S-SGD: Symmetrical Stochastic Gradient Descent with Weight Noise Injection for Reaching Flat Minima
This is an incremental improvement for deep learning practitioners seeking better generalization in neural network training.
The paper tackles the problem of SGD converging to sharp minima with poor generalization by proposing a symmetrical weight noise injection method that evaluates loss at two adjacent points to avoid sharp minima. The method shows superior performance compared to conventional SGD and previous weight-noise injection methods across various batch sizes and learning rates.
The stochastic gradient descent (SGD) method is most widely used for deep neural network (DNN) training. However, the method does not always converge to a flat minimum of the loss surface that can demonstrate high generalization capability. Weight noise injection has been extensively studied for finding flat minima using the SGD method. We devise a new weight-noise injection-based SGD method that adds symmetrical noises to the DNN weights. The training with symmetrical noise evaluates the loss surface at two adjacent points, by which convergence to sharp minima can be avoided. Fixed-magnitude symmetric noises are added to minimize training instability. The proposed method is compared with the conventional SGD method and previous weight-noise injection algorithms using convolutional neural networks for image classification. Particularly, performance improvements in large batch training are demonstrated. This method shows superior performance compared with conventional SGD and weight-noise injection methods regardless of the batch-size and learning rate scheduling algorithms.