On the Analysis of Ellipsoid Approximation of Nested Sampling
Add to your list(s)
Download to your calendar using vCal
If you have a question about this talk, please contact Ryan Prescott Adams.
In 2004, Skilling proposed an efficient monte carlo method, called nested sampling, to numerically calculate the evidence in model comparison tasks. One important technical problem is that, in each iteration, we have to perform a uniform sampling under a hard constrain.
In 2006, Mukherjee, Liddle and Parkinson proposed the method called an “ellipsoid approximation” to solve this technical problem. However, a new problem, arising from the ellipsoid method itself, is to select its parameter called the “expansion factor”.
This talk shall breifly give some theoretical and experimental results based on random matrix theory about how to select the appropriate parameter. This is an in-progress work.
This talk is part of the Inference Group series.
This talk is included in these lists:
Note that ex-directory lists are not shown.
|