The Year in AI Papers
The year/Independent research

Paper 2605.29250

OmniRetrieval: Unified Retrieval across Heterogeneous Knowledge Sources

Jinheon Baek, Soyeong Jeong, Sangwoo Park, Woongyeong Yeo, Minki Kang, Patara Trirat, Heejun Lee, Sung Ju Hwang

Published
May 2026
Research lab
Independent
Citations
0
GitHub
33 stars

01 In brief

Summary

OmniRetrieval is a framework for unified retrieval across heterogeneous knowledge sources (unstructured text, relational databases, RDF knowledge graphs, and labeled property graphs).

Instead of collapsing sources into a shared representation, it takes a natural-language query, selects relevant sources via a long-context LLM reading all source descriptors, formulates native queries (SQL, SPARQL, Cypher, or free-form text) per source, executes them, and consolidates results via cross-source evidence selection.

Evaluated on 13 datasets and 309 knowledge bases, it outperforms single-backend baselines and KB Routing across five LLM backbones (e.g., average retrieval accuracy 44.34% vs.

39.98% for KB Routing).

Analyses show that broad exploration at source selection with deferred commitment to evidence selection drives gains; larger candidate lists improve retrieval but expose evidence selection as a bottleneck.

Unified-representation methods are infeasible at scale and underperform when constrained.

The framework preserves each source's structural operators and allows adding sources by registration alone.

02 From the paper

Abstract

Real-world information needs require access to structurally diverse knowledge sources, from unstructured text and relational tables to knowledge graphs and property graphs. Existing retrievers, however, operate over one source at a time under a fixed query language, leaving the broader landscape of available knowledge fragmented behind incompatible interfaces. A natural attempt at unification would collapse these sources into a shared space, but this erases the structural affordances (such as schemas, ontologies, compositional operators) that give each source its expressive power. Effective retrieval over diverse knowledge, therefore, requires not homogenization but an overarching layer that meets each source on its own terms. To achieve this, we present OmniRetrieval, a framework that takes any natural-language query, identifies appropriate knowledge sources, and dispatches source-native queries to their native execution engines. Across an extensive benchmark spanning 13 datasets and 309 distinct knowledge bases over text, relational, and graph-structured sources, OmniRetrieval exceeds single-source baselines, demonstrating that it can serve as a general-purpose interface to the heterogeneous sources while preserving the structural distinctions that make each source valuable.