SurveyG: A Multi-Agent LLM Framework with Hierarchical Citation Graph for Automated Survey Generation

2025-10-08

AI Summary

Existing survey generation methods overlook structural relationships among papers, resulting in logically incoherent content lacking evolutionary trajectories. To address this, we propose a hierarchical citation-graph-based multi-agent survey generation framework. First, we construct a multilayer knowledge graph by jointly modeling citation dependencies and semantic similarity. Second, we design cooperative agents that perform longitudinal traversal (from foundational → developmental → frontier works) and lateral thematic expansion. Finally, systematic outlines are generated via dynamic path search, iterative summary aggregation, and factual consistency verification. Compared to state-of-the-art approaches, our framework significantly improves structural coherence, knowledge coverage completeness, and domain alignment—validated through double-blind evaluations by human experts and large language models. It establishes a novel, interpretable, and traceable paradigm for automated academic survey generation.

Technology Category

• Multiagent Systems: Agent-Based Simulation and Emergent Behavior
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Web Mining and Content Analysis: Models for Web evolution

Abstract

Large language models (LLMs) are increasingly adopted for automating survey paper generation cite{wang2406autosurvey, liang2025surveyx, yan2025surveyforge,su2025benchmarking,wen2025interactivesurvey}. Existing approaches typically extract content from a large collection of related papers and prompt LLMs to summarize them directly. However, such methods often overlook the structural relationships among papers, resulting in generated surveys that lack a coherent taxonomy and a deeper contextual understanding of research progress. To address these shortcomings, we propose extbf{SurveyG}, an LLM-based agent framework that integrates extit{hierarchical citation graph}, where nodes denote research papers and edges capture both citation dependencies and semantic relatedness between their contents, thereby embedding structural and contextual knowledge into the survey generation process. The graph is organized into three layers: extbf{Foundation}, extbf{Development}, and extbf{Frontier}, to capture the evolution of research from seminal works to incremental advances and emerging directions. By combining horizontal search within layers and vertical depth traversal across layers, the agent produces multi-level summaries, which are consolidated into a structured survey outline. A multi-agent validation stage then ensures consistency, coverage, and factual accuracy in generating the final survey. Experiments, including evaluations by human experts and LLM-as-a-judge, demonstrate that SurveyG outperforms state-of-the-art frameworks, producing surveys that are more comprehensive and better structured to the underlying knowledge taxonomy of a field.

Problem

• Automating survey generation with structural paper relationships
• Addressing incoherent taxonomy in existing LLM-based survey methods
• Capturing research evolution through hierarchical citation graphs

Innovation

• Multi-agent LLM framework for automated survey generation
• Hierarchical citation graph embedding structural knowledge
• Multi-level summaries with validation for comprehensive surveys