.png){kind=icon}
Return
Neural Hyper Flow: Parameterized prediction for control systems
.png){kind=icon}
PRE
AI
DOI：10.1016/j.engappai.2025.110413.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Neural operators for systems of differential equations have been used as surrogates instead of potentially computationally expensive solution algorithms. However, the application of these operators is unsuitable as a model for control loop design. The main reason is that most operators consider a specific initial condition for each training. In particular, the control application needs to update its initial condition and recompute the control law. This paper aims to provide a novel description of the neural operator that could be seamlessly applied as a surrogate with different initial conditions. By utilizing hypernetworks and selecting appropriate parameters, this work can be applied to a future control application.
Keywords:
Operator
Neural Operator
Ordinary differential equation
Control design
Hypernetworks
Journal
IF:
8
Papers:
5.8K
Citations:
3.5W
Organization
Cited Papers
ShareSaveShareSaveShareSaveShareSaveA review of uncertainty quantification in deep learning: Techniques, applications and challenges.png)
INFORMATION FUSION
OAAI
INFORMATION FUSIONIF15.5
OAAIShareSaveLearning nonlinear operators via DeepONet based on the universal approximation theorem of operatorsOAAI
OAAIShareSave
View more