Explore a groundbreaking synthesis algorithm presented at POPL 2024 that efficiently searches programs with local variables. Dive into the concept of lifted interpretation, which enables simultaneous evaluation of all programs from a grammar, allowing for effective reduction of the search space for programs with local variables. Learn how this innovative approach overcomes limitations of prior bottom-up synthesis algorithms, particularly in evaluating programs with free local variables. Discover the application of these ideas in web automation through Arborist, a tool that outperforms state-of-the-art techniques like WebRobot and Helena in automating complex tasks. Gain insights into the algorithm's potential to significantly expand the range of automatable tasks in web environments.
Overview
Syllabus
[POPL'24] Efficient Bottom-Up Synthesis for Programs with Local Variables
Taught by
ACM SIGPLAN