Variational Training of Neural Network Approximations of Solution Maps for Physical Models
发布者：程梦琴发布时间：2019-06-13浏览次数：120
| Speaker: | Yingzhou Li, Duke University |
| Host: | Rujun Jiang, School of Data Science, Fudan University |
| Time: | 10:00-11:00, June 13, 2019 |
| Location: | Room S301, Zibin Building, Fudan University |
| Abstract: | A novel solve-training framework is proposed to train neural network in representing low dimensional solution maps of physical models. Solve-training framework uses the neural network as the ansatz of the solution map and train the network variationally via loss functions from the underlying physical models. Solve-training framework avoids expensive data preparation in the traditional supervised training procedure, which prepares labels for input data, and still achieves effective representation of the solution map adapted to the input data distribution. The efficiency of solve-training framework is demonstrated through obtaining solutions maps for linear and nonlinear elliptic equations, and maps from potentials to ground states of linear and nonlinear Schrodinger equations. |
| Bio: | Yingzhou Li is currently working as Phillip Griffiths research assistant professor at Duke University. Before joining Duke, Yingzhou received his Ph.D. degree from Institute for Computational and Mathematical Engineering at Stanford University in 2017. He specializes in scientific computing and numerical analysis. In particular, his research concerns the design of fast numerical algorithms for problems in computational physics, computational chemistry, and scientific computing. |
