arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2604.23765cs.LGcs.NEmath.FA
收藏

Kolmogorov-Arnold 网络普适性的充要条件

Necessary and sufficient conditions for universality of Kolmogorov-Arnold networks

vugar E. Ismailov

Vugar Ismailov

AI总结:

本文分析了 Kolmogorov-Arnold 网络(KAN)的普适逼近性质,证明了当所有边缘函数为仿射时普适性不成立,而添加一个非仿射函数即可恢复普适性,并给出了深层和两层 KAN 的充要条件。

摘要:

我们从边缘函数的角度分析了 Kolmogorov-Arnold 网络(KAN)的普适逼近性质。如果这些函数都是仿射的,那么普适性显然不成立。除了仿射函数之外,还需要多少非仿射函数才能保证普适性?我们证明一个就足够了。更精确地说,我们证明对于每个紧集 $K\subset\mathbb{R}^n$,所有边缘函数要么是仿射的,要么等于一个固定的连续函数 $\sigma$ 的深层 KAN 在 $C(K)$ 中稠密,当且仅当 $\sigma$ 是非仿射的。相比之下,对于恰好有两个隐藏层的 KAN,普适性成立当且仅当 $\sigma$ 是非多项式的。我们进一步证明,并不需要完整的仿射函数类;它可以用一个有限集代替而不影响普适性。特别地,在非多项式情况下,当深度任意时,一个由五个仿射函数组成的固定族就足够了。更一般地,对于每个连续的非仿射函数 $\sigma$,存在一个有限的仿射族 $A_\sigma$,使得边缘函数在 $A_\sigma\cup\{\sigma\}$ 中的深层 KAN 仍然是普适的。我们还证明,采用 Liu 等人~\cite{Liu2024} 引入的基于样条的边缘参数化的 KAN 是经典意义上的普适逼近器,即使样条次数和节点序列是预先固定的。

英文摘要:

We analyze the universal approximation property of Kolmogorov-Arnold Networks (KANs) in terms of their edge functions. If these functions are all affine, then universality clearly fails. How many non-affine functions are needed, in addition to affine ones, to ensure universality? We show that a single one suffices. More precisely, we prove that deep KANs in which all edge functions are either affine or equal to a fixed continuous function $σ$ are dense in $C(K)$ for every compact set $K\subset\mathbb{R}^n$ if and only if $σ$ is non-affine. In contrast, for KANs with exactly two hidden layers, universality holds if and only if $σ$ is nonpolynomial. We further show that the full class of affine functions is not required; it can be replaced by a finite set without affecting universality. In particular, in the nonpolynomial case, a fixed family of five affine functions suffices when the depth is arbitrary. More generally, for every continuous non-affine function $σ$, there exists a finite affine family $A_σ$ such that deep KANs with edge functions in $A_σ\cup\{σ\}$ remain universal. We also prove that KANs with the spline-based edge parameterization introduced by Liu et al.~\cite{Liu2024} are universal approximators in the classical sense, even when the spline degree and knot sequence are fixed in advance.

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。

↑