Researchers have developed ConVer, a new tool that uses large language models (LLMs) to aid in the formal verification of large C programs. ConVer employs a top-down compositional approach, synthesizing function contracts from system properties and iteratively refining them through a CEGAR-CEGIS loop. The tool has demonstrated significant success rates across various benchmark suites, including Frama-C, X.509 parser, LF2C-Simple, and VerifyThis, with some configurations achieving over 90% verification success. AI
IMPACT This research demonstrates how LLMs can be integrated into formal verification processes, potentially improving the scalability and efficiency of ensuring software correctness.
RANK_REASON The cluster contains a research paper detailing a new tool and methodology for software verification.
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
