BEGIN:VCALENDAR VERSION:2.0 PRODID:researchseminars.org CALSCALE:GREGORIAN X-WR-CALNAME:researchseminars.org BEGIN:VEVENT SUMMARY:Armita KazemiNajafabadi (Northeastern University) DTSTART:20251110T140000Z DTEND:20251110T150000Z DTSTAMP:20260404T110742Z UID:IEEECSS_TCSP/2 DESCRIPTION:Title: Adversarial Strategies Against Multi-Agent AI Defenses in Cybe r-Physical Networks\nby Armita KazemiNajafabadi (Northeastern Universi ty) as part of Rising Star Symposium on Cyber-Physical Systems Security\, Resilience\, and Privacy\n\n\nAbstract\nMulti-agent reinforcement learning (MARL) has emerged as a promising approach for adaptive and scalable cybe r defense\, enabling distributed agents to learn coordinated defense polic ies in complex and uncertain network environments. Its ability to adaptive ly optimize decisions across multiple agents makes MARL appealing for defe nding against evolving cyber threats. In a broader area of artificial inte lligence\, existing adversarial examples have shown that even sophisticate d models can be misled\, causing classifiers or learned policies to make i ncorrect decisions. This raises important questions for cyber defense: cou ld similar adversarial strategies undermine MARL-based defenders\, and how resilient are they when facing intelligent and coordinated deception? \n\ nIn our work\, we investigate the design of AI-powered adversaries that ch allenge MARL defense policies in distributed network environments. By mode ling defenders’ interactions as a decentralized decision-making process under uncertainty\, we develop new classes of adversarial strategies that intelligently manipulate feedback\, disrupt information flow\, and interfe re with coordination—operating under both resource and stealth constrain ts to systematically degrade groups of AI decision makers. Our findings re veal critical gaps in MARL defense mechanisms and motivate next-generation security frameworks that explicitly account for deception-aware adversari es\, advancing the robustness of AI-driven cyber defense in dynamic and ad versarial environments.\n LOCATION:https://stable.researchseminars.org/talk/IEEECSS_TCSP/2/ END:VEVENT BEGIN:VEVENT SUMMARY:Sibasis Padhi (Walmart Global Tech) DTSTART:20251121T140000Z DTEND:20251121T150000Z DTSTAMP:20260404T110742Z UID:IEEECSS_TCSP/3 DESCRIPTION:Title: Agentic AI for Secure and Resilient FinTech Microservices: An Industry Perspective\nby Sibasis Padhi (Walmart Global Tech) as part o f Rising Star Symposium on Cyber-Physical Systems Security\, Resilience\, and Privacy\n\n\nAbstract\nThis talk explores how principles of cyber-phys ical system (CPS) security and resilience can be effectively translated in to the design of secure\, autonomous\, and self-healing microservices with in the FinTech sector. Drawing from real-world industry experience\, I wil l demonstrate how agentic AI\, performance-aware microservices\, and zero- trust principles can be used to safeguard high-volume financial transactio ns from disruptions\, anomalies\, and threats. The session will cover prac tical architectures for self-tuning systems\, discuss failure domains in d istributed transaction systems\, and present secure-by-design strategies f or cloud-native applications. By aligning industry-grade systems with CPS- like resilience models\, this talk aims to foster discussion between theor etical control systems security and its real-world application in modern d igital finance infrastructure.\n LOCATION:https://stable.researchseminars.org/talk/IEEECSS_TCSP/3/ END:VEVENT BEGIN:VEVENT SUMMARY:Ya-Ting Yang (New York University) DTSTART:20251125T140000Z DTEND:20251125T150000Z DTSTAMP:20260404T110742Z UID:IEEECSS_TCSP/4 DESCRIPTION:Title: Cross-Layered Design for Security and Resilience in AI-Driven Cyber Physical Human Systems\nby Ya-Ting Yang (New York University) as part of Rising Star Symposium on Cyber-Physical Systems Security\, Resili ence\, and Privacy\n\n\nAbstract\nModern societies increasingly rely on AI -driven cyber-physical-human systems (CPHSs)\, such as intelligent transpo rtation\, industrial automation\, and other critical infrastructure. While these systems promise efficiency and intelligence\, they also introduce n ew vulnerabilities where security\, privacy\, and resilience are tightly c oupled with human trust. A central question arises: how can we design soci o-technical systems that remain trustworthy and resilient even in the pres ence of adversarial manipulation and the cognitive biases of human decisio n-makers? In this talk\, we will present a research agenda that develops principled and computationally tractable frameworks for understanding trus t in CPHSs. We will walk through four key perspectives: assessing trust vi a meta-game analysis of human–CPS interactions\, building trust in AI th rough crowd auditing and accountability mechanisms\, exploiting trust in a dversaries through defensive deception\, and maintaining user trust under misinformation with information design strategies. These frameworks will b e illustrated through case studies in critical CPHS domains. We will concl ude by outlining future directions toward resilient\, cognitive-aware CPHS s.\n LOCATION:https://stable.researchseminars.org/talk/IEEECSS_TCSP/4/ END:VEVENT BEGIN:VEVENT SUMMARY:Peng Wu (Northeastern University) DTSTART:20251217T140000Z DTEND:20251217T150000Z DTSTAMP:20260404T110742Z UID:IEEECSS_TCSP/5 DESCRIPTION:Title: Bayesian Data Fusion for Distributed Learning\nby Peng Wu (Northeastern University) as part of Rising Star Symposium on Cyber-Physic al Systems Security\, Resilience\, and Privacy\n\n\nAbstract\nBayesian dat a fusion offers a principled route to distributed learning under privacy a nd uncertainty. This talk develops a unifying framework that clarifies how local beliefs should be combined when priors are shared. We analyze the C onditionally Independent Likelihood (CIL) and Conditionally Independent Po sterior (CIP) rules\, identify the prior double-counting pitfall in naïve posterior multiplication\, and derive corrections that preserve coherence while characterizing accuracy as a function of client count and prior inf ormativeness\, beyond Gaussian models. Building on this foundation\, we in troduce federated posterior sharing for multi-agent systems\, in which age nts exchange posteriors rather than data to construct a global belief and act. The method supports single-shot or periodic synchronization\, avoids prior reuse\, and improves reward and sample efficiency under uncertainty and heterogeneity. Finally\, we present a Bayesian formulation of clustere d federated learning that treats client–cluster assignment as latent dat a association\, yielding practical approximations that handle non-IID feat ure and label skew and outperform standard clustered FL. Together\, these results provide a coherent recipe—fuse beliefs\, correct for shared prio rs\, and quantify uncertainty—for privacy-preserving learning and decisi on making at scale.\n LOCATION:https://stable.researchseminars.org/talk/IEEECSS_TCSP/5/ END:VEVENT BEGIN:VEVENT SUMMARY:Hanjiang Hu (Carnegie Mellon University) DTSTART:20260113T140000Z DTEND:20260113T150000Z DTSTAMP:20260404T110742Z UID:IEEECSS_TCSP/6 DESCRIPTION:Title: Verified Safety with Neural Barrier Functions: From Dynamical Systems to Language Models\nby Hanjiang Hu (Carnegie Mellon University ) as part of Rising Star Symposium on Cyber-Physical Systems Security\, Re silience\, and Privacy\n\nAbstract: TBA\n LOCATION:https://stable.researchseminars.org/talk/IEEECSS_TCSP/6/ END:VEVENT BEGIN:VEVENT SUMMARY:Joowon Lee (Seoul National University) DTSTART:20260203T140000Z DTEND:20260203T150000Z DTSTAMP:20260404T110742Z UID:IEEECSS_TCSP/7 DESCRIPTION:Title: Design of Controllers Having Integer Coefficients for Encrypte d Control\nby Joowon Lee (Seoul National University) as part of Rising Star Symposium on Cyber-Physical Systems Security\, Resilience\, and Priv acy\n\nAbstract: TBA\n LOCATION:https://stable.researchseminars.org/talk/IEEECSS_TCSP/7/ END:VEVENT BEGIN:VEVENT SUMMARY:Helena Calatrava (Northeastern University) DTSTART:20260317T130000Z DTEND:20260317T140000Z DTSTAMP:20260404T110742Z UID:IEEECSS_TCSP/8 DESCRIPTION:Title: Statistical Signal Processing for Resilient Positioning and Tr acking\nby Helena Calatrava (Northeastern University) as part of Risin g Star Symposium on Cyber-Physical Systems Security\, Resilience\, and Pri vacy\n\nAbstract: TBA\n LOCATION:https://stable.researchseminars.org/talk/IEEECSS_TCSP/8/ END:VEVENT BEGIN:VEVENT SUMMARY:Kushal Chakrabarti (Tata Consultancy Services Private Ltd) DTSTART:20260217T140000Z DTEND:20260217T150000Z DTSTAMP:20260404T110742Z UID:IEEECSS_TCSP/9 DESCRIPTION:Title: Who Owns the Model? Protecting Model Confidentiality in Federa ted Learning Against Eavesdroppers\nby Kushal Chakrabarti (Tata Consul tancy Services Private Ltd) as part of Rising Star Symposium on Cyber-Phys ical Systems Security\, Resilience\, and Privacy\n\nAbstract: TBA\n LOCATION:https://stable.researchseminars.org/talk/IEEECSS_TCSP/9/ END:VEVENT BEGIN:VEVENT SUMMARY:Krishna Muvva (University of Nebraska - Lincoln) DTSTART:20260303T160000Z DTEND:20260303T170000Z DTSTAMP:20260404T110742Z UID:IEEECSS_TCSP/10 DESCRIPTION:Title: Learn to Fly: Enabling Deep Learning based Perception & Contr ol in Aerial Robotics\nby Krishna Muvva (University of Nebraska - Linc oln) as part of Rising Star Symposium on Cyber-Physical Systems Security\, Resilience\, and Privacy\n\nAbstract: TBA\n LOCATION:https://stable.researchseminars.org/talk/IEEECSS_TCSP/10/ END:VEVENT BEGIN:VEVENT SUMMARY:Abbas Yazdinejad (University of Regina) DTSTART:20260331T140000Z DTEND:20260331T150000Z DTSTAMP:20260404T110742Z UID:IEEECSS_TCSP/11 DESCRIPTION:Title: Toward Human-Aware Autonomous Cyber Defense: Cognitive–Phys iological Intelligence for Adaptive Security Operations\nby Abbas Yazd inejad (University of Regina) as part of Rising Star Symposium on Cyber-Ph ysical Systems Security\, Resilience\, and Privacy\n\nAbstract: TBA\n LOCATION:https://stable.researchseminars.org/talk/IEEECSS_TCSP/11/ END:VEVENT END:VCALENDAR