WorldDB : A Vector Graph-of-Worlds Memory Engine with Ontology-Aware Write-Time Reconciliation
Harish Santhanalakshmi Ganesan · 2026
Read WorldDB : A Vector Graph-of-Worlds Memory Engine with Ontology-Aware Write-Time Reconciliation by Harish Santhanalakshmi Ganesan (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-18478.txt.
Description
Ganesan presents WorldDB, a memory engine that treats every node as a nested world with its own subgraph and ontology scope. Flat vector stores fragment facts and lose identity across sessions, while flat knowledge graphs lack recursive composition. WorldDB makes nodes content-addressed and immutable, giving a Merkle-style audit trail, and turns edges into write-time programs that handle supersession, contradiction, and merges. On LongMemEval-s, it reaches 96.4 percent overall accuracy with Claude Opus 4.7 answering, beating the prior best system.
Catalog record
| CREATOR | Harish Santhanalakshmi Ganesan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 40385 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.18478