The Naturalness of Software

Software, like other forms of human expression, tends to be repetitive; this repetition can be captured in statistical models and exploited for various tasks. We lead a large, multi-institution consortium studying the application of computational NLP methods to software code. UC Davis is the lead; our partners include the Language Technologies Institute at Carnegie-Mellon University, and the Software Engineering Group at Iowa State Unviersity. This project has it’s own webpage (under construction)