Facebook funds project to fix integrity and privacy issues
University of Adelaide researchers are pushing the boundaries of bug detection through social testing as Facebook develops a cyber-cyber digital twin of its social media platform called WW.
WW is a Web Enabled Simulation (WES), in which user communities and their interactions and behaviours are simulated using machine learning, while the real platform infrastructure is both directly executed by the simulation and also is, itself, simulated in various off-line and synthetic versions.
Associate Professor Markus Wagner and Dr Christoph Treude from the School of Computer Science, are one of only five projects that have been successful in receiving USD$92,784 funding for their research titled "Socialz – Multi-Objective Automated Social Fuzz Testing" which will simulate the interactions of a large community of users.
Their work comprises three activities:
- capturing real-world user interactions at an abstract level
- evolution of diverse bots (including those with atypical behaviour), and
- injection of invalid, unexpected, and random behaviour into the created bots for the purpose of social fuzz testing of the system with respect to multiple objectives.
Christoph and Markus (both ARC DECRA holders) have been long-term collaborators in the greater space of automated software engineering. While Christoph will soon move on to another university, they will conduct this project together and at the University of Adelaide.