Human-AI interaction: Machine learning-based geostatistical hybrid models
Bibliographic record
Abstract
Intelligent methods for estimating mineral grades have been developed but some tasks cannot be completely automated through artificial intelligence. Human-in-the-loop (HITL) approaches are being increasingly utilized, where the strengths of both human expertise and artificial intelligence are combined to improve outcomes. This study integrates HITL models with machine learning (ML) based geostatistical hybrid modelling and ensembling approaches for mineral grade estimation and ore sorting. In the hybrid modelling approach, ML models such as an elliptical radial basis neural network (ERBFN), locally weighted support vector regression (LWSVR), kernel density estimated trend (KDET), and a convolutional neural network (CNN) are incorporated as secondary variables within intrinsic collocated cokriging (ICCK). Additionally, the study utilizes two types of ensemble models—global (GWE) and local weighting-based (LWE) ensembles. These ensembles integrate outputs from hybrid models, applying global and local weights based on each model’s cross-validation performance. Depending on their level of expertise, humans are integrated as either (1) novice practitioners considered as human-as-feedback (HAF) systems where they act as model checks and key parameter validators, without the ability to influence ML training or (2) expert practitioners considered as systems where model parameters are actively adjusted, model structures are tuned, and the learning process is guided by human experts. The effectiveness of the HAF and HAC systems is evaluated using data from multiple blast areas obtained from an open-pit copper mine. Compared to fully automated modelling, the HAF system improved estimation accuracy in terms of R 2 values by between 3.6% ( ICCK CNN ) and 5.9% (GWE) across hybrid and ensemble models. Meanwhile, the HAC system demonstrated more significant enhancements, with R 2 values increase ranging from 5.0% ( ICCK CNN ) to 16.5% (GWE) for these same models. This advancement suggests the potential for more precise and effective decision-making in mining operations using HITL systems.
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 | 0.001 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 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 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".