v0.2.0 — Head-to-head: individual vs population reference

v0.2.0 — Head-to-head: individual vs population reference

A reproducible synthetic head-to-head addressing the open empirical question in the framework paper (§6, Q1): does the individual baseline actually outperform the population reference for welfare-relevant change?

This is a proof-of-concept on fully synthetic data — not validation on real dogs. No real animals, sensors, or clinical outcomes; the decline signal is injected and known. Not diagnostic, not production.

What's new

• examples/head_to_head_validation.py — individual- vs population-referenced detection on a heterogeneous synthetic cohort, built on the existing barkley package (same ICF baseline, same CUSUM detector; only the reference frame differs).
• results/barkley_head_to_head.png — 4-panel figure: mean ROC ± SD, detection-lag distributions, heterogeneity sweep with error bars, and a calm-dog false-reassurance illustration.
• results/HEAD_TO_HEAD_RESULTS.md — method, fairness contract, and full results.
• README — new "Worked example" section.

Method (fairness contract)

Both arms use the same detector (CUSUM), the same robust statistic (median + MAD), and the same 56-day calibration window — the population arm additionally gets the pooled-cohort advantage. Healthy controls are included and false alarms are measured (threshold-independent ROC + a matched operating point). Every condition is averaged over 30 independent synthetic seeds (mean ± SD).

Result (synthetic, illustrative)

As between-dog heterogeneity grows, the population "normal" band widens until a calm dog's genuine decline stays inside it — the false-reassurance mechanism the paper names the reference-class trap. At a matched 10% false-alarm rate, the individual reference holds 100% ± 0% detection across all heterogeneity levels tested, while the population reference falls from ~100% to 56% ± 22% as heterogeneity grows. AUC: individual 0.988 ± 0.010 vs population 0.935 ± 0.020. At low heterogeneity the two are comparable (population marginally better by AUC) — the advantage is conditional, not universal.

The real-data head-to-head remains the open validation step.

---

Run: python examples/head_to_head_validation.py → results/barkley_head_to_head.png
© 2026 Élodie Aishwarya P. Remoissenet — Barkley AI

v0.1.1 — Zenodo archiving + ecosystem links

Patch release to trigger Zenodo DOI archiving.

Changes since v0.1.0:

• Added Reference Architecture link in Drift Explorer footer
• Updated Hugging Face dataset card with companion repository
• Updated barkley-canine-cognition-lab hub with Public Research Artifacts section

v0.1.0 — Initial public release

Release Notes

v0.1.0 — Initial public release (2026-05-24)

First public, source-available release of the Barkley Reference Architecture —
the minimal, runnable computational backbone behind the Barkley research thesis:

A dog can remain normal for its breed while drifting away from its own
behavioral baseline.

What's included

This release demonstrates four research pillars plus the structuring schema,
on fully synthetic data:

• Individual Baseline / ICF (barkley/individual_baseline.py) — per-dog
  median + scaled-MAD baseline with slow exponential maintenance.
• Multi-Resolution Temporal Binning (barkley/temporal_binning.py) —
  circadian (Bin A), weekly (Bin B), and quarterly (Bin C) windows, including a
  from-scratch two-sample Kolmogorov–Smirnov distribution-shift test.
• Rate of Drift / Behavioral Drift (barkley/rate_of_drift.py) — OLS drift
  slope, two-sided CUSUM, and a lightweight BOCPD-BLS baseline-shift detector.
• Sovereignty of Silence / Missingness (barkley/missingness.py) — the
  Artefactual / Contextual / Informative taxonomy and feature encoding.
• DogGraph event schema (barkley/schema.py) — the memory and structuring
  layer that ties the pillars into a typed, serialisable behavioral event.

Runnable reference pipeline

examples/run_reference_pipeline.py runs the full pipeline end-to-end on one
synthetic dog and prints a reproducible summary:

events_generated:            309
baseline_window_days:        56
drift_detected:              true
missingness_classified:      true
ks_distribution_shift:       0.564
research_demonstrator_only:  true

Tests

16 tests covering all four pillars and the schema. Runnable two ways:

python tests/test_reference_pipeline.py     # no external dependency
python -m pytest tests/ -v                  # pytest

Requirements

Python 3.10+, numpy, pandas. No GPU, no internet, no external services.

Scope and limitations

• Fully synthetic data only; no real animal data.
• All numerical parameters are illustrative research defaults, not validated
  operational thresholds.
• Research demonstrator only; not diagnostic, not production.
• This is the minimal backbone; it does not reproduce or expose the full
  Barkley system.

Citation

This release accompanies the Barkley framework paper
(DOI: 10.5281/zenodo.20060327). A separate software DOI will be minted once this
release is archived on Zenodo; CITATION.cff will be updated at that point.

---

Research demonstrator only. Synthetic data only. Not diagnostic.
labs@getbarkley.com