A Short Survey on Formalising Software Requirements using Large Language Models

TL;DR

This survey reviews how large language models are used to help formalize software requirements, highlighting key research, examples, and future directions in this emerging field.

Contribution

It provides a comprehensive overview of 35 studies on LLM-assisted formal requirements specification, including practical examples and insights for future research.

Findings

Identifies key papers and methodologies in LLM-based formalization

Highlights practical examples in Dafny, C, and Java

Suggests future research directions for LLM applications

Abstract

This paper presents a focused literature survey on the use of large language models (LLM) to assist in writing formal specifications for software. A summary of thirty-five key papers is presented, including examples for specifying programs written in Dafny, C and Java. This paper arose from the project VERIFAI - Traceability and verification of natural language requirements that addresses the challenges in writing formal specifications from requirements that are expressed in natural language. Our methodology employed multiple academic databases to identify relevant research. The AI-assisted tool Elicit facilitated the initial paper selection, which were manually screened for final selection. The survey provides valuable insights and future directions for utilising LLMs while formalising software requirements.