Date: Friday, Nov 20, 2020 Time : 5:00 pm - 6:00 pm (Indian Standard Time) Join Zoom Meeting https://us02web.zoom.us/j/82286271984?pwd=VHVUM2pFVFExU2owaEhXU3lGZSs3UT09 Meeting ID: 822 8627 1984 Passcode: 455327 Scalable Monte Carlo algorithms for Bayesian inference Deborshee Sen Postdoctoral Associate, SAMSI, and Department of Statistical Science, Duke University, USA. 20-11-20 Abstract The Bayesian framework is an appealing technique for conducting statistical inference with accurate uncertainty quantification. Numerical methods are routinely deployed since analytical solutions are often not available. Among them, Monte Carlo algorithms are arguably the most popular approach. However, such algorithms become computationally challenging as the size of the data increases. This has motivated significant recent attention to the development and study of scalable Monte Carlo algorithms for Bayesian inference. In this talk, I will first give a background to the problem of Bayesian inference for large data. I will then discuss scalable algorithms for inference, including my own contributions to the area.
|