Learning-Based Assume-Guarantee Regression Verification

Due to enormous resource consumption, model checking each revision of evolving systems repeatedly is impractical. To reduce cost in checking every revision, contextual assumptions are reused from assumeguarantee reasoning. However, contextual assumptions are not always reusable. We propose a fine-grained learning technique to maximize the reuse of contextual assumptions. Based on fine-grained learning, we develop a regressional assume-guarantee verification approach for evolving systems. We have implemented a prototype of our approach and conducted extensive experiments (with 1018 verification tasks). The results suggest promising outlooks for our incremental technique.