D-ISN: TRACK 1: Characterizing the Global Illicit Trade in Energy-Critical Materials using Machine Learning, Remote Sensing, and Qualitative Research

这项为期五年的破坏非法供应网络（D-ISN）项目的目标是绘制和描述关键能源矿物（ECM）供应链中非法来源材料的数量。ecm对可再生能源、核能和化石能源发电至关重要，被列入美国政府列出的35种“关键材料”清单，但它们的供应链仍然不透明，容易受到非法活动的影响。目前还没有对ECM贸易流量的合法-非法组成或其随时间演变的全球测量。为了解决这一问题，本项目试图根据几个来源国、过境国和目的地国的原始研究，绘制全球ECM流动地图并建立模型。本研究的发现和开发的工具将改善非法来源ECM的发现和可追溯性，确定几个ECM供应链上的弱点，并生成其动态预测模型，以确定有效的中断策略。其结果将以从开放和专有数据集中获得的数据以及在国家和国家以下各级收集的数据为依据，并在广泛的实地研究中进行测试。该项目将提高我们国家识别能源关键材料供应链脆弱性的能力，促进改变游戏规则的创新技术工具和多学科方法，从而提高检测和破坏影响ECM供应链的非法活动的能力。这是确保透明、安全和可持续供应链的关键一步。该项目整合了机器学习、遥感和多语言定性研究等方法，通过描述钴、锂、铌、铂族金属（PGM）、稀土元素（REE）和钽等几种ecm的全球非法特征，检测和破坏非法活动。利用远耦合框架内的复杂系统理论，该项目将(1)描述和模拟合法和非法ECM贸易流交汇的过程，并分析自2000年以来相关的贸易、政策和土地使用动态；(2)查明贸易违规行为的来源；(3)利用遥感、人类专家分析和人工智能计算官方特许经营范围以外的采矿情况；(4)利用人在环（human-in-the-loop）机器学习建立非法国际贸易流量与当地土地利用变化相关性的预测模型，然后利用人文地理学的定性方法进行因果机制分析，以确定非法ECM贸易流量变化的“方式和原因”。该项目解决了一个重大缺口，并将通过提供全球评估来产生可操作的科学：供应链分析所依据的官方贸易数据的可信度；非法商品在维持全球ECM流动方面的作用，包括其在特定目的地市场的相对发生率；在可测绘的采矿活动范围内发生的非法采矿活动的程度。该项目通过融合科学方法推进科学理论，将社会、物理、数据和计算机科学置于同等地位，以形成非法ECM流动的完整图景，并开发了首个同类多模式方法，以确定非法ECM贸易流动与全球范围内特定非法生产地点之间的联系。在这个项目下所制订的方法将能够侦测非法来源的ECM，将适用于其他矿物商品的流动，并将为继续研究非法矿物生产和贸易的复杂动态提供坚实的基础。该项目由IIS和促进竞争研究的既定计划（EPSCoR）共同资助。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The objective of this five-year Disrupting Operations of Illicit Supply Networks (D-ISN) project is to map and characterize the volume of illicitly-sourced materials in energy-critical minerals (ECM) supply chains. ECMs are essential to renewable, nuclear, and fossil energy generation and are included in the US government’s list of 35 ‘critical materials’, yet their supply chains remain opaque and vulnerable to illicit activity. There are currently no global measurements of the licit-illicit composition of ECM trade flows or their evolution over time. To address the problem, this project seeks to map and model global ECM flows based on original research in several source, transit, and destination countries. The findings and tools developed under this study will improve discovery and traceability of illicitly sourced ECM, identify vulnerable points along several ECM supply chains, and generate predictive models of their dynamics in order to identify effective disruption strategies. The results will be informed by data drawn from open and proprietary datasets, as well as data gathered at national and subnational levels, and tested under extensive field research. The project will advance our national ability to identify vulnerabilities in the supply chains of energy-critical materials, catalyze game-changing innovative technological tools and multidisciplinary methodologies that enhance capacity for the detection and disruption of illicit activities affecting ECM supply chains. This is a key step for ensuring transparent, secure, and sustainable supply chains.The project integrates methods from machine learning, remote sensing, and multilingual qualitative research to detect and disrupt illicit activities by characterizing the global illicit in several ECMs: Cobalt, Lithium, Niobium, Platinum Group Metals (PGM), Rare Earth Elements (REE), and Tantalum. Using complex systems theory within a telecoupling framework, the project will (1) characterize and model the processes through which licit and illicit ECM trade flows converge and analyze linked trade, policy, and land use dynamics since 2000 (the what); (2) identify sources of trade irregularities; (3) use remote sensing, human expert analysis, and AI to calculate mining occurring beyond official concessions (the where); (4) use human-in-the-loop machine learning to generate a predictive model of correlations between illicit international trade flows and local land use change, and then use qualitative methods from human geography to conduct a causal mechanism analysis to identify “the how and the ¬why” of changing illicit ECM trade flows. The project addresses a major gap and will generate actionable science by providing a global assessment of: the credibility of official trade data upon which supply chain analyses are based; the role of illicit commodities in sustaining global ECM flows, including their relative occurrence in specific destination markets, and; the level of illicit mining activity occurring within the mappable extent of mining activities. This project advances scientific theory through a convergence science approach that places social, physical, data and computer sciences on equal footing in order to form a complete picture of illicit ECM flows, and develops a first-of-its-kind multi-modal approach to identify links between illicit ECM trade flows and specific locations of illicit production on a global scale. The methods developed under this project will enable detection of illicitly sourced ECM, will be applicable to other mineral commodity flows, and will provide a strong foundation for continuing research on the complex dynamics of illicit mineral production and trade. This project is jointly funded by IIS and the Established Program to Stimulate Competitive Research (EPSCoR).This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.