Dr. Jeff Huang, associate professor in the Department of Computer Science and Engineering at Texas A&M University, was recently honored with a Young Faculty Award (YFA) from the Defense Advanced Research Projects Agency (DARPA).
The YFA program aims to identify and engage rising stars in junior research positions and expose them to the needs of the Department of Defense and DARPA’s program development process.
Huang received the award for his project titled “SmartScript: A Learning-based Approach to Static Type and Semantic Inconsistency Analysis of Dynamic Scripts.”
“The goal of this project is to develop SmartScript, a novel static code analysis that automatically detects type and semantic inconsistency errors in dynamic scripts, such as shell scripts, Python, LUA, Javascript, etc,” Huang said.
To achieve the goal, Huang created a machine-learning technique that leverages inconsistency interference and valuable natural language information in the code such as variable names to discover hidden semantic errors.
“The novel insight is that code semantics conveyed through natural language information are typically ignored by rule-based machines, but can be spotted easily by human developers,” Huang said. “By learning such nonstructural semantics, SmartScript can train a classifier with human intelligence to detect hidden semantic errors significantly beyond the power of existing tools.”
Additionally, the detected inconsistency errors will be explained with high signals whenever an inconsistency is identified.
“We anticipate that the tool and techniques resulting from this research will transition to industry and enable millions of developers to continuously reveal hidden bugs early in the software development cycle,” Huang said.
Huang is also collaborating with Dr. Ruihong Huang, assistant professor in the Department of Computer Science and Engineering and an expert in natural language processing (NLP), to create advanced NLP techniques for this project.