Lizuo Liu: Multiscale DNN for Stationary Navier Stokes Equations with Oscillatory Solutions

In this talk, we develop new multi-scale deep neural network to compute oscillatory flows for stationary Navier-Stokes equation in complex domains. The multiscale neural network is the structure that can convert the high frequency components in target to low frequency components thus accelerate the convergence of training of neural network. Navier-Stokes flow with high frequency components in 2-D domain are learned by the multiscale deep neural network. The results show that the new multiscale deep neural network can be trained fast and accurate.