GSI Agent: Domain Knowledge Enhancement for Large Language Models in Green Stormwater Infrastructure

TL;DR

This paper introduces GSI Agent, a framework that enhances large language models with domain-specific knowledge for better performance in green stormwater infrastructure tasks, combining fine-tuning, retrieval, and reasoning strategies.

Contribution

The paper presents a novel domain-enhanced LLM framework for GSI, integrating supervised fine-tuning, retrieval-augmented generation, and an agent-based reasoning pipeline.

Findings

Significant improvement in GSI task performance (BLEU-4 from 0.090 to 0.307)

Maintains general knowledge capabilities (performance stable on common datasets)

Constructed a new GSI dataset aligned with real-world scenarios

Abstract

Green Stormwater Infrastructure (GSI) systems, such as permeable pavement, rain gardens, and bioretention facilities, require continuous inspection and maintenance to ensure long-term performance. However, domain knowledge about GSI is often scattered across municipal manuals, regulatory documents, and inspection forms. As a result, non-expert users and maintenance staff may struggle to obtain reliable and actionable guidance from field observations. Although Large Language Models (LLMs) have demonstrated strong general reasoning and language generation capabilities, they often lack domain-specific knowledge and may produce inaccurate or hallucinated answers in engineering scenarios. This limitation restricts their direct application to professional infrastructure tasks. In this paper, we propose GSI Agent, a domain-enhanced LLM framework designed to improve performance in GSI-related tasks. Our approach integrates three complementary strategies: (1) supervised fine-tuning (SFT) on a curated GSI instruction dataset, (2) retrieval-augmented generation (RAG) over an internal GSI knowledge base constructed from municipal documents, and (3) an agent-based reasoning pipeline that coordinates retrieval, context integration, and structured response generation. We also construct a new GSI Dataset aligned with real-world GSI inspection and maintenance scenarios. Experimental results show that our framework significantly improves domain-specific performance while maintaining general knowledge capability. On the GSI dataset, BLEU-4 improves from 0.090 to 0.307, while performance on the common knowledge dataset remains stable (0.304 vs. 0.305). These results demonstrate that systematic domain knowledge enhancement can effectively adapt general-purpose LLMs to professional infrastructure applications.