Support
During my fellowship, I focused on the design and analysis of regular expressions (regex), which are symbolic notations used for defining search patterns in text. While widely used in programming and data validation, poorly written regex can introduce serious performance vulnerabilities, particularly Regular Expression Denial of Service (ReDoS) attacks. These occur when specific input patterns cause excessive backtracking and resource exhaustion.
My primary goal was to identify a general methodology for writing regex that balances expressiveness with safety and maintainability. I analyzed common ReDoS-prone patterns and developed strategies to rewrite them into functionally equivalent, yet computationally safe, alternatives. This required a deep understanding of regex mechanics, formal language theory, and algorithmic optimization. The outcome was a set of practical guidelines for constructing robust regex expressions that mitigate ReDoS risks without sacrificing functionality or clarity.