SoftTarget Regularization: An Effective Technique to Reduce Over-Fitting in Neural Networks
This addresses over-fitting for neural network practitioners, offering an incremental improvement over existing regularization techniques.
The paper tackles over-fitting in deep neural networks by introducing SoftTarget regularization, which adjusts labels using a weighted average of real labels and past soft-targets, achieving performance comparable to Dropout without reducing model capacity.
Deep neural networks are learning models with a very high capacity and therefore prone to over-fitting. Many regularization techniques such as Dropout, DropConnect, and weight decay all attempt to solve the problem of over-fitting by reducing the capacity of their respective models (Srivastava et al., 2014), (Wan et al., 2013), (Krogh & Hertz, 1992). In this paper we introduce a new form of regularization that guides the learning problem in a way that reduces over-fitting without sacrificing the capacity of the model. The mistakes that models make in early stages of training carry information about the learning problem. By adjusting the labels of the current epoch of training through a weighted average of the real labels, and an exponential average of the past soft-targets we achieved a regularization scheme as powerful as Dropout without necessarily reducing the capacity of the model, and simplified the complexity of the learning problem. SoftTarget regularization proved to be an effective tool in various neural network architectures.