DFG-Project on Learning Data Structure Behaviour

The Software Technologies Group is pleased to announce that it has been awarded a DFG grant on the topic of "Learning Data Structure Behaviour from Executions of Pointer Programs", which will start in June 2014. Dynamic data structures are a common source of error and complexity in pointer programs and, thus, analyses focusing on these artifacts provide great utility. Our approach aims to combine machine learning and pattern recognition to automatically learn information of interest to developers and maintainers for the purpose of program comprehension and reverse engineering. In addition, this project will study how the learned information may be used to enhance program verification by, e.g., automatically generating verification annotations such as pre-/post-conditions and invariants.

The funds provided by the grant will support both a PhD student, who will focus full-time on this interesting research area, as well as continued collaboration with our partners at the University of York, UK and K.U. Leuven, Belgium. A brief project description can be found on the SWT Research Pages and the full case for support is available from Prof. Lüttgen's Homepage.