|
Abstract:
|
Deep generative model such as variational auto-encoder or generative adversarial networks have received much attention recently for modeling complex high-dimensional data such as images. Statistically, deep generative models are tools to model complex nonlinear dependencies of high dimensional random vectors. In this talk, I will talk about using deep generative model for modeling dependencies in financial time series. In particular, we consider how to modify standard deep generative models for time series data which are collected sequentially.
|
