Graph-Augmented Reasoning with Large Language Models for Tobacco Pest and Disease Management

TL;DR

This paper introduces a graph-augmented reasoning framework that enhances large language models with structured domain knowledge for more accurate and evidence-based tobacco pest and disease management recommendations.

Contribution

It presents a novel integration of domain-specific knowledge graphs with LLMs, improving reasoning and evidence retrieval in plant disease management tasks.

Findings

Significant performance improvements on multi-hop reasoning questions.

Enhanced domain consistency in generated recommendations.

Effective mitigation of hallucinations in LLM outputs.

Abstract

This paper proposes a graph-augmented reasoning framework for tobacco pest and disease management that integrates structured domain knowledge into large language models. Building on GraphRAG, we construct a domain-specific knowledge graph and retrieve query-relevant subgraphs to provide relational evidence during answer generation. The framework adopts ChatGLM as the Transformer backbone with LoRA-based parameter-efficient fine-tuning, and employs a graph neural network to learn node representations that capture symptom-disease-treatment dependencies. By explicitly modeling diseases, symptoms, pesticides, and control measures as linked entities, the system supports evidence-aware retrieval beyond surface-level text similarity. Retrieved graph evidence is incorporated into the LLM input to guide generation toward domain-consistent recommendations and to mitigate hallucinated or inappropriate treatments. Experimental results show consistent improvements over text-only baselines, with the largest gains observed on multi-hop and comparative reasoning questions that require chaining multiple relations.