Can Graph Neural Networks Learn to Solve the MaxSAT Problem? (Student Abstract)

The paper presents an attempt to bridge the gap between machine learning and symbolic reasoning. We build graph neural networks (GNNs) predict solution of Maximum Satisfiability (MaxSAT) problem, optimization variant SAT. Two closely related representations are adopted, we prove their theoretical equivalence. also show that GNNs can achieve attractive performance solve hard MaxSAT problems in certain distributions even compared with state-of-the-art solvers through experimental evaluation.