arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.19221cs.LGcs.CVcs.NE
收藏

Cauchy激活函数与XNet

Cauchy activation function and XNet

Northwestern University(西北大学);Oxford Suzhou Centre for Advanced Research(牛津苏州高等研究院);Great Bay University(大湾区大学);University of Massachusetts Amherst(马萨诸塞大学阿默斯特分校)

Xin Li, Zhihong Xia, Hongkun Zhang

AI总结:

本文提出源自Cauchy积分定理的Cauchy激活函数,并构建新型神经网络XNet,在图像分类与求解偏微分方程等高维任务上显著优于既有基准和PINNs。

摘要:

我们开发了一种名为Cauchy激活函数的新型激活函数。该函数源自复分析中的Cauchy积分定理,专为需要高精度的问题而设计。这一创新催生了一类新型神经网络,我们称之为(Comple)XNet或简称XNet。我们将证明XNet对图像分类和求解偏微分方程(PDEs)等高维挑战特别有效。我们的评估表明,在计算机视觉领域,XNet显著优于MNIST和CIFAR-10等既有基准,并且在低维和高维PDE场景中,相比物理信息神经网络(PINNs)具有显著优势。

英文摘要:

We have developed a novel activation function, named the Cauchy Activation Function. This function is derived from the Cauchy Integral Theorem in complex analysis and is specifically tailored for problems requiring high precision. This innovation has led to the creation of a new class of neural networks, which we call (Comple)XNet, or simply XNet. We will demonstrate that XNet is particularly effective for high-dimensional challenges such as image classification and solving Partial Differential Equations (PDEs). Our evaluations show that XNet significantly outperforms established benchmarks like MNIST and CIFAR-10 in computer vision, and offers substantial advantages over Physics-Informed Neural Networks (PINNs) in both low-dimensional and high-dimensional PDE scenarios.

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。

↑