Automated Functional Synthesis: An Ideal Meeting Ground for Symbolic Reasoning and Machine Learning

Explore the groundbreaking approach to functional synthesis presented in this 34-minute talk by Kuldeep Meel from the University of Toronto. Delve into the Manthan method, which ingeniously combines machine learning and symbolic reasoning to tackle one of computer science's most fundamental problems. Discover how this data-driven approach views functional synthesis as a classification problem, leveraging advances in constrained sampling for data generation and automated reasoning for proof-guided refinement and verification. Gain insights into the significant performance improvements achieved and the exciting future prospects at the intersection of machine learning and symbolic reasoning in system synthesis.