Towards Understanding the Impacts of Textual Dissimilarity on Duplicate Bug Report Detection
Bibliographic record
Abstract
About 40% of software bug reports are duplicates of one another, which pose a major overhead during software maintenance. Traditional techniques often focus on detecting duplicate bug reports that are textually similar. However, in bug tracking systems, many duplicate bug reports might not be textually similar, for which the traditional techniques might fall short. In this paper, we conduct a large-scale empirical study to better understand the impacts of textual dissimilarity on the detection of duplicate bug reports. First, we collect a total of 92,854 bug reports from three open-source systems and construct two datasets containing textually similar and textually dissimilar duplicate bug reports. Then we determine the performance of three existing techniques in detecting duplicate bug reports and show that their performance is significantly poor for textually dissimilar duplicate reports. Second, we analyze the two groups of bug reports using a combination of descriptive analysis, word embedding visualization, and manual analysis. We found that textually dissimilar duplicate bug reports often miss important components (e.g., expected behaviors and steps to reproduce), which could lead to their textual differences and poor performance by the existing techniques. Finally, we apply domain-specific embedding to duplicate bug report detection problems, which shows mixed results. All these findings above warrant further investigation and more effective solutions for detecting textually dissimilar duplicate bug reports.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 teacher head, 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".