Agentic LLM Methodology Automates Requirement Formalization for Safety-Critical Systems

arXiv·medium signal

Proposes an agentic LLM pipeline that automatically converts natural language specifications for safety-critical systems into formally verifiable properties, going beyond syntactic correctness to ensure semantic alignment between informal requirements and formal artifacts. Unlike one-shot LLM formalization, the agentic approach iteratively refines until formal properties are verification-ready. Relevant for teams using LLMs in regulated domains where formal verification is required.