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 2602.07800cs.LGcs.NAcs.NEmath.NA
收藏

用深度神经网络和变换器近似矩阵函数

Approximating Matrix Functions with Deep Neural Networks and Transformers

Department of Mathematics and Statistics(数学与统计学系)

Rahul Padmanabhan, Simone Brugiapaglia

AI总结:

本文研究了用深度神经网络和变换器近似矩阵函数,证明了ReLU网络的精度界限,并展示了变换器在近似矩阵函数中的有效性。

摘要:

变换器已革新自然语言处理,但其在数值计算中的应用却较少受到关注。我们研究了使用神经网络(包括变换器)来近似矩阵函数的方法,这些函数将标量函数映射到矩阵。我们专注于将方阵映射到相同维度方阵的函数。这类矩阵函数在科学计算中广泛出现,例如连续时间马尔可夫链中的矩阵指数和动态系统稳定性分析中的矩阵符号函数。在本文中,我们做出了两项贡献。首先,我们证明了使用ReLU网络近似矩阵指数以任意精度所需的宽度和深度的界限。其次,我们通过实验表明,使用适当的数值编码的变换器编码器-解码器可以以5%的相对误差概率近似某些矩阵函数。我们的研究揭示了编码方案对性能有显著影响,不同的方案对不同的函数效果不同。

英文摘要:

Transformers have revolutionized natural language processing, but their use for numerical computation has received less attention. We study the approximation of matrix functions, which map scalar functions to matrices, using neural networks including transformers. We focus on functions mapping square matrices to square matrices of the same dimension. These types of matrix functions appear throughout scientific computing, e.g., the matrix exponential in continuous-time Markov chains and the matrix sign function in stability analysis of dynamical systems. In this paper, we make two contributions. First, we prove bounds on the width and depth of ReLU networks needed to approximate the matrix exponential to an arbitrary precision. Second, we show experimentally that a transformer encoder-decoder with suitable numerical encodings can approximate certain matrix functions at a relative error of 5% with high probability. Our study reveals that the encoding scheme strongly affects performance, with different schemes working better for different functions.

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。

↑