Neural Language Models and Few-Shot Learning for Systematic Requirements Processing in Model-Driven Software Engineering

Explore the application of neural language models and few-shot learning for systematic requirements processing in Model-Driven Software Engineering (MDSE). This 18-minute conference talk from ACM SIGPLAN addresses the challenges of handling increasing numbers of requirements in systems engineering, particularly in the automotive domain. Learn how domain-specific language constructs can help avoid ambiguities and increase formality in requirements. Discover the main contribution of adopting and evaluating few-shot learning with large pretrained language models for automated translation of informal requirements to structured languages, such as a requirement Domain-Specific Language (DSL). Gain insights into overcoming the challenges of translating masses of requirements and training requirements engineers when introducing formal requirement notations in existing projects.

[SLE] Neural Language Models and Few Shot Learning for Systematic Requirements Processing in MDSE