arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2307.07810cs.LGmath.COmath.RTstat.ML
收藏

图自同构群等变神经网络

Graph Automorphism Group Equivariant Neural Networks

Imperial College London(帝国理工学院)

Edward Pearce-Crump, William J. Knottenbelt

AI总结:

本文针对图数据,提出构建对自同构群等变的神经网络,通过刻画层间线性等变函数并找到生成矩阵集,利用Frucht定理表明该方法适用于任意有限对称群。

摘要:

排列等变神经网络通常用于学习位于图上的数据。然而,对于任何具有 $n$ 个顶点的图 $G$,使用对称群 $S_n$ 作为其对称群并未考虑顶点之间存在的关联。鉴于实际的对称群是自同构群 Aut$(G)$,我们展示了如何通过获得层间可学习的、线性的、Aut$(G)$-等变函数的完整刻画来构建对 Aut$(G)$ 等变的神经网络,其中层间函数是 $\mathbb{R}^{n}$ 的某个张量幂。特别地,我们在 $\mathbb{R}^{n}$ 的标准基下找到了这些层函数的生成矩阵集。这一结果对于从对称群为有限群的数据中学习具有重要意义,因为 Frucht (1938) 的一个定理表明,任何有限群都同构于某个图的自同构群。

英文摘要:

Permutation equivariant neural networks are typically used to learn from data that lives on a graph. However, for any graph $G$ that has $n$ vertices, using the symmetric group $S_n$ as its group of symmetries does not take into account the relations that exist between the vertices. Given that the actual group of symmetries is the automorphism group Aut$(G)$, we show how to construct neural networks that are equivariant to Aut$(G)$ by obtaining a full characterisation of the learnable, linear, Aut$(G)$-equivariant functions between layers that are some tensor power of $\mathbb{R}^{n}$. In particular, we find a spanning set of matrices for these layer functions in the standard basis of $\mathbb{R}^{n}$. This result has important consequences for learning from data whose group of symmetries is a finite group because a theorem by Frucht (1938) showed that any finite group is isomorphic to the automorphism group of a graph.

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。

↑