KUMARI
Knowledge Operating System for Tamil Civilizational Heritage
What It Is
A living world model and knowledge operating system for documented Tamil heritage sites. Phase 2 extension of THIRUVAASAL — generalizing single-site probabilistic reconstruction into a multi-site, evidence-aware, versioned, queryable infrastructure. Exposes evidence submissions, reconstruction posteriors, provenance, temporal slices, and confidence fields as API contracts before any generation happens.
The Problem
Cultural-heritage reconstruction efforts are either institutional (paywalled, brittle, one-off) or generative (unbounded, unverifiable). There is no shared knowledge infrastructure that exposes evidence + uncertainty + provenance + rights state across multiple sites with a query model that heritage reviewers can audit.
Approach
- Evidence API schema — what counts as evidence, at what tier, with what rights status
- Reconstruction API response schema — exposes posterior distributions, marginals, uncertainty, and provenance per element
- Query model — evidence, sites, elements, posterior state, temporal slices, version diffs
- Versioning and provenance metadata — every reconstruction state is addressable and auditable
- Rights and review workflow states — explicit institutional lifecycle, not implicit
Results
- Planned — KUMARI becomes infrastructure only after THIRUVAASAL's underlying method is validated on a single site with holdout
- First implementation track: institutional data and API contracts, not generation
- Epistemic boundary: the name honors cultural memory; the method is bounded by evidence — KUMARI explicitly does NOT assert a singular vast submerged Tamil continent as geological fact