D-ISN: An AI-augmented Framework to Detect, Disrupt, and Dismantle Opioid Trafficking Networks

该项目旨在设计和开发一个人工智能（AI）增强框架，以检测、破坏和拆除阿片类药物贩运网络，从而加强公共安全、健康和福利。阿片类药物贩运是一场严重的国家危机，由于使用现代技术而成为可能。在暗网和流行的社交媒体平台上托管的非法毒品市场已成为阿片类药物交易的重要媒介，阿片类药物贩运也与虚拟产品交易等其他贩运网络混在一起。该项目采用以系统为重点的整体框架，以更好地了解这些贩运网络的动态和运作，并将促进积极主动的应对战略。本项目开发的人工智能增强框架将能够系统地汇总和分析由暗网和社交媒体平台托管的非法毒品市场产生的大量数据。具体而言，该项目有四个研究目标：(1)开发新的深度图学习技术，对多源、多模态数据进行全面建模，并随时间演变，用于在线阿片类药物贩运者检测；(2)通过新的双重学习模型揭示阿片类药物贩运网络的组织结构（如贩毒集团、关键参与者）；(3)发展新的可解释的贸易流推理，以预测阿片类药物贸易路线并了解运作模式；(4)开发一种新的自适应强化学习范式，使专家（例如，医疗保健/行业合作伙伴、政府和执法机构）能够促进阻断阿片类药物贩运网络的互动过程。该研究将通过一种融合的方法来推进科学理论，该方法将部署计算机和数据科学、工程学、公共卫生和社会科学，以解决国家阿片类药物危机。该项目的成果将向公众开放并广泛分发。该项目将通过新课程开发、学生指导活动和代表性不足群体的参与，将研究与教育结合起来，培养后代掌握预防阿片类药物的跨学科研究方法。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。

This Disrupting Operations of Illicit Supply Networks (D-ISN) project aims to design and develop an artificial intelligence (AI)-augmented framework to detect, disrupt, and dismantle opioid trafficking networks, thereby enhancing public safety, health and welfare. Opioid trafficking is a serious national crisis that has been enabled by the use of modern technologies. Illicit drug markets hosted in darknet and popular social media platforms have emerged as important mediums for trading opioids, and opioid trafficking has also co-mingled with other trafficking networks such as virtual product trade. This project employs a holistic, systems-focused framework to better understand the dynamics and operations of these trafficking networks and will facilitate proactive response strategies. The AI-augmented framework developed in this project will enable systematic aggregation and analysis of the large volume of data generated from both illicit drug markets hosted in darknet and social media platforms. Specially, this project has four research objectives: (1) develop novel deep graph learning techniques to comprehensively model multi-source, multi-modal data with evolution over time for online opioid trafficker detection; (2) uncover the organizational structures (e.g., drug cartel, key players) of opioid trafficking networks through a new dual learning model; (3) develop novel interpretable trade flow reasoning to predict opioid trade routes and understand operational patterns; and (4) develop a novel adaptive reinforcement learning paradigm to enable expert-in-the-loop (e.g., healthcare/industry partners, government and law enforcement agencies) to facilitate an interactive process for opioid trafficking network interdiction. The research will advance scientific theory through a convergent approach that deploys computer and data sciences, engineering, public health, and social science to address the national opioid crisis. The outcomes of this project will be made publicly accessible and broadly distributed. The project will integrate research with education through novel curriculum development, student mentoring activities, and participation of underrepresented groups to train future generations in interdisciplinary research methodologies for opioid prevention.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.