新聞標題： ( 2017-12-27 )
演講主題：Diffuse to fuse sensors
摘要內容：Understanding the intrinsic structure from a given massive dataset, which is often nonlinear and complex, is a common challenge shared in almost all scientific fields, and the problem is becoming more challenging when the data are from multiple sensors with heterogenous data types. The diffusion geometry is a flexible framework that has led to several convincing results with solid theoretical backup. We will discuss how to apply the diffusion geometry, particularly the common manifold model and alternating diffusion, to deal with the sensor fusion problem. Its application to the sleep dynamics analysis will be shown.