HugRAG: Hierarchical Causal Knowledge Graph Design for RAG

2026-02-04

AI Summary

This work addresses critical limitations in existing graph-based retrieval-augmented generation (RAG) approaches, which rely on superficial node matching and lack explicit causal modeling, often yielding unfaithful answers. Furthermore, their modular graph structures create information silos that hinder cross-module reasoning and scalability. To overcome these issues, the authors propose a hierarchical causal knowledge graph that explicitly encodes causal relationships through a causal gating mechanism, effectively suppressing spurious correlations. By integrating graph neural networks with retrieval-augmented generation, the framework enables global, structured causal reasoning. Experimental results demonstrate that the proposed method significantly outperforms current graph RAG baselines across multiple datasets, achieving substantial improvements in answer faithfulness, reasoning capability, and scalability.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Reasoning under Uncertainty: Causality
• Knowledge Representation and Reasoning: Action, Change, and Causality

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Retrieval augmented generation (RAG) has enhanced large language models by enabling access to external knowledge, with graph-based RAG emerging as a powerful paradigm for structured retrieval and reasoning. However, existing graph-based methods often over-rely on surface-level node matching and lack explicit causal modeling, leading to unfaithful or spurious answers. Prior attempts to incorporate causality are typically limited to local or single-document contexts and also suffer from information isolation that arises from modular graph structures, which hinders scalability and cross-module causal reasoning. To address these challenges, we propose HugRAG, a framework that rethinks knowledge organization for graph-based RAG through causal gating across hierarchical modules. HugRAG explicitly models causal relationships to suppress spurious correlations while enabling scalable reasoning over large-scale knowledge graphs. Extensive experiments demonstrate that HugRAG consistently outperforms competitive graph-based RAG baselines across multiple datasets and evaluation metrics. Our work establishes a principled foundation for structured, scalable, and causally grounded RAG systems.

Problem

• Retrieval Augmented Generation
• Causal Reasoning
• Knowledge Graph
• Graph-based RAG
• Scalability

Innovation

• causal reasoning
• hierarchical knowledge graph
• retrieval-augmented generation
• graph-based RAG
• causal gating