講者：Ping-Han Huang博士候選人（School of Mathematical and Statistical Sciences, Arizona State University）
時間：115年1月13日（星期二）11:00
地點：台大次震宇宙館601室
講題：Design of Experiments for Sparse Functional Data.
摘要：Functional measurements are often contaminated with errors or collected at irregularly sampled time points due to practical constraints. A primary focus on design problems has thus revolved around finding the best time points to collect observations from subjects to better facilitate downstream analysis. Following this focus, this talk consists of two parts that discuss two stages of an experiment respectively. The first part focuses on formulating a good pilot-study design to facilitate identifying optimal designs for future data collection and making statistical inference from pilot studies. The second part focuses on developing adaptive designs for subsequent experiments. A Bayesian hierarchical model is constructed with the concept of active learning and a new form of utility function. Simulations show that the designs in these two parts outperform other existing designs. Going forward, this talk will be concluded with possible future connections to artificial intelligence.