.png){kind=icon}
返回
Fractional deep neural network via constrained optimization
.png){kind=icon}
OA
AI
DOI：10.1088/2632-2153/aba8e7.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
This paper introduces a novel algorithmic framework for a deep neural network (DNN), which in a mathematically rigorous manner, allows us to incorporate history (or memory) into the network-it ensures all layers are connected to one another. This DNN, called Fractional-DNN, can be viewed as a time-discretization of a fractional in time non-linear ordinary differential equation (ODE). The learning problem then is a minimization problem subject to that fractional ODE as constraints. We emphasize that an analogy between the existing DNN and ODEs, with standard time derivative, is well-known by now. The focus of our work is the Fractional-DNN. Using the Lagrangian approach, we provide a derivation of the backward propagation and the design equations. We test our network on several datasets for classification problems. Fractional-DNN offers various advantages over the existing DNN. The key benefits are a significant improvement to the vanishing gradient issue due to the memory effect, and better handling of nonsmooth data due to the network's ability to approximate non-smooth functions.
Keyword:
deep learning
deep neural network
fractional time derivatives
constrained optimization
fractional neural network
期刊
M
IF:
4.6
论文数:
1.1K
被引数:
3.4K
机构
引用论文
Differential effects of plant root systems on nickel, copper and silver bioavailability in contaminated soil.png)
Chemosphere
ChemosphereIF0
PREAI
分享收藏Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations物理信息神经网络: 一种用于解决涉及非线性偏微分方程的正反问题的深度学习框架
OAAI
OAAI分享收藏
查看更多内容