Using Language Models for Enhancing the Completeness of Natural-language Requirements

TL;DR

This paper explores using BERT, a language model, to identify missing information in natural-language requirements by predicting omitted content, with a focus on optimizing prediction strategies and reducing noise for better completeness detection.

Contribution

It introduces an optimal number of predictions per mask for BERT and a machine learning filter to improve the accuracy of missing content detection in requirements.

Findings

BERT effectively predicts missing terminology in requirements.

The filter significantly reduces noise in BERT's predictions.

Optimal predictions per mask balance discovery and noise.

Abstract

[Context and motivation] Incompleteness in natural-language requirements is a challenging problem. [Question/problem] A common technique for detecting incompleteness in requirements is checking the requirements against external sources. With the emergence of language models such as BERT, an interesting question is whether language models are useful external sources for finding potential incompleteness in requirements. [Principal ideas/results] We mask words in requirements and have BERT's masked language model (MLM) generate contextualized predictions for filling the masked slots. We simulate incompleteness by withholding content from requirements and measure BERT's ability to predict terminology that is present in the withheld content but absent in the content disclosed to BERT. [Contribution] BERT can be configured to generate multiple predictions per mask. Our first contribution is to determine how many predictions per mask is an optimal trade-off between effectively discovering omissions in requirements and the level of noise in the predictions. Our second contribution is devising a machine learning-based filter that post-processes predictions made by BERT to further reduce noise. We empirically evaluate our solution over 40 requirements specifications drawn from the PURE dataset [1]. Our results indicate that: (1) predictions made by BERT are highly effective at pinpointing terminology that is missing from requirements, and (2) our filter can substantially reduce noise from the predictions, thus making BERT a more compelling aid for improving completeness in requirements.