A DSL and model transformations to specify learning corpora for modeling assistants
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
Software engineering undergraduate students spend a significant time learning various topics related to software design, including notably model-driven engineering (MDE), where different types of structural and behavioral models are used to design, implement, and validate an application. MDE instructors spend a lot of time covering modeling concepts, which is more difficult with ever-increasing class sizes. Online resources, such as learning corpora for domain modeling, can aid in this learning process by serving as a more dynamic textbook alternative or as part of a larger interactive application with domain modeling exercises and tutorials. A Learning Corpus (LC) is an extensible list of entries representing possible mistakes that could occur when defining a model, e.g., Missing Abstraction-Occurrence pattern in the case of a domain model. Each LC entry includes progressive levels of feedback, including written responses, quizzes, and references to external resources. To make it easy for instructors to customize the entries as well as add their own, we propose a novel, simple, and intuitive approach based on an internal domain-specific language that supports features such as context-specific information and concise arbitrary metamodel navigation with shorthands. Transformations to source code as well as Markdown and LATEX enable use of the LC entries in different contexts. These transformations as well as the integration of the generated code in a sample Modeling Assistant application verify and validate the LC metamodel and specification.
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 | 0.000 | 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 it