🤖 AI Summary
In software requirements engineering, manual translation of high-level abstract features into testable functional requirements (FRs) suffers from low efficiency and poor interpretability. To address this, we propose EasyFR—a framework that formalizes FR generation as a structured slot-filling task guided by Semantic Role Labeling (SRL). Our key contributions are: (1) a novel configurable dual-variable SRL template system; (2) the Key2Temp model, which automatically maps features to optimal template variants; and (3) the first integration of an SRL-guided mechanism into the decoding process of pretrained language models. Experiments across four public benchmarks demonstrate that EasyFR significantly outperforms state-of-the-art NLG models, including GPT-4. Ablation studies confirm that SRL template recommendation critically enhances generation quality. This work establishes a new paradigm for automated, reusable, and interpretable FR synthesis.
📝 Abstract
The translation of high-level abstract features into clear, and testable functional requirements (FRs) is a crucial step in software development, bridging the gap between user needs and technical specifications. In engineering practice, significant expert effort is needed for this translation. Our approach, EasyFR, streamlines the process by recommending Semantic Role Labeling (SRL) sequences for the given abstract features to guide Pre-trained Language Models (PLMs) in producing cohesive FR statements. By analyzing ten diverse datasets, we induce two variable SRL templates, each including two configurable parts. For concrete features, our proposed Key2Temp model can construct the appropriate variant of the SRL template by identifying a variable SRL template and placing the feature tokens in the appropriate slots. In this way, our approach reframes the process of requirement generation into a structured slot-filling activity. Experimental validation on four open datasets demonstrates that EasyFR outperforms three advanced Natural language generation (NLG) approaches, including GPT4, particularly when existing FRs are available for training. The positive influence of our SRL template variant recommendations is further confirmed through an ablation study. We believe that our results indicate a notable step forward in the realm of automated requirements synthesis, holding potential to improve the process of requirements specification in future software projects.