貝氏變分法到參數估算重力波天文學


物理系學生會議:貝氏變分法到參數估算重力波天文學

Talk:- Variational Bayesian to Parameter Estimation in Gravitational Wave Astronomy

The traditional way of conducting parameter estimation in gravitational-wave data analysis is done by obtaining the posterior and evidence through Markov-Chain Monte Carlo methods. These computational methods are time consuming as they rely on wandering randomly in parameter spaces of high dimensions. Instead of randomly sampling, variational Bayesian makes use of an approximated probability distributions, which resembles the true distribution. Minimizing the differences between the true and approximated distributions can then be a more directional way of conducting parameter estimation. In this presentation, we will discuss the possibility of implementing variational Bayesian to parameter estimation in gravitational-wave data analysis. 
 
時間:2017.05.20(六) 14:00 ~ 14:15
地點:香港中文大學 科學館LT4
講者:Adrian, K.W. Chung
語言:英語

#131

此活動由香港中文大學物理系主辦。

發表回覆