KGFR: A Foundation Retriever for Generalized Knowledge Graph Question Answering

TL;DR

This paper introduces KGFR, a scalable and generalizable knowledge graph retriever that collaborates with LLMs to improve reasoning on knowledge-intensive questions, especially on unseen graphs.

Contribution

The paper presents a novel LLM-KGFR framework that encodes relations with descriptions, initializes entities based on questions, and employs APP for efficient large graph retrieval, enabling zero-shot generalization.

Findings

Achieves strong performance on KG reasoning tasks.

Maintains scalability on large, unseen graphs.

Supports controllable, iterative reasoning with LLMs.

Abstract

Large language models (LLMs) excel at reasoning but struggle with knowledge-intensive questions due to limited context and parametric knowledge. However, existing methods that rely on finetuned LLMs or GNN retrievers are limited by dataset-specific tuning and scalability on large or unseen graphs. We propose the LLM-KGFR collaborative framework, where an LLM works with a structured retriever, the Knowledge Graph Foundation Retriever (KGFR). KGFR encodes relations using LLM-generated descriptions and initializes entities based on their roles in the question, enabling zero-shot generalization to unseen KGs. To handle large graphs efficiently, it employs Asymmetric Progressive Propagation (APP)- a stepwise expansion that selectively limits high-degree nodes while retaining informative paths. Through node-, edge-, and path-level interfaces, the LLM iteratively requests candidate answers, supporting facts, and reasoning paths, forming a controllable reasoning loop. Experiments demonstrate that LLM-KGFR achieves strong performance while maintaining scalability and generalization, providing a practical solution for KG-augmented reasoning.