A decade-old Android testing paper takes ISSTA's impact honor
Ke Mao, Mark Harman and Yue Jia have received the ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA) 2026 Most Influential Paper Award for their 2016 work, “Sapienz: multi-objective automated testing for Android applications.” The research originated at University College London and later became testing infrastructure at Meta.
Sapienz searches Android apps for crashes automatically and shrinks the reproduction path so engineers can investigate failures with fewer steps. Its original design pursued three objectives at once: covering more of an app, uncovering more crashes, and shortening the sequences that trigger them.
In its evaluation on 1,000 popular Google Play apps, Sapienz found 558 previously unknown crashes.
From paper to daily runs at Meta
Deployment at Meta followed within 17 months of publication. The system now executes millions of tests per day across Meta's platforms, reports thousands of unique crashes weekly, and covers hundreds of app releases every month as part of the company's reliability and quality work.
Crash detection was only the starting point. Using test accounts holding synthetic conversations, Sapienz also served as the app-exploration engine behind dynamic taint analysis at WhatsApp.
What the keynote says about where testing goes next
Ke Mao presented Sapienz's trajectory in an ISSTA 2026 keynote titled “Finding the Bugs Users Would Find: From Sapienz to Autonomous Agents.” The talk points to LLM agents that go past crash discovery: they can explore an app and judge whether its behavior lines up with user expectations, stretching automated testing from reliability toward functional correctness and, increasingly, automated repair. Multi-objective exploration and actionable fault revelation — Sapienz's founding ideas — transfer directly to that setting, and a decade later they remain valid.
Meta frames the work as part of its broader privacy and risk management program, which aims to place privacy considerations at the center of product development, surface risks earlier in the lifecycle, apply safeguards more consistently, and build protections into new innovations from the outset.



