Contrastive Weight Regularization for Large Minibatch SGD
This paper offers an incremental improvement to the generalization problem for researchers and practitioners using large-batch SGD in deep learning.
This paper addresses the poor generalization of large-batch SGD by introducing Distinctive Regularization (DReg), which replicates a network layer and encourages parameter diversity. This method significantly boosts convergence and improves generalization performance for large-batch SGD, including with momentum.
The minibatch stochastic gradient descent method (SGD) is widely applied in deep learning due to its efficiency and scalability that enable training deep networks with a large volume of data. Particularly in the distributed setting, SGD is usually applied with large batch size. However, as opposed to small-batch SGD, neural network models trained with large-batch SGD can hardly generalize well, i.e., the validation accuracy is low. In this work, we introduce a novel regularization technique, namely distinctive regularization (DReg), which replicates a certain layer of the deep network and encourages the parameters of both layers to be diverse. The DReg technique introduces very little computation overhead. Moreover, we empirically show that optimizing the neural network with DReg using large-batch SGD achieves a significant boost in the convergence and improved generalization performance. We also demonstrate that DReg can boost the convergence of large-batch SGD with momentum. We believe that DReg can be used as a simple regularization trick to accelerate large-batch training in deep learning.