中国海关采用人工智能_Al_+和机器学习_ML_的案例研究-英文版_32页_3mb

Summary of the SMART CUSTOMS PROJECT

Core Content

The SMART CUSTOMS PROJECT is a case study on the adoption of Artificial Intelligence (AI) and Machine Learning (ML) in the General Administration of China Customs (GACC). The project is part of the World Customs Organization (WCO) initiative to identify and support the use of disruptive technologies in Customs operations. It highlights the strategic vision, implementation framework, data management, legal and policy arrangements, and real-world applications of AI and ML in China Customs.

Main Views

• The SMART CUSTOMS initiative is driven by the "3S" concept (Smart Customs, Smart Borders, Smart Connectivity), which emphasizes the integration of modern technologies and innovative thinking to enhance customs operations.
• The Smart Customs Cooperation Partnership (SCCP) is a key mechanism for fostering international collaboration and promoting the modernization of customs systems through shared knowledge and joint efforts.
• The "1 + N + X" implementation scheme outlines the strategic approach to developing and deploying AI and ML technologies across 22 operational fields and 132 specific scenarios.
• The GACC has established a comprehensive data management framework, including the development of a big data lake, intelligent model databases, and standardized data governance processes.
• A multi-level governance structure involving central departments, local offices, and field operations ensures the effective development and application of AI technologies.
• Legal and policy frameworks are in place to govern data security, intellectual property rights, and the use of AI in customs operations.

Key Information

I. Background

• The WCO and GACC signed a Cooperation Agreement in December 2023, initiating the Smart Customs Project.
• The project focuses on disruptive technologies such as AI, blockchain, IoT, and big data, aiming to improve trade security, facilitation, and efficiency.
• The study mission in China, conducted from 28 October to 1 November 2024, collected data on AI/ML implementation, including on-site visits to Huangpu and Guangzhou Customs Districts.

II. Introduction to GACC

• The General Administration of Customs of the People's Republic of China (GACC) is the central authority overseeing customs operations.
• It manages 42 customs districts and 678 customs houses, with a total of 100,000 staff.
• The GACC has integrated AI, IoT, and big data into its operations, including border health checks, inspection, and quarantine.

III. Strategy and Vision

• The "3S" concept is central to the GACC's vision for modernizing customs operations.
• Smart Customs aims to enhance control, internal management, and institutional innovation through technology.
• The SCCP is a framework for international cooperation, promoting information sharing, stakeholder collaboration, and capacity building.
• The "1 + N + X" scheme ensures a structured and scalable approach to implementing AI and ML technologies.

IV. Data Management

• The GACC has implemented a big data lake containing over 261 billion data records and 15,000 data tables.
• Data governance and sharing mechanisms have been optimized, with a focus on internal and external data integration.
• Intelligent model databases are built using advanced algorithms like Catboost, targeting areas such as fraud detection, food safety, and animal/plant quarantine.

V. Leadership and Governance

• The Risk Management Department leads the development of intelligent models and databases.
• A centralized and decentralized model is used for algorithm deployment, combining "grand unification" with "small scale autonomy".
• The Model Operational Management Centre in Huangpu Customs coordinates demand, project management, and performance evaluation.
• A lifecycle management approach is used for model development, including research, deployment, iteration, and removal.

VI. Policy Arrangements

• The "14th Five-Year Plan for Big Data Application in Customs" (2021) institutionalizes big data management.
• The "Overall Plan for Smart Customs" (2023) and "Construction Plan for Intelligent Models" outline the strategic goals and key projects for AI.
• The GACC has established a unified model algorithm interface and a research and development environment for intelligent models.

VII. Legal Framework

• National laws such as the Data Security Law, Cybersecurity Law, and Personal Information Protection Law support data governance.
• The GACC has issued regulations for customs data classification, security, and application, including data sharing, audit, and model management.
• Intellectual Property Rights (IPR) are managed by the GACC, with all AI-developed content owned by the organization.

VIII. Use Cases

• TianXuan (Merak) is an AI-based risk assessment system that has been deployed in 260 ports across China.
• Smart Image Analysis is applied to large container/vehicle scanners, passenger vehicle scanners, and CT units, enhancing automated inspection and risk detection.
• Intelligent models are used in four areas: external law enforcement, enterprise services, internal administrative management, and compliance audit.
• 24 models are currently in operation, with 20 more under development, targeting various enforcement and service areas.

IX. Technologies and Tools

• The GACC uses AI algorithms, big data platforms, cloud computing, blockchain, and IoT.
• OCR and natural language processing (NLP) are applied for document review and compliance audits.
• The "Smart Office Model" automates information collection, document drafting, and error correction.

X. Skills and Training

• A dedicated team for AI model development has been established at the central level.
• Training and capacity building are emphasized to support the technical and operational requirements of AI implementation.

XI. Stakeholder Engagement and Communication

• The GACC collaborates with government agencies, international organizations, private sector, and academia.
• Community Portals and knowledge sharing are promoted to enhance international cooperation and innovation.

XII. Costs

• The project involves significant investment in technology infrastructure, data management, and model development.
• Cost reduction is a key objective, aiming to improve trade efficiency and economic development.

XIII. Innovation

• The GACC emphasizes innovation in AI and ML applications, with a focus on automated control, data analytics, and intelligent decision-making.
• Agile project management is used to minimize deployment time and enhance model adaptability.

XIV. Future Plans

• The GACC plans to expand AI and ML applications across more customs districts and specific business scenarios.
• Continued international collaboration and capacity building are expected to drive global supply chain interconnectivity and trade facilitation.