Towards Reproducible Machine Learning-Based Process Monitoring and Quality Prediction Research for Additive Manufacturing
Bibliographic record
Abstract
Abstract Machine learning (ML)-based monitoring systems have been extensively developed to enhance the print quality of additive manufacturing (AM). In-situ and in-process data acquired using sensors can be used to train ML models that detect process anomalies, predict part quality, and adjust process parameters. However, the reproducibility of the proposed AM monitoring systems has not been investigated. There has not been a method to evaluate and improve reproducibility in the joint domain of AM and ML. Consequently, some crucial information for reproducing the research is usually missing from the publications; thus, systems reproduced based on the publications often cannot achieve the claimed performance. This paper establishes the definition of reproducibility in this domain, proposes a reproducibility investigation pipeline, and composes a reproducibility checklist. A research is reproducible if a performance comparable to the original research can be obtained when reproduced by a different team using a different experiment setup. The reproducibility investigation pipeline sequentially guides the readers through all the necessary reproduction steps, during which the reproducibility checklist will help extract the reproducibility information from the publication. A case study that reproduced a vision-based warping detection system demonstrated the usage and validated the efficacy of the proposed pipeline and checklist. It has been observed that the reproducibility checklist can help the authors verify that all the information critical to reproducibility is provided in the publications. The investigation pipeline can help identify the missing reproducibility information, which should be acquired from the original authors to achieve the claimed performance.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.104 | 0.186 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.003 |
| Bibliometrics | 0.006 | 0.004 |
| Science and technology studies | 0.001 | 0.004 |
| Scholarly communication | 0.009 | 0.008 |
| Open science | 0.005 | 0.006 |
| Research integrity | 0.003 | 0.004 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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 source (direct Gemma or distilled Codex), 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".