MétaCan
Menu
Back to cohort
Record W4210408749 · doi:10.46713/jdst.004.01

Military Dataset Processing Approaches or Trauma Risk Mitigation in Machine Learning Practitioners

2022· article· en· W4210408749 on OpenAlexaff
Mélanie Breton, Valérie Lavigne, Malek Djaffri, Maxime Dionne

Bibliographic record

VenueJournal of Defence & Security Technologies · 2022
Typearticle
Languageen
FieldComputer Science
TopicAdversarial Robustness in Machine Learning
Canadian institutionsDefence Research and Development Canada
Fundersnot available
KeywordsOffensiveComputer sciencePreprocessorArtificial intelligenceMachine learningWork (physics)Motion (physics)Data scienceEngineeringOperations research

Abstract

fetched live from OpenAlex

To develop new computer vision capabilities leveraging artificial intelligence, we will increasingly need to use operationally realistic training and validation datasets. Although operational full motion video and imagery datasets present information regarding their provenance and classification level, these designations are often not indicative of the presence of potentially offensive or traumatizing content. As machine learning and data scientists increasingly need to work with operational unsanitized operational video and imagery data, they will have a higher risk of being exposed to sensitive and traumatic content. In this paper, we first raise awareness about this risk within the Defense community. Then, we propose several approaches for mitigating machine learning practitioner's exposure to offensive and traumatizing media, including dataset preprocessing procedures and viewing tool design considerations.

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 imitation

Not 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.

metaresearch head score (Codex)0.017
metaresearch head score (Gemma)0.048
Version: metacan-v3-hybrid-931329e0061cValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Not applicable · Consensus signal: none
GenreCandidate signal: Empirical · Consensus signal: none
Teacher disagreement score0.017
Threshold uncertainty score0.090

Distilled classifier scores by category (both heads)

CategoryCodexGemma
Metaresearch0.0170.048
Meta-epidemiology (narrow)0.0010.001
Meta-epidemiology (broad)0.0010.001
Bibliometrics0.0020.002
Science and technology studies0.0010.002
Scholarly communication0.0040.006
Open science0.0020.003
Research integrity0.0020.002
Insufficient payload (model declined to judge)0.0070.003

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 head0.027
GPT teacher head0.267
Teacher spread0.240 · 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

Classification

machine, unvalidated

Machine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.

The models applied no category: nothing in the taxonomy fit this work.
Study designNot applicable
Domainnot available
GenreEmpirical

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".

Quick stats

Citations1
Published2022
Admission routes1
Has abstractyes

Explore more

Same venueJournal of Defence & Security TechnologiesSame topicAdversarial Robustness in Machine LearningFrench-language works237,207