Speaker: Prof. Chii-Ruey Hwang(Academia Sinica)
Title: Monte Carlo Markov Processes
Time: 2016-04-07 (Thu.)  15:00 - 16:00
Place: Auditorium, 6 Floor, Institute of Mathematics (NTU Campus)
Abstract: Let $\pi$ be an underlying probability (density) known up to a norming constant, Markov processes (chains), regarded as 'conceptual algorithms', may be used to approximate sample from $\pi$. The evaluation of the approximation is investigated under various comparison criteria, e.g. asymptotic variance, spectral gap, convergence exponent in variational norm. The worst-case analysis, the average-case analysis, uniform comparison, antisymmetric perturbations are considered. Some related problems in analysis and geometry are discussed. This is a survey of our work.
  || Close window ||