Inductive Program Synthesis via Iterative Forward-Backward Abstract Interpretation
Offered By: ACM SIGPLAN via YouTube
Course Description
Overview
Explore a novel approach to inductive program synthesis presented at the PLDI 2023 conference in this 17-minute video. Delve into the innovative technique of iterative forward-backward abstract interpretation, designed to address the challenge of gigantic search spaces in example-based program synthesis. Learn how this method combines forward abstract interpretation to compute possible outputs given inputs with backward abstract interpretation to determine possible inputs given outputs. Discover how this alternating approach effectively prunes the search space and improves the efficiency of syntax-guided synthesis (SyGuS). Examine the experimental results that demonstrate significant performance improvements over state-of-the-art approaches. Gain insights into the application of sophisticated abstract interpretation techniques for a wide range of inductive synthesis tasks.
Syllabus
[PLDI'23] Inductive Program Synthesis via Iterative Forward-Backward Abstract Interpretation
Taught by
ACM SIGPLAN
Related Courses
Stanford Seminar - Concepts and Questions as ProgramsStanford University via YouTube DreamCoder- Growing Generalizable, Interpretable Knowledge With Wake-Sleep Bayesian Program Learning
Yannic Kilcher via YouTube A Neural Network Solves and Generates Mathematics Problems by Program Synthesis - Paper Explained
Aleksa Gordić - The AI Epiphany via YouTube EI Seminar - Recent Papers in Embodied Intelligence
Massachusetts Institute of Technology via YouTube Using Program Synthesis to Build Compilers
Simons Institute via YouTube