Wednesday, July 1, 2015
Speaker: Thomas Pani
Venue: IST Austria
Starts right after Thomas Ferrère’s talk
In this talk, we consider empirical metrics for software source code, which can predict the performance of verification tools on specific types of software. Our metrics comprise variable usage patterns, loop patterns, as well as indicators of control-flow complexity and are extracted by simple data-flow analyses. We demonstrate that our metrics are powerful enough to devise a machine-learning based portfolio solver for software verification. Our experiments show that this portfolio solver would be the (hypothetical) overall winner of both the 2014 and 2015 International Competition on Software Verification (SV-COMP), which gives strong empirical evidence for the predictive power of our metrics.