KP
Chanakya v3.5 · Gothamಕರ್ನಾಟಕ ರಾಜ್ಯ ಪೊಲೀಸ್

OFFICER ID

S. Kulkarni

SK

ಗುಪ್ತಚರ ಎಂಜಿನ್‌ಗಳು· Intelligence Engines [GOTHAM_v4]

5 Volumes · MIRE · EDGE · XAI · LEADS · KG

Master Identities

0

Scored Edges

0

Evidence Items

0

Inferred Facts

0

Open Leads

0

Vol 1 · MIRE

Master Identity Resolution

Jaro-Winkler + weighted attribute scoring · every merge audited

Live Merge Decision · PER-000000123100

✓ Auto-merge · ≥ 95

  • aadhaarXXXX-XXXX-4429 · both sources+100
  • mobile+91 9876543210 (normalised)+35
  • nameJaro-Winkler 0.983 · Ramesh Kumar ⇔ Rameshkumar+20
  • fatherS. Gowda+15

Master Registry · 0

Vol 2 · Edge Confidence

Evidence-Backed Relationships

Every edge is a defensible sum of unique evidence

Vol 3 · Explainable AI

Match Reasoning · Court-Ready

Every score defended by a structured evidence chain

Vol 4 · Recommendations

Investigator Lead Queue

Ranked, deduplicated, explainable actions

Vol 5 · KG Reasoning

Inferred Facts · Provenance Preserved

Forward-chained inference with replayable proof chains

All inferences emitted as immutable DomainEvents · replayable through the Sovereign Vault ledger

Vol 5 · Rete Forward Chain

Incremental Materialisation · α → β → Production

Stream observed facts; only rules touching the changed predicate re-evaluate

α-Memory · Observed Facts

  1. Working memory empty — assert facts to trigger the network

Production Firings · Inferred Facts

  1. No rules fired yet · β-network unmatched
Rete network · 5 rules · every firing emitted as FACT_INFERRED DomainEvent

Vol 5 · Backward Chain

Goal-Driven Justification · Court-Ready Proof

Ask a specific question; receive a full chain of antecedents that justify the answer

Query Goal
PER-A · Ramesh Gowdaperson_of_interest_forSCENE-D · KA-BLR-2026-0142?
Same InferenceRule model as forward chain · Volume-3 Explanation ready