eigenvalue. Notably, our theorem shows that a chain of multiple skip-connections can improve the optimization landscape, whereas existing results study direct skip-connections to the last hidden layer or output layer. Finally, we complement our results by showing benign properties of the "near-identity regions" of deep ResNets, showing depth-independent upper bounds for the risk attained at critical points as well as the Rademacher complexity.
15 pages. NeurIPS 2019 Camera-ready version
References in corpus (18)
- Spectrally-normalized margin bounds for neural networks
- The loss surface of deep and wide neural networks
- No bad local minima: Data independent training error guarantees for multilayer neural networks
- ResNet with one-neuron hidden layers is a Universal Approximator
- Spurious Local Minima are Common in Two-Layer ReLU Neural Networks
- Norm-Based Capacity Control in Neural Networks
- Diverse Neural Network Learns True Target Functions
- Depth with Nonlinearity Creates No Bad Local Minima in ResNets
- On the loss landscape of a class of deep neural networks with no bad local valleys
- Understanding the Loss Surface of Neural Networks for Binary Classification
- Adding One Neuron Can Eliminate All Bad Local Minima
- Local minima in training of neural networks
- Are ResNets Provably Better than Linear Predictors?
- Critical Points of Neural Networks: Analytical Forms and Landscape Properties
- Small nonlinearities in activation functions create bad local minima in neural networks
- Representing smooth functions as compositions of near-identity functions with implications for deep network optimization
- The Multilinear Structure of ReLU Networks
- Efficiently testing local optimality and escaping saddles for ReLU networks