AxProf is an algorithmic profiling framework for randomized approximate programs. Randomized approximate programs are used in multiple fields, such as big data analysis and approximate hardware simulation. Reasoning about the correctness of such programs requires the proper usage of appropriate statistical tests. AxProf accepts an accuracy specification as a formula in a mathematical notation and generates statistical reasoning code. See how to use AxProf in our tutorial.
AxProf is implemented in Python 3. It is distributed on GitHub under the Apache-2.0 license.View on Github
- Statistical Algorithmic Profiling for Randomized Approximate Programs
Keyur Joshi, Vimuth Fernando, Sasa Misailovic
41st ACM/IEEE International Conference on Software Engineering (ICSE 2019)
Full Text | Presentation