Multi-Label Ambiguity Detection in Software Requirements Using Language Models
Bibliographic record
Abstract
In software engineering, requirements specification plays an important role in capturing and analyzing stakeholders' needs. However, as software requirements are typically written in natural language, they usually have ambiguity. Identifying and understanding the ambiguity in requirements enables development teams to prevent misunderstandings, address defects earlier, and reduce project costs and failures. While most recent studies focus on detecting only one or two types of ambiguity, this study investigates whether language models (LMs) can detect multiple types of ambiguity, specifically structural ambiguity, scope ambiguity, anaphoric ambiguity, and vagueness. We collected 1,266 real-world requirements from the public requirements (PURE) dataset, which includes software requirements specifications from several domains, and labeled them using established methods. Two LMs, DistilBERT and RoBERTa, were fine-tuned using the labeled dataset (split into training and validation sets). Model performance was evaluated on an independent test set of 255 manually annotated requirements. RoBERTa achieved a higher macroaveraged F1-score (75.61%) compared to DistilBERT (73.15%). Applying back-translation (BT) data augmentation to address class imbalance further improved RoBERTa's macro-averaged F1-score to 78.84%. Compared to existing techniques, our multilabel method achieved approximately 12% and 7% higher F1- scores in scope and structural ambiguity detection, respectively, while demonstrating slightly lower results for anaphoric ambiguity (about 1.8%) and vagueness detection (about 3%).
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.006 | 0.019 |
| Meta-epidemiology (narrow) | 0.002 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.004 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".