An Empirical Study of Self-Admitted Technical Debt in Machine Learning Software
Bibliographic record
Abstract
The emergence of open source ML libraries such as TensorFlow and Google Auto ML has enabled developers to harness state-of-the-art ML algorithms with minimal overhead. However, during this accelerated ML development process, said developers may often make sub-optimal design and implementation decisions, leading to the introduction of technical debt that, if not addressed promptly, can significantly impact on the quality of ML-based software. Developers frequently acknowledge these sub-optimal design and development choices through code comments written during development. These comments, which often highlight areas requiring additional work or refinement in the future are known as self-admitted technical debt (SATD) . While prior research has demonstrated that SATD can serve as a reliable indicator of technical debt and has extensively studied SATD in traditional (non-ML) software, little attention has been given to this issue in the context of ML. This article aims to investigate the occurrence of SATD in ML code by analyzing 318 open source ML projects across five domains, along with 318 non-ML projects. We detected SATD in source code comments in various snapshots of the studied projects, conducted a manual analysis of a sample of the identified SATD to comprehend the nature of technical debt in the ML code, and performed a survival analysis of the SATD to understand the evolution dynamics of such debts. Our analyses yielded the following observations: (i) ML projects have a median percentage of SATD that is twice that of non-ML projects. (ii) ML pipeline stages for data preprocessing and model generation logic are more susceptible to debt than model validation and deployment stages. (iii) SATDs appear in ML projects earlier in the development process compared to non-ML projects. (iv) Long-lasting SATDs are typically introduced during extensive code changes that span multiple files, which exhibit low complexity. Our research contributes to the understanding of technical debt in an ML context and underscores the need for targeted debt management strategies. This contribution is particularly relevant for developers and stakeholders in ML projects by aiding them in identifying and addressing technical debt proactively and paving the way for future research in developing automated tools and methodologies for managing SATD in an ML environment.
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 | 0.005 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 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 | 0.001 |
| 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".