Family
LightGBM meta on 6 forensic layers
productionApache 2.0visionboth
Forgery Detector.
PR-trio of audit roadmap.
Foundation · LightGBM meta on 6 forensic layers
What it does
The honest description.
PR-trio of audit roadmap.
Architecture
How it works.
P(forged) ∈ [0, 1].
Training data & drift
What it learned from, and how we watch it.
Source
Drift baseline · current quarter
Drift baseline not yet documented for this model.
What it's NOT
- The 6 layers stay (rule-based forensic feature extractors); the trainer learns optimal combination weights from data.
- Heldout is synthetic (real forged-doc corpora are DPDP-sensitive).
Who calls this model
Across the house and the API.
Internal routes
No public routes documented yet.
Featured in
/features/documents-kycVersion history · immutable once published
Every release, traceable.
Data Mukodi holds about you is subject to DPDP §12 erasure (use /consent in the app), and any concern about this model can be raised via [email protected] under DPDP §13 grievance, or via /security.