Agile Teaching: Automated Student Support and Feedback Generation
Bibliographic record
Abstract
The software engineering industry prioritizes efficiency through the use of tools and processes such as version control and Agile methodologies. Automation has allowed developers to be more productive and deliver superior results. However, automation is not as widespread in software engineering education. To improve educational outcomes and provide more effective feedback to students, this research implements software engineering methods and automation techniques in software engineering education and evaluates their effectiveness. Our focus is on establishing a connection between source code and natural language, allowing us to generate a natural language description from a given source code sample. To generate feedback, we employ deep learning models to learn code representations and gain a deeper understanding of code. However, the complexity of code makes it challenging to learn its representation accurately. After learning about code, we compare students' code with the instructor-provided solution based on configurable thresholds and generate comments to provide guidance. This work extends the Transformer model, GraphCodeBERT, which is a pre-trained model for programming languages that incorporates the inherent structure of code. We utilize both syntax-level information, such as abstract syntax trees, and semantic-level information, such as data flow, during pre-training. The data flow graph has nodes representing variables and edges indicating the “where-the-value-comes-from” relationship between variables. Our model is based on the Transformer neural architecture and uses a gated graph neural network model to learn code embeddings. This function incorporates code structure, a copy mechanism, and relative position representations, allowing the model to better understand the semantics of code. We evaluated our approach on the Java dataset in terms of BLEU, METEOR, and ROUGE-L metrics and compared it with the state-of-the-art code comment generation model, GTrans. Our model demonstrated improvements in two metrics of METEOR and ROUGE-L.
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 | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| 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".