Research

AI research systems for cultural knowledge.

CogLab research products turn PhD questions into working systems: Kartvelian knowledge graphs, source-grounded historical AI, entity linking, and provenance-aware data infrastructure.

Research

Developed demo

Kartvelian Knowledge Universe

An immersive historical graph interface for people, places, events, inscriptions, manuscripts, institutions, and corpora from AuraDB.

Neo4j AuraKartvelian DHGraph UI
Research concept

Graph-Grounded Historian

A historical AI assistant that answers only through visible graph paths, source citations, confidence notes, and human-reviewable evidence.

GraphRAGCitationsExplainable AI
In Development
Under development

Historical Georgian Entity Linking

A workflow for linking people, places, dates, manuscript references, orthographic variants, and named entities across Georgian sources.

NERDisambiguationReview states
In Development
Under development

Source Ingestion Pipeline

A provenance-aware import layer for corpora, manuscripts, inscriptions, prosopography, and bibliographic datasets into Neo4j.

ProvenanceETLTEI/EpiDoc
In Development

Next research work

The Research button now has a home. The next step is to make each product line independently explorable and deployable, with a shared graph schema behind every product.

01

Stabilize graph schema

Define common fields for source, date, confidence, provenance, and entity aliases.

02

Move demo toward WebGL

Rebuild the current SVG universe as a Three.js experience after data fields stabilize.

03

Research workflows

Add question-to-evidence workflows for PhD notes, citations, and graph-grounded answers.

Contact

For research collaboration and PhD-aligned project conversations, connect with Cognitive Systems Lab through LinkedIn.