ID: 1778
Presenting Author: siyan Cun
Session: 598 - Engaging Communities: Digital Storytelling for Trustworthy IA
Status: pending
We employ AI (RAG, Knowledge Graphs) and a CSI to intelligently assess and optimize Environmental Permit Lists. This enhances cross-tier policy consistency and quantifies pollution-carbon synergy.
To tackle key challenges within China’s vast system of environmental permit lists—notably their overwhelming volume, qualitative focus, and the complexity of assessing completeness and consistency—this study introduces an Artificial Intelligence (AI) framework for intelligent policy evaluation and refinement.
The methodology establishes a specialized industry knowledge base and utilizes a Retrieval-Augmented Generation (RAG) model that incorporates rerank optimization. This approach successfully mitigates the low accuracy issue inherent when direct Large Language Models (LLMs) attempt to extract specific technical measures from the policy text.
The research evaluates the policy outcomes across three critical dimensions: 1. Policy Consistency: A Knowledge Graph is employed for the formalization of policy stipulations, enabling automated conflict detection and spatial validation across provincial, municipal, and unit levels. 2. List Completeness: Gaps in technical coverage are identified by systematically benchmarking the listed measures against the industry’s Best Available Techniques (BAT). 3. Synergy Quantification: A Comprehensive Synergy Index (CSI) is developed to precisely measure the integrated environmental and economic co-benefits, specifically the coordination of pollution reduction and carbon mitigation per unit of marginal cost.
Focusing on the Beijing-Tianjin-Hebei (Jing-Jin-Ji) region as a case study, this work aims to provide robust technical support for highly differentiated and precise ecological environment zoning and control implementation.
Siyan Cun is a Master's candidate at the School of Environment, Tsinghua University, China. Her research focuses on integrating Artificial Intelligence (AI) with ecological zoning and control policies
Coauthor 1: mai Su