arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2108.05643cs.LGmath.STstat.TH
收藏

浅ReLU网络的极小表示

On minimal representations of shallow ReLU networks

S. Dereich, S. Kassing

摘要:

浅层ReLU网络的实现函数是连续的分段仿射函数$f:\mathbb R^d\to \mathbb R$,其中域$\mathbb R^{d}$被一组$n$超平面划分为$f$在其上仿射的单元。我们发现,$f$的最小表示使用$n$,$n+1$或$n+2$神经元,我们的特点,每三种情况下。在输入层是一维的特定情况下,最小表示总是最多使用$n+1$神经元,但在所有更高维的设置中,存在需要$n+2$神经元的函数。然后证明了表示$f$的极小网络的集合构成$C^\infty$ -子流形$M$,并导出了$M$的连通分支的维数和个数.此外,我们给出了超平面的一个标准,该标准保证所有连续的分段仿射函数都是适当的ReLU网络的实现函数。

英文摘要:

The realization function of a shallow ReLU network is a continuous and piecewise affine function $f:\mathbb R^d\to \mathbb R$, where the domain $\mathbb R^{d}$ is partitioned by a set of $n$ hyperplanes into cells on which $f$ is affine. We show that the minimal representation for $f$ uses either $n$, $n+1$ or $n+2$ neurons and we characterize each of the three cases. In the particular case, where the input layer is one-dimensional, minimal representations always use at most $n+1$ neurons but in all higher dimensional settings there are functions for which $n+2$ neurons are needed. Then we show that the set of minimal networks representing $f$ forms a $C^\infty$-submanifold $M$ and we derive the dimension and the number of connected components of $M$. Additionally, we give a criterion for the hyperplanes that guarantees that all continuous, piecewise affine functions are realization functions of appropriate ReLU networks.

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。

↑