All you need is a good init
This addresses the challenge of training very deep networks efficiently for researchers and practitioners in machine learning, though it is incremental as it builds on existing initialization techniques.
The paper tackles the problem of weight initialization for deep neural networks by proposing LSUV initialization, a simple two-step method that normalizes layer outputs to unit variance, resulting in test accuracy better or equal to standard methods and achieving state-of-the-art or close performance on datasets like MNIST, CIFAR-10/100, and ImageNet.
Layer-sequential unit-variance (LSUV) initialization - a simple method for weight initialization for deep net learning - is proposed. The method consists of the two steps. First, pre-initialize weights of each convolution or inner-product layer with orthonormal matrices. Second, proceed from the first to the final layer, normalizing the variance of the output of each layer to be equal to one. Experiment with different activation functions (maxout, ReLU-family, tanh) show that the proposed initialization leads to learning of very deep nets that (i) produces networks with test accuracy better or equal to standard methods and (ii) is at least as fast as the complex schemes proposed specifically for very deep nets such as FitNets (Romero et al. (2015)) and Highway (Srivastava et al. (2015)). Performance is evaluated on GoogLeNet, CaffeNet, FitNets and Residual nets and the state-of-the-art, or very close to it, is achieved on the MNIST, CIFAR-10/100 and ImageNet datasets.