A systematic mapping study on the employment of neural networks on software engineering projects: Where to go next?
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Abstract Deep learning has recently experienced explosive growth in use, largely due to advances in neural networks and the availability of large corpora of domain data. Project management activities generate and handle large volumes of data. Software engineering closely relates to project management, so software engineering projects must be prone to the use of neural networks. We seek to obtain an accurate vision of how neural networks are being used in software engineering projects through a systematic mapping study. We confirm that neural networks have already made their way into these projects; however, we show that their current uses are limited to certain repetitive and legacy tasks. Given uncovered ample room for expansion, we point out a few directions the industry and academy can lean toward to in the next years for taking better advantage of neural networks in software engineering projects and immediately advancing the field. We investigate if, how, and to what extent have neural networks been employed to the advancement of software engineering projects. As such, a systematic mapping study was conducted, which led to the conclusion that even though these algorithms have indeed been employed on several software engineering tasks, this employment so far has been shy, mostly relying on legacy types of neural networks. More modern variants, namely, deep learning algorithms, are slowly gaining momentum and should be the trend going forward.
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.
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 | 0.004 |
| 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 | 0.001 | 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 it