This talk will address the critical challenge of ensuring the robustness of graph learning models by walking through a structured, three-step processing pipeline. We will begin by dissecting adversarial attacks, exploring how graph-structured data can be subtly manipulated to mislead model predictions. The discussion will then pivot to defense strategies, first covering empirical defense methods that have shown practical promise but lack formal guarantees. Subsequently, we will delve into the more rigorous field of certified...

Modern global supply chains exhibit complex, multi-tier dependencies that remain largely invisible to decision-makers, creating vulnerabilities to cascading disruptions. This presentation introduces a novel AI-driven methodology that combines data scraping, generative AI, and network science to illuminate deep-tier supply chain networks and quantify systemic risks. The approach leverages diverse data sources—including 10-K filings, earnings calls, and trade databases—to construct large-scale supply chain networks spanning over 55,000 firms, 230,000 relationships, 93 countries, and 278 industries, enabling...

Unlike previous generations of wireless networks, which were primarily designed to meet the requirements of human communications, 5G networks enable extensive data collection from machines. As we transition to 6G, the emphasis moves beyond connectivity toward leveraging this machine-generated data for a new spectrum of control applications, such as UAV swarms, collaborative robots, and cooperative autonomous vehicles. Designing communication systems for these advanced control applications introduces a distinct set of challenges. These include meeting stringent...

针对电子计算能耗墙、内存墙等瓶颈，本研究提出基于黑磷各向异性光电特性的激光驱动光子神经网络架构革新方案，实现能效与智能双突破：1、动态存算一体：利用黑磷层数可调带隙特性（0.3-2.0 eV），提出激光诱导光致荧光动态调控机制，建全光推理架构，指纹识别准确率96.34%；2、联结学习神经网络计算：通过双光子聚合效应模拟STDP学习规则，结合激光时序调控实现类脑条件反射验证，模式识别率>97%；3、跨模态低功耗计算：基于应力发光特性，构建激光-机械跨模态感知-决策系统，手势识别能耗趋零（<1nW）；4、超快光逻辑门阵列：利用黑磷非线性克尔效应，开发七种激光驱动光逻辑门（延迟<100ps，能耗<10fJ/OP）等。构建了从材料物性调控到智能光芯片集成的全链条创新体系，相关成果获广东省自然科学一等奖。未来将重点突破高密度光子集成、动态权重重构算法及多模态感知融合，为我国光计算芯片自主可控提供核心支撑。

How can we coordinate self-interested autonomous agents toward desired system outcomes? This question, long central to the field of multi-agent systems, has gained renewed urgency due to the rise of today’s intelligent agents and agentic technologies. We approach this challenge from an optimisation perspective, aiming to find ways to steer agents to optimal correlated equilibria in stochastic environments. I will introduce our recent algorithmic results for this problem. At the core of our approach is...

The number of people working remotely has increased substantially given advancements in video conferencing software.  It has now become easier for individuals to have an immersive online meeting experience, even while traveling. This is due in part to increasingly autonomous features in the car that allow the human driver more opportunities to actively engage in virtual meetings while driving. However, these virtual meetings may also divert the driver’s attention from the road at inopportune times,...

Modeling human driving behavior in highly interactive traffic scenarios is challenging due to the complex game-theoretic interactions and inherent heterogeneity of traffic participants. In this talk, Dr. Chen will first introduce his research on lane-changing (LC) scenarios. A complete LC maneuver consists of both a decision-making stage and an implementation stage. In the decision-making stage, he proposed a Deep Markov Cognitive Hierarchy Model (DMCHM) that integrates both temporal dependencies and game-theoretic processes for lane-changing decisions....

The emergence of connected and automated vehicles (CAVs) is revolutionizing transportation systems by enabling real-time information sharing among traffic participants. The shared data, e.g., photometric images and movement trajectories, offer invaluable opportunities to develop various transportation applications for smart cities. However, most applications rely on deterministic models, ignoring uncertainties in estimations. Given limited data, imperfect models, and inherent dynamics in transportation systems, such point estimators may lead to biased models and suboptimal solutions in system...

Due to the dramatic increase in high data rate services and in order to meet the demands of the sixth-generation (6G) wireless networks, researchers from both academia and industry have been exploring advanced transmission techniques, new network architectures and new frequency spectrum such as the millimeter wave, the terahertz, the infrared, and the visible light spectrum. Light Fidelity (LiFi), in particular, is an emerging technology that has been introduced as a promising solution for 6G...

Climate change and population growth pose unprecedented challenges to the transportation system. Meanwhile, the proliferation of crowdsensing techniques, such as mobile phones, vehicles, social media, and cameras, has generated vast spatiotemporal data for understanding human activities and their interaction with the external environment. Effectively managing such massive, multi-structured spatiotemporal data, extracting valuable travel information, and tailoring solutions to various transportation issues are more crucial than ever. In this talk, I will first delve into my...