DiffWatch: Watch Out for the Evolving Differential Testing in Deep Learning Libraries
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
Testing deep learning libraries is ultimately important for ensuring the quality and safety of many deep learning applications. As differential testing is commonly used to help the creation of test oracles, its maintenance poses new challenges. In this tool demo paper, we present DiffWatch, a fully automated tool for Python, which identifies differential test practices in DLLs and continuously monitors new changes of external libraries that may trigger the updates of the identified differential tests.Our evaluation on four DLLs demonstrates that DiffWatch can detect differential testing with a high accuracy. In addition, we demonstrate usage examples to show DiffWatch’s capability of monitoring the development of external libraries and alert the maintainers of DLLs about new changes that may trigger the updates of differential test practices. In short, DiffWatch can help developers adequately react to the code evolution of external libraries. DiffWatch is publicly available and a demo video can be found at https://www.youtube.com/watch?v=gR7m5QQuSqE.
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 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.005 | 0.002 |
| Research integrity | 0.000 | 0.002 |
| 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