UltRAG: A universal simple scalable recipe for knowledge graph RAG

Dobrik Georgiev Georgiev, Kheeran K. Naidu, Alberto Cattaneo, Federico Monti, Carlo Luschi, Daniel Justus

Keywords: factuality, knowledge graphs, RAG, KGRAG, LLM, GraphRAG, graph reasoning, NAR, GNN, GFM, Graph Neural Networks, Graph Foundational Models

TL;DR: We show executing queries against a knowledge graph can be quite competitive provided that your executor is neural.

Abstract: Large language models (LLMs) frequently generate confident yet factually incorrect content when used for language generation (a phenomenon often known as _hallucination_). Retrieval augmented generation (RAG) tries to reduce factual errors by identifying information in a knowledge corpus and putting it in the context window of the model. While this approach is well-established for document-structured data, it is non-trivial to adapt it for Knowledge Graphs (KGs), especially for queries that require multi-node/multi-hop reasoning on graphs. We introduce UltRAG, a general framework for retrieving information from Knowledge Graphs that shifts away from classical RAG. By endowing LLMs with off-the-shelf _neural_ query executing modules, we highlight how readily available language models can achieve state-of-the-art results on Knowledge Graph Question Answering (KGQA) tasks _without any retraining of the LLM or executor involved_. In our experiments, UltRAG achieves better performance when compared to state-of-the-art KG-RAG solutions, and it enables language models to interface with Wikidata-scale graphs (116M entities, 1.6B relations) at comparable or lower costs.

Submission Number: 28