What is it about?
Several computer applications consume vast amounts of random numbers, and we showcase how to produce these incredibly quickly, and offer huge speed improvements without compromising accuracy. This is achieved through a combination of mathematical shortcuts and programming tricks.
Featured Image
Photo by Markus Winkler on Unsplash
Why is it important?
Many fields rely on generating random numbers, covering everything from weather forecasting to pricing financial products. Our findings show how we can speedup these applications by orders of magnitude without compromising the final accuracy, by utilising random numbers which are incredibly quick to produce. This can be used to improve weather forecasting accuracy, or understand risk when constructing an investment portfolio.
Perspectives
In our work we outline a mathematical and computational recipe for arriving at the result of a simulation much quicker without compromising on the accuracy. We do this for an "easy" and hugely common family of random numbers, and another considerably more "difficult" family, showing the same recipe yields huge speed improvements in both scenarios. We hope this recipe can be widely adopted by practitioners to speed up their applications.
Oliver Sheridan-Methven
Read the Original
This page is a summary of: Approximating inverse cumulative distribution functions to produce approximate random variables, ACM Transactions on Mathematical Software, June 2023, ACM (Association for Computing Machinery),
DOI: 10.1145/3604935.
You can read the full text:
Contributors
The following have contributed to this page