Suman Jana Awarded an ARO Young Investigator Award
Suman Jana, assistant professor of computer science, has been selected by the Army Research Office (ARO) to receive a Young Investigator Program (YIP) award for his proposal, “Building Efficient Fuzzers using Automata Learning.” Among the most prestigious honors awarded by the Army to researchers, ARO YIP awards recognize early-career faculty who pursue research in areas relevant to the Army. The award both supports their research and encourages their teaching and research careers.
The four-year, $360,000 award will support Jana’s research into incorporating machine learning methods to improve methods for finding and fixing software bugs before they become security vulnerabilities.
For this project, Jana will research and build a more efficient type of fuzzer, a software testing method. Fuzzers work by providing unexpected and randomized inputs, called “fuzz,” that are fed into software programs to find test cases that cause crashes, memory leaks, and other problematic behaviors. How those inputs are generated is key to maximizing coverage and ensuring all parts of the code are tested. Sophisticated fuzzers, such as Google’s American Fuzzy Lop (AFL) and libFuzzer, incorporate feedback from each test case to learn the format of the input to increase the number of test cases. This evolutionary approach, though popular, is too unsystematic and too dependent on previous test cases to find other, deeper bugs unrelated to previous test cases.
“The funding provided by ARO will allow us to explore the best ways to incorporate automatamodels to test software,” says Jana, who is also a member of Columbia’s Data Science Institute. “Such models are able to learn behaviors on all prior inputs and use this information to generate new input, allowing us to test unexplored parts of the program and continually refine the learned model. We are grateful to the ARO for making the project possible.”
An assistant professor of computer science at Columbia since 2016, Suman Jana earned his PhD in 2014 from the University of Texas, where he was supported by a Google PhD Fellowship.