arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.03084cs.LGcs.AI
收藏

关于Transformer在Maxout网络和连续分段线性函数中的表达能力

On the Expressive Power of Transformers for Maxout Networks and Continuous Piecewise Linear Functions

School of Mathematics, Sun Yat-sen University(中山大学数学学院);School of Information Science, Guangdong University of Finance and Economics(广东财经大学信息学院)

Linyan Gu, Lihua Yang, Feng Zhou

AI总结:

本文研究了Transformer在Maxout网络和连续分段线性函数中的表达能力,通过分析线性区域数量揭示了Transformer的理论表达性。

摘要:

Transformer网络在广泛的应用中取得了显著的实证成功,但其理论表达能力仍不够充分。在本文中,我们研究了Transformer架构的表达能力。我们首先通过Transformer网络显式地近似Maxout网络,同时保持相似的模型复杂度。作为结果,Transformers继承了ReLU网络的通用近似能力,只要在相似的复杂度约束下。基于这一联系,我们开发了一个框架来分析Transformer对连续分段线性函数的近似能力,并通过线性区域的数量来定量表征其表达性,该数量随着深度呈指数增长。我们的分析建立了标准前馈神经网络近似理论与Transformer架构之间的理论桥梁。它还提供了对Transformers的结构洞察:自注意力层实现max型操作,而前馈层实现token-wise的仿射变换。

英文摘要:

Transformer networks have achieved remarkable empirical success across a wide range of applications, yet their theoretical expressive power remains insufficiently understood. In this paper, we study the expressive capabilities of Transformer architectures. We first establish an explicit approximation of maxout networks by Transformer networks while preserving comparable model complexity. As a consequence, Transformers inherit the universal approximation capability of ReLU networks under similar complexity constraints. Building on this connection, we develop a framework to analyze the approximation of continuous piecewise linear functions by Transformers and quantitatively characterize their expressivity via the number of linear regions, which grows exponentially with depth. Our analysis establishes a theoretical bridge between approximation theory for standard feedforward neural networks and Transformer architectures. It also yields structural insights into Transformers: self-attention layers implement max-type operations, while feedforward layers realize token-wise affine transformations.

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。

↑