Boolean satisfiability is a fundamental problem in computerscience with a wide range of applications including planning, configurationmanagement, design and verification of software/hardware systems. Modern SAT solvers achieve scalability and ro-bustness with sophisticated heuristics that are challenging to understandand explain. We propose to view modern conflict-driven clause learning (CDCL) solvers as a composition of classifiers and regressors for different tasks such as branching, clause memory management, and restarting. The current version of CrystalBall focuses on deriving a classifier to keep or throw away a learned clause. In a departure from recent machine learning based techniques, CrystalBall employs supervised learning anduses extensive, multi-gigabyte data extracted from runs of a single SAT solver to perform predictive analytics. Read this blog post for more details.
Relevant Papers:
