捆绑调教学术报告：从自动化到韧性：在人工智能时代重新思考海事安全-捆绑调教

Zaili Yang教授是英国利物浦约翰穆尔斯大学（LJMU）的海洋运输学教授。Yang教授的研究兴趣是分析并建模运输网络的安全性、弹性和可持续性，特别是海洋和物流系统。Yang教授已获得超过1400万英镑的外部资助（其中800万英镑由他作为项目负责人获得），这些资助来自欧盟、英国EPSRC和英国DTI，用于支持他的研究，包括一项著名的ERC整合者资助和一项英国EPSRC PBIAA资助。杨教授已成功完成12个博士后项目和47个博士项目。他目前在海洋安全、物流运营和港口优化等领域有5名博士后研究人员和11名博士生在他的指导下进行研究。他的研究结果已发表在450篇同行评审的论文中，这些论文涉及风险和供应链领域，包括5本书和250篇被SCI/SSCI引用的期刊论文。Yang教授是14种交通/海洋期刊的编辑委员会成员（例如《运输研究E辑》）。他还曾担任美国、英国、欧盟、加拿大、挪威和中国国家研究委员会的评审委员会成员。他已获得20项论文奖和5项研究奖（例如2018年东北亚物流奖）。

Zaili Yang is Professor of Maritime Transport at Liverpool John Moores University (LJMU), UK. Prof. Yang’s research interests are analysis and modelling of safety, resilience and sustainability of transport networks, particularly maritime and logistics systems. Prof. Yang has received more than £14m external grants (£8m as the PI) from the EU, UK EPSRC and UK DTI to support his research, including a prestigious ERC consolidator grant and a UK EPSRC PBIAA grant. Prof. Yang has successfully completed 12 postdoctoral and 47 PhD projects. He currently has 5 postdoctoral researchers and 11 PhD students under his supervision in the research areas of maritime safety, logistics operation and port optimisation. His research findings have been published in 450 refereed papers in risk and supply chain areas, including 5 books and 250 SCI/SSCI cited journal papers. Prof. Yang is an editorial board member of 14 transport/maritime journals (e.g. Transport Research Part E). He has also served as a member of review boards for national research councils of USA, UK, EU, Canada, Norway and China. He has received 20 paper awards and 5 research awards (e.g. Northeast Asia Logistics Award 2018).

报告内容简介：

人工智能与自主化技术正在重塑海上安全与运营格局。本次主题演讲首先探讨人工智能在应对气候变化、恐怖主义、网络攻击等非传统风险中的作用，重点阐释数据驱动贝叶斯网络（DDBNs）与大型语言模型（LLMs）如何协同突破传统方法的局限，通过有效管理数据不确定性将海事安全分析提升至全新高度。

尽管自主化技术为航运带来更安全的航行与更高效率，却也引发关键挑战：包括航线安全、测试场景可靠性，以及复杂人机交互中的人类可靠性及系统韧性。英国海事研究创新中心（MarRI-UK）支持的研究展望了一个以人为中心的技术进步愿景，提出融合机器学习/深度学习、复杂网络理论与心理学洞见的跨学科解决方案，以推动这一快速演进的研究前沿。

演讲继而聚焦集装箱运输网络的韧性研究。在疫情与地缘冲突等事件影响下，此类网络日益面临级联失效风险。基于欧洲研究理事会项目成果，我们构建了创新的二维韧性分析框架，该框架能捕捉多尺度网络特征与中断动态，评估从国家系统到单个港口等不同主体如何通过启动或响应连锁效应与负荷重分配，实现对外部冲击的抵御与适应。

综合这些视角，我们勾勒出人工智能驱动下海事领域的发展路径——通过自动化与数字化推动新一轮产业变革，构建更安全、智能、韧性的航运新生态。

Artificial intelligence and autonomy are reshaping maritime safety and operations. This keynote begins by exploring the role of AI in addressing non-classical risks such as climate change, terrorism, and cyberattacks. It highlights how Data-Driven Bayesian Networks (DDBNs) and Large Language Models (LLMs) can jointly deliver new solutions to managing data uncertainty and boosting maritime safety analysis to a new high level, where existing methods fail to reach.

While autonomy promises safer navigation and improved efficiency, it also introduces critical challenges in shipping, ranging from route safety and testing scenario reliability to human reliability and resilience in complex human-machine interactions. Research supported by MarRI-UK envisions a future where autonomy is not only technologically advanced but also human-centred. Multidisciplinary solutions by combining machine/deep learning, complex network theory, and psychological insights are proposed to advance this fast-evolving research frontier.

The talk then turns to the resilience of container shipping networks, increasingly vulnerable to cascading failures from events such as pandemics and geopolitical conflicts. The findings from a European Research Council project reveal a novel two-dimensional resilience analysis framework, capable of capturing multi-scale network characteristics and disruption dynamics. This framework assesses how entities from national systems to individual ports can withstand and adapt to shocks by initiating or responding to ripple effects and load redistribution.

Together, these perspectives chart a path toward an AI-driven safer, smarter, and more resilient maritime sector, where automation and digitalisation drive the next wave of transformation.

欢迎有兴趣的老师和同学参加！