Stochastic Modelling, Big Data and Deep Learning

发布者:文明办发布时间:2024-06-17浏览次数:101

主讲人:毛学荣 英国Strathclyde大学数学与统计系教授


时间:2024年6月21日14:00


地点:三号楼332室


举办单位:数理学院


主讲人介绍:毛学荣,英国Strathclyde大学数学与统计系教授、1969统计学首席教授、爱丁堡皇家学会(即苏格兰皇家学院)院士、“英国沃弗森研究功勋奖”获得者。


内容介绍:In this talk we will explain how the ordinary differential equations (ODEs) are not enough to model the underlying stochastic quantity and why stochastic differential equations (SDEs) appear naturally. Several well-known SDE models will be presented including the Nobel prize winning model in finance, stochastic SIS epidemic model, stochastic Lotka-Volterra model in population dynamics. We will then explain how SDE models differ significantly from ODE models and reveal the crucial role of noise. We will then emphasise that the use of SDE models depend on the estimation of system parameters. In the case when the model has only a few parameters, we show how they can be estimated by the classical statistical methods, e.g., the least-square method; while when there are lots of parameters, we will show how the deep learning plays its crucial role.