A5056594196- Model Reduction and Neural Networks
- Fluid Dynamics and Turbulent Flows
- Lattice Boltzmann Simulation Studies
- Geophysical Methods and Applications
- Probabilistic and Robust Engineering Design
- Privacy-Preserving Technologies in Data
- Nuclear Engineering Thermal-Hydraulics
- Control Systems and Identification
- Adversarial Robustness in Machine Learning
University of Stavanger
2023-2024
Gradient inversion attacks can reconstruct the victim's private data once they have access to model and gradient. However, existing research is still immature, many are conducted in ideal conditions. It unclear how damaging such really be effectively defended. In this paper, we first summarize current relevant researches their limitations. Then design a general gradient attack framework, which both FedSGD FedAVG. We propose approaches enhance label inference image restoration, respectively....
Machine learning methods have in various ways emerged as a useful tool for modeling the dynamics of physical systems context partial differential equations (PDEs). Nonlinear conservation laws (NCLs) form ut+f(u)x=0 play vital role within family PDEs. A main challenge with NCLs is that solutions contain discontinuities. That is, one or several jumps (uL(t),uR(t)) uL≠uR may move space and time such information about f(u) interval associated this jump not present observation data. Moreover,...
In contemporary research, neural networks are being used to derive Ordinary Differential Equations (ODEs) from observations. However, parameterized ODEs pose a more significant challenge than non-parameterized since the required understand roles of parameters, i.e., structure equations. This paper proposes novel approach by combining Symbolic Neural Network (S-Net) with ODE Solver solve this issue. First, S-Net learns and then predicts dynamics based on new parameters initial states. To...
Abstract In recent years, there has been an increasing interest in utilizing deep learning-based techniques to predict solutions various partial differential equations. this study, we investigate the identification of unknown flux function and diffusion coefficient a one-dimensional convection-diffusion equation. The is allowed vanish on intervals implying that generally possess low regularity, i.e., are discontinuous. Therefore, must be interpreted sense entropy which combine weak...