A Graph-Retrieval-Augmented Generation Framework Enhances Decision-Making in the Circular Economy
This work addresses the need for reliable, regulatory-ready decision-making in the circular economy, though it is incremental as it builds on existing RAG methods with domain-specific enhancements.
The paper tackled the problem of LLMs hallucinating industrial codes and emission factors in sustainable manufacturing by introducing CircuGraphRAG, a framework that grounds LLM outputs in a domain-specific knowledge graph, achieving ROUGE-L F1 scores up to 1.0 compared to baseline scores below 0.08 and improving efficiency by halving response time and reducing token usage by 16%.
Large language models (LLMs) hold promise for sustainable manufacturing, but often hallucinate industrial codes and emission factors, undermining regulatory and investment decisions. We introduce CircuGraphRAG, a retrieval-augmented generation (RAG) framework that grounds LLMs outputs in a domain-specific knowledge graph for the circular economy. This graph connects 117,380 industrial and waste entities with classification codes and GWP100 emission data, enabling structured multi-hop reasoning. Natural language queries are translated into SPARQL and verified subgraphs are retrieved to ensure accuracy and traceability. Compared with Standalone LLMs and Naive RAG, CircuGraphRAG achieves superior performance in single-hop and multi-hop question answering, with ROUGE-L F1 scores up to 1.0, while baseline scores below 0.08. It also improves efficiency, halving the response time and reducing token usage by 16% in representative tasks. CircuGraphRAG provides fact-checked, regulatory-ready support for circular economy planning, advancing reliable, low-carbon resource decision making.