Agenda

Utilities face a range of cybersecurity challenges due to increasing legacy architecture, regulatory changes, and increasing IT/OT convergence in addition to the evolving threat landscape. Siloed cybersecurity tools add to the complexity. These challenges can be addressed with a comprehensive approach, maintaining compliance with evolving regulations, and fostering a strong security culture across the organization.

As smart cities evolve, the need for resilient and sustainable energy infrastructure becomes critical. This talk explores new research in cooperative energy systems, focusing on collaboration among distributed energy resources (DERs), critical buildings, and human stakeholders. By applying recent advances in Cooperative AI, we demonstrate how intelligent systems can collectively manage energy resources to optimize social value and enhance energy security.

We examine the application of multi-agent reinforcement learning and game-theoretic approaches, such as regret minimization, in facilitating energy resource sharing and pooling, particularly during emergencies like power outages or cyberattacks. These methods enable systems to learn optimal cooperation strategies, balancing individual and collective interests.

Despite the basic premise that next-generation wireless networks (e.g., 6G) will embrace artificial intelligence (AI) integration, current efforts mostly extend existing "AI for wireless" paradigms qualitatively or incrementally. Creating AI-native wireless networks faces technical hurdles due to data-driven, training-intensive AI limitations, such as black-box models, limited reasoning and adaptability, data dependency, and energy inefficiency. In this talk, we propose a forward-looking framework grounded in causal reasoning that fosters explainable, reasoning-aware AI-native wireless networks to overcome these challenges. Further, we also discuss that incorporating neuro-symbolic AI into future wireless networking holds great promise, as it combines the understanding of the relations among intricate wireless concepts (symbolic component) with the expressive power of the neural networks. Further, we illustrate the potential of causal reasoning and neuro-symbolic AI frameworks through the example of an emerging field such as semantic communications that craft a nuanced semantic language between the communicating nodes. This language aims to compute a minimalistic and generalizable semantic representation that enhances communication efficiency by incorporating advanced reasoning components at both ends of the communication process. Finally, we touch upon the fundamental principles of developing "universal foundation models" (aka wireless specific generative AI models) that are driven by three distinct characteristics: 1) integration of multi-modal sensing data, 2) grounding sensory input via causal reasoning and retrieval-augmented generation (RAG), and 3) instructibility to environmental feedback through logical and mathematical reasoning enabled by neuro-symbolic AI. 

Deep learning (DL) models are widely employed in modeling cyber-physical and critical infrastructure systems like power grids, wireless communication and smart manufacturing, for quality control, often in contexts with computational and memory constraints. Accurate decision pipelines can be achieved by coupling the memory constrained (edge) context with less constrained fog and cloud contexts enabling the operation of larger DL models. Such application contexts with multiple distributed DL models with varying representational power and computational costs are often termed `Hierarchy-of-Compute` (HoC) contexts. Effectively leveraging the multiple DL models with different costs and complexities requires an intelligent decision agent that balances task accuracy with cost. Developing such an intelligent agent enables minimizing decision cost and maximizing decision accuracy. In this talk, we will discuss the design of an intelligent decision pipeline in HoC contexts for smart manufacturing and wireless communication.

In recent years, connected vehicles have gradually become integrated into our daily lives, relying on vehicular networks to generate and share diverse types of sensor data, enhancing both travel safety and efficiency. However, due to the open nature of vehicular networks, these sensing data could be erroneous, which may be caused by various reasons, ranging from an onboard device (OBD) sensor malfunctioning and reporting incorrect reading to the sensing data being tampered by a malicious vehicle. This talk will address the research question of how to enhance security and trust in vehicular networks and discuss how blockchain and AI technologies help achieve that. We will first introduce the research background regarding security and trust management, and then provide a general overview of the blockchain-and-AI-enabled trust management approach for vehicular networks.

Gamification, including game-based learning (GBL), is a recognized pedagogical approach for teaching and reinforcing cybersecurity knowledge and skills. An innovative form of GBL that is gaining popularity across various educational levels, from secondary schools to professional development, is escape room-style education. This talk will cover the design and development of escape room activities tailored for teaching web and software security concepts to undergraduate students at York College. Additionally, the potential of utilizing LLMs in the design and development of escape rooms will be discussed.

Networked Cyber-Physical Systems (N-CPS) consist of intelligent computing and control (“cyber”) elements and real-world (“physical”) subsystems, interconnected (“networked”) through various communication and physical processes. These systems are pervasive in our environment, playing an integral role in critical infrastructure, such as vehicular platoons, multi-robot systems, microgrids and supply chain networks, as well as in complex processes like biochemical reactions, interacting populations and epidemic control. Therefore, careful design of such N-CPS is paramount, mainly to promote their overall stability, robustness and resilience. To this end, distributed controllers and the network topologies (communication/physical) associated with N-CPS can be treated as design components. Traditionally, the design of controllers and topologies has been approached as separate problems in the literature, often under restrictive assumptions about the other component. However, research on co-designing controllers and topologies remains relatively sparse, primarily due to the increased complexity introduced by considering both design components simultaneously. 

In my research, this challenge is attacked from two fronts: decentralization and dissipativity. In particular, I have developed an innovative framework for the efficient co-design of controllers and topology in N-CPS. This co-design framework is not only computationally efficient but also operationally compositional, enabling it to handle large-scale changes in N-CPS efficiently, thus enhancing resilience. Moreover, the proposed co-design framework not only enforces stability but also optimizes the impact of disturbances on the performance of N-CPS. Overall, the developed co-design framework provides insightful control and topology solutions and is broadly applicable across a wide range of N-CPS, thanks to the generality of the decentralization and dissipativity concepts employed. In this talk, I will start by introducing the decentralization and dissipativity concepts, followed by a discussion of the proposed co-design framework and its application to various N-CPS scenarios. I will conclude by highlighting key results, including the co-design framework’s performance, computational efficiency, and the enhanced robustness and resilience it offers.