Is BERT the New Silver Bullet? - An Empirical Investigation of Requirements Dependency Classification
Bibliographic record
Abstract
Bidirectional Encoder Representations from Transformers (BERT) is a successful transformer-based Machine Learning technique for Natural Language Processing (NLP) based tasks developed by Google. It has taken various domains by storm, and Software Engineering is one among them. But does this mean that BERT is the new Silver Bullet? It is certainly not. We demonstrate it through an empirical investigation of the Requirements Dependency Classification (RDC). In general, based on various criteria used for evaluation, decisions on classification method preference may vary. For RDC, we go beyond traditional metrics such as the F1 score and consider Return-on-Investment (ROI) to evaluate two techniques for such decision making. We study RDC-BERT (fine-tuned BERT using data specific to requirements dependency classification) and compare with Random Forest, our baseline. For RDC and data from FOSS system Redmine, we demonstrate how decisions on method preference vary based on (i) accuracy, (ii) ROI, and (iii) sensitivity analysis. Results show that for all the three scenarios, method preference decisions depend on learning and evaluation parameters. Although these results are with respected to the chosen data sets, we argue that the proposed methodology is a prospective approach to study similar questions for data analytics, in general.
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 | 0.018 | 0.116 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.005 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.003 |
| Insufficient payload (model declined to judge) | 0.002 | 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".