Skip to content

Congratulations to Shirley on her work on predicting path satisfiability

We’ve been working on Android app analysis using symbolic execution for a number of years now. One thing we’ve noticed is that many static paths in Android applications turn out to be infeasible — that is no satsifying inputs can ever execute the path. Spending analysis time on such paths is a complete waste of resources. In her forthcoming paper, Accelerating symbolic analysis for Android apps, Shirley describes how machine learning can help predict paths orders of magnitude faster than solving for satisfying inputs, reducing the number of paths to analyze by 51% and analysis time by 14%.