arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2010.04596cs.LGmath.STstat.TH
收藏

带sigmoid激活函数的深度神经网络逼近光滑函数

Approximating smooth functions by deep neural networks with sigmoid activation function

Technische Universität Darmstadt(达姆施塔特工业大学)

Sophie Langer

摘要:

我们研究了具有sigmoid激活函数的深度神经网络(DNN)的能力。最近,研究表明,DNN近似于紧致集上的任何$d$维光滑函数,其阶数为$W^{-p/d}$,其中$W$是网络中非零权重的数量,$p$是函数的光滑度。不幸的是,这些比率仅适用于一类特殊的稀疏连接DNN。我们问自己,对于更简单和更一般的类,我们是否可以显示相同的近似率,即,DNN仅由其宽度和深度定义。在这篇文章中,我们证明了具有固定深度和$M^d$阶宽度的DNN实现了$M^{-2p}$的近似率。作为结论,我们定量地描述了DNN的逼近能力在网络中的总权重$W_0$方面,并显示了$W_0^{-p/d}$的逼近率。这个更一般的结果最终帮助我们理解哪种网络拓扑保证了特定的目标精度。

英文摘要:

We study the power of deep neural networks (DNNs) with sigmoid activation function. Recently, it was shown that DNNs approximate any $d$-dimensional, smooth function on a compact set with a rate of order $W^{-p/d}$, where $W$ is the number of nonzero weights in the network and $p$ is the smoothness of the function. Unfortunately, these rates only hold for a special class of sparsely connected DNNs. We ask ourselves if we can show the same approximation rate for a simpler and more general class, i.e., DNNs which are only defined by its width and depth. In this article we show that DNNs with fixed depth and a width of order $M^d$ achieve an approximation rate of $M^{-2p}$. As a conclusion we quantitatively characterize the approximation power of DNNs in terms of the overall weights $W_0$ in the network and show an approximation rate of $W_0^{-p/d}$. This more general result finally helps us to understand which network topology guarantees a special target accuracy.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑