Neural network architectures
Neural network architectures
复制标题
神经网络架构
DOI：
10.1201/b10604-9
复制
发表时间：
2001
期刊：
影响因子：
--
通讯作者：
Bogdan M. Wilamowski
中科院分区：
文献类型：
--
作者：
Bogdan M. Wilamowski
神经网络是目前最现代的创造性和有趣的知识领域之一。使用神经网络可以解决许多问题时，使用其他方法几乎是不可能的，由于其复杂性。目前，有大量的神经网络架构可用，包括简单的感知器和具有数万亿参数的神经网络。神经网络架构的多样性取决于它们所面临的任务。此外，神经网络的范围也在不断增长，因为随着数据量和计算能力的增加，其应用的可能性也在不断增加。本文概述了主要的神经网络架构，这些架构是现代神经网络开发的基础，可用于日常生活的许多方面。
Neural networks are currently one of the most modern creative and interesting areas of knowledge. The use of neural networks allows solving many problems when the use of other approaches is almost impossible due to their complexity. Todaty, there are a large number of neural network architectures available, including simple ones like perceptron and neural networks with trillions of parameters. Such a variety of neural network architectures is determined by the tasks they face. In addition, the scope of neural networks is also growing due to the emergence of the possibility of their application in connection with increasing data volumes and computing power. The article provides an overview of the main of neural networks architectures, which serve as the basis for development of the modern neural networks, that can be used in many aspects of everyday life.
可解释的多项式神经常微分方程。
DOI：
10.1063/5.0130803
发表时间：
2023
期刊：
Chaos (Woodbury, N.Y.)
影响因子：
--
作者：
Fronk,Colby;Petzold,Linda
通讯作者：
Petzold,Linda