Prediction of successful training outcomes for drug detection dogs using subjective ratings and behavioral test measures: A case study in Japan customs
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
Drug detection dogs, primarily employed by customs and police forces, play a crucial role in preventing the spread of illegal drugs worldwide. To minimize training costs, accurately predicting which dogs will succeed in scent detection training is essential. Local training organizations seek validated behavioral assessment methods for this purpose, but the wide range of methods used and the lack of scientific verification pose challenges. Previous research on detection dogs in Japan focused on genetics, but behavioral assessment methods for training have been understudied. To bridge the gap, the current study aimed to outline and evaluate the predictive validity of behavioral assessment systems used for drug detection dogs at Japan Customs. We compared the relative effectiveness of two different behavioral assessment methods: subjective ratings by chief trainers and behavioral measures in a novel test situation. For subjective ratings, we used subscales of Training Focus (i.e., interest in play, independence, concentration, activity, and boldness) and Tolerance (i.e., friendliness to humans and tolerance to dogs) to characterize a dog’s personality. For behavioral tests, a simple behavioral test measured a dog’s approach behavior and reactivity to an unfamiliar person. Data from 196 dogs (159 Labrador Retrievers and 37 German Shepherds) showed high inter-rater agreement for both methods. A GLMM model revealed that Training Focus subscale scores significantly predicted training success of candidate dogs. On the other hand, Tolerance scores and behavioral test measures were poor predictors for scent detection work. Dog breed and sex did not significantly influence final training outcomes. Receiver Operating Characteristic (ROC) curves indicated that Training Focus scales' classification performance for training success is comparable to or better than previous reports for assistance and detection dogs. These findings demonstrate the predictive validity of subjective Training Focus ratings, aiding in the selection of drug detection dogs at Japan Customs. While generalizability to other detection dog populations and identification of alternative behavioral predictors remains uncertain, this study provides valuable insights into the predictive accuracy of trainer ratings in a dog behavior assessment system. • Trainer’s ratings of Training Focus strongly predicted the success of detection dogs. • Trainer’s ratings of Tolerance did not predict the success of detection dogs. • Behavioral test measures did not predict the success of drug detection dogs. • Dog breed and sex did not significantly influence final training outcomes.
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 | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| 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