A new local time-decoupled squared Wasserstein-2 method for training stochastic neural networks to reconstruct uncertain parameters in dynamical systems
This addresses parameter uncertainty in dynamical systems modeling, though it appears incremental as it builds on existing Wasserstein and neural network approaches.
The paper tackles the problem of reconstructing uncertain parameter distributions in dynamical systems by proposing a local time-decoupled squared Wasserstein-2 method for training stochastic neural networks, demonstrating effectiveness through numerical examples.
In this work, we propose and analyze a new local time-decoupled squared Wasserstein-2 method for reconstructing the distribution of unknown parameters in dynamical systems. Specifically, we show that a stochastic neural network model, which can be effectively trained by minimizing our proposed local time-decoupled squared Wasserstein-2 loss function, is an effective model for approximating the distribution of uncertain model parameters in dynamical systems. Through several numerical examples, we showcase the effectiveness of our proposed method in reconstructing the distribution of parameters in different dynamical systems.