Schedule
MINDS SEMINAR
MINDS Seminar Series | Kyongmin Yeo (IBM) - Data-driven Forecast of Random Dynamical System Using Recurrent Generative Adversar
MINDS SEMINAR
period : 2021-10-19 ~ 2021-10-19
time : 10:00:00 ~ 11:00:00
개최 장소 : Online streaming (Zoom)
Topic : Data-driven Forecast of Random Dynamical System Using Recurrent Generative Adversarial Network
개요
| Date | 2021-10-19 ~ 2021-10-19 | Time | 10:00:00 ~ 11:00:00 |
| Speaker | Kyongmin Yeo | Affiliation | IBM |
| Place | Online streaming (Zoom) | Streaming link | ID : 688 896 1076 / PW : 54321 |
| Topic | Data-driven Forecast of Random Dynamical System Using Recurrent Generative Adversarial Network | ||
| Contents | We present a deep learning model for data-driven simulations of random dynamical systems without a distributional assumption. The deep learning model consists of a recurrent neural network, which captures the temporal dynamics, and a generative adversarial network to learn and sample from the probability distribution of the random dynamical system. Although generative adversarial networks provide a powerful tool to model a complex probability distribution, it is well known that the training often fails without a proper regularization. Here, we propose a regularization strategy for a time series problem based on the consistency conditions of the predictive distributions. First, the maximum mean discrepancy (MMD) is used to enforce the consistency between conditional and marginal distributions. Then, the marginal distributions of multiple-step predictions are regularized by using MMD or from a multiple discriminators. The behavior of the proposed model is studied by using three stochastic processes with complex noise structures. | ||
MinDS
·
2021-09-16 10:01 ·
Views 625