A common computational challenge in data science is the estimation of the expected value some probabilistic process. This process may represent a physical phenomena (from biochemistry to financial asset pricing) or probabilities may represent the level of uncertainty in some model parameter to be estimated.
In this presentation, Dr David Warne will discuss the multilevel Monte Carlo approach to computing expectations that utilises a clever telescoping summation of approximations to obtain accurate estimates that can be orders of magnitude faster to compute than standard approaches.
About the presenter
Dr David Warne is a postdoctoral Research Fellow in the Centre for Data Science at QUT. He develops mathematical models and statistical methodologies to tackle challenging real work problems involving interacting populations, especially in ecology, biology and epidemiology. He also utilises his background in high performance computing to get to most out of available computational resources.
Event recording:
Resources:
Details:
| Start Date: | 13/05/2021 [add to calendar] |
| Start Time: | 2pm (AEST) |