In this paper, we develop an alternating direction method of multipliers (ADMM) for deep neural networks training with sigmoid-type activation functions (called \textit{sigmoid-ADMM pair}), mainly motivated by the gradient-free nature of ADMM in avoiding the saturation of sigmoid-type activations and the advantages of deep neural networks with sigmoid-type activations (called deep sigmoid nets) over their rectified linear unit (ReLU) counterparts (called deep ReLU nets) in terms of approximation. In particular, we prove that the approximation capability of deep sigmoid nets is not worse than that of deep ReLU nets by showing that ReLU activation function can be well approximated by deep sigmoid nets with two hidden layers and finitely many free parameters but not vice-verse. We also establish the global convergence of the proposed ADMM for the nonlinearly constrained formulation of the deep sigmoid nets training from arbitrary initial points to a Karush-Kuhn-Tucker (KKT) point at a rate of order ${\cal O}(1/k)$. Besides sigmoid activation, such a convergence theorem holds for a general class of smooth activations. Compared with the widely used stochastic gradient descent (SGD) algorithm for the deep ReLU nets training (called ReLU-SGD pair), the proposed sigmoid-ADMM pair is practically stable with respect to the algorithmic hyperparameters including the learning rate, initial schemes and the pro-processing of the input data. Moreover, we find that to approximate and learn simple but important functions the proposed sigmoid-ADMM pair numerically outperforms the ReLU-SGD pair.
翻译:在本文中,我们为深神经网络的培训开发了一个交替方向的倍数(ADMM)方法,该方法主要出于ADM的无梯度性质,以避免模拟类型激活的饱和性,以及深神经网络的优点,即具有类类激活(所谓的深类网),而不是经纠正的线性单元(RELU)(所谓的深重ReLU网),在近似方面。特别是,我们证明深类网络的近距离能力并不比深类网络更差(称为\ textit{sigmod-AD 配对对),其动机主要是由于ADMD的无梯度性质,而避免了Sigmod型激活的渐渐变性,我们还建立了拟议的ADMU网络的全球性趋同性组合,从任意的初始点到Karush-Kuhn-Tuck(KTral-ral-ral-ral-ral-rational-ral-ral-ral-ral-ral-ral-ral-ral-ral-ral-ral-ral-lation slation-ral-ral-ral-lation slation slation-lation-lation-lation-lation-lational-lational-lational-lational-leval-lational-l),我们证明,我们证明, 和Sl-sal-sal-sal-leval-leval-lation-lation-lation-lation-lation-lation-lation-lation-lational-ld-ld-de-ldal-ld-ldal-ld-ld-ld-ldal-ld-ldal-ld-ld-ld-ld-ldal-ldal-ldal-ld-ldal-ldal-ldal-ld-ld-ld-ldal-ld-ld-ldal-ld-ld-ld-ld-ld-ld-ld-ld-ldal-ld-ld-l化,用于G)