Bibliographic record
Abstract
Building machine learning (ML) tools, or systems, for use in manufacturing environments is a challenge that extends far beyond the understanding of the ML algorithm. Yet, these challenges, outside of the algorithm, are less discussed in literature. Therefore, the purpose of this work is to practically illustrate several best practices, and challenges, discovered while building an ML system to detect tool wear in metal CNC machining. Namely, one should focus on the data infrastructure first; begin modeling with simple models; be cognizant of data leakage; use open-source software; and leverage advances in computational power. The ML system developed in this work is built upon classical ML algorithms and is applied to a real-world manufacturing CNC dataset. The best-performing random forest model on the CNC dataset achieves a true positive rate (sensitivity) of 90.3% and a true negative rate (specificity) of 98.3%. The results are suitable for deployment in a production environment and demonstrate the practicality of the classical ML algorithms and techniques used. The system is also tested on the publicly available UC Berkeley milling dataset. All the code is available online so others can reproduce and learn from the results.
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 | 0.022 | 0.059 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.005 | 0.006 |
| Science and technology studies | 0.001 | 0.004 |
| Scholarly communication | 0.007 | 0.009 |
| Open science | 0.006 | 0.003 |
| Research integrity | 0.004 | 0.007 |
| Insufficient payload (model declined to judge) | 0.002 | 0.005 |
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".