Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore a 19-minute conference talk from POPL 2025 that presents an innovative method for conducting algebraic program analysis (APA) incrementally when programs change. Learn about the two main contributions: an efficient tree-based data structure for representing path expressions and techniques for quickly updating program properties when path expressions change. The presenters, Chenyu Zhou, Yuzhou Fang, Jingbo Wang, and Chao Wang from USC and Purdue University, demonstrate how their approach significantly outperforms baseline APA and state-of-the-art methods, achieving speedups ranging from 160X to 4761X across thirteen Java applications from the DaCapo benchmark suite. The talk covers key concepts in data-flow analysis, side-channel analysis, and incremental algorithms that are essential for efficient program analysis.
