From Requirements as AMR-like Graphs to Automata-Based Reasoning
Aurélien Lamercerie, Annie Foret · HAL (Le Centre pour la Communication Scientifique Directe) · 2021
Requirement documents, including natural language statement sets, form the basis for system design. Their validation is a critical task. Formal methods exist to assist the designer in this way, but they appear too complex for practical use. This paper presents a new methodology that aims at reducing this gap. It is based on graphbased meaning representations and compositional transduction operations, so as to extract and formalize requirements in suitable formats. Our targets are automata enabling efficient formal reasoning. This proposal is also intended to be simple, reliable and transparent.