Automatic Software Bug Detection and Fixing by Learning from Large Code Examples

This project seeks to build the first system to automatically detect and fix software bugs by learning from massive code examples. It will do so by combining natural language processing and compiler-based code analysis technologies to first extract the past development knowledge by mining open-sourced projects, and then use the learned knowledge to automatically detect and repair bugs for new programs. This project will create new methods, analyses, techniques and tools for an exciting vision that software development is highly automated. If successful, this project can free millions of software developers from the time-consuming, error-prone process of software debugging in the long-term. 

If your application is successful you will join the Distributed Systems and Services research group at the School of Computing at University of Leeds under the supervision of Dr Zheng Wang.