Machine Learning-Based Delineation of Geodomain Boundaries: A Proof-of-Concept Study Using Data from the Witwatersrand Goldfields
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
Abstract Machine-aided geological interpretation provides an opportunity for rapid and data-driven decision-making. In disciplines such as geostatistics, the integration of machine learning has the potential to improve the reliability of mineral resources and ore reserve estimates. In this study, inspired by existing geostatistical approaches that use radial basis functions to delineate domain boundaries, we reformulate the problem into a machine learning task for automated domain boundary delineation to partition the orebody. We use an actual dataset from an operating mine (Driefontein gold mine, Witwatersrand Basin in South Africa) to showcase our new method. Using various machine learning algorithms, domain boundaries were created. We show that based on a combination of in-discipline requirements and heuristic reasoning, some algorithms/models may be more desirable than others, beyond merely cross-validation performance metrics. In particular, the support vector machine algorithm yielded simple (low boundary complexity) but geologically realistic and feasible domain boundaries. In addition to the empirical results, the support vector machine algorithm is also functionally the most resemblant of current approaches that makes use of radial basis functions. The delineated domains were subsequently used to demonstrate the effectiveness of domain delineation by comparing domain-based estimation versus non-domain-based estimation using an identical automated workflow. Analysis of estimation results indicate that domain-based estimation is more likely to result in better metal reconciliation as compared with non-domained based estimation. Through the adoption of the machine learning framework, we realized several benefits including: uncertainty quantification; domain boundary complexity tuning; automation; dynamic updates of models using new data; and simple integration with existing machine learning-based workflows.
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 | 0.003 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.002 | 0.001 |
| 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 it