MétaCan
Menu
Back to cohort

An Application of Automated Machine Learning Within a Data Farming Process

2022· article· en· W4317792675 on OpenAlex

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.

affAt least one author lists a Canadian institution in the pinned OpenAlex snapshot.

Bibliographic record

Venue2022 Winter Simulation Conference (WSC) · 2022
Typearticle
Languageen
FieldDecision Sciences
TopicSimulation Techniques and Applications
Canadian institutionsDepartment of National DefenceDefence Research and Development Canada
Fundersnot available
KeywordsMetamodelingComputer scienceRobustness (evolution)Machine learningArtificial intelligenceProcess (computing)Software engineeringProgramming language

Abstract

fetched live from OpenAlex

Data farming is a simulation-based methodology used within the defense community to analyze complex systems and provide insights to decision makers. It can produce very large, multi-dimensional data sets that require sophisticated analysis tools, such as metamodeling. Advances in explainable artificial intelligence have expanded the types of metamodels that can be considered; however, constructing a well-fitting machine learning metamodel involves many tasks that can become time consuming for an analyst. Automated machine learning (autoML) can save an analyst time by automating metamodel training, tuning and testing. Using outputs of an agent-based simulation of a military ground-based air defense scenario, we compared the performance of metamodels trained using autoML and different experimental designs. We found that autoML can reasonably automate the construction of metamodels and adds robustness to the analysis by considering multiple types of metamodels; however, the type and size of experimental design can significantly impact metamodel 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.

Full frame distilled prediction

Teacher imitation

Not 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.

metaresearch head score (Codex)0.002
metaresearch head score (Gemma)0.001
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesInsufficient payload (model declined to judge)
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Simulation or modeling · Consensus signal: Simulation or modeling
GenreCandidate signal: Empirical · Consensus signal: none
Teacher disagreement score0.764
Threshold uncertainty score0.999

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0020.001
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0000.000
Bibliometrics0.0000.001
Science and technology studies0.0010.000
Scholarly communication0.0000.001
Open science0.0020.001
Research integrity0.0000.000
Insufficient payload (model declined to judge)0.0020.000

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.

Opus teacher head0.173
GPT teacher head0.466
Teacher spread0.293 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it