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:20260404T083746Z UID:IEEECSS_TCSP/2 DESCRIPTION:Title: Adversarial Strategies Against Multi-Agent AI Defenses in Cyber-Physi cal Networks\nby Armita KazemiNajafabadi (Northeastern University) as part of Rising Star Symposium on Cyber-Physical Systems Security\, Resilie nce\, and Privacy\n\n\nAbstract\nMulti-agent reinforcement learning (MARL) has emerged as a promising approach for adaptive and scalable cyber defen se\, enabling distributed agents to learn coordinated defense policies in complex and uncertain network environments. Its ability to adaptively opti mize decisions across multiple agents makes MARL appealing for defending a gainst evolving cyber threats. In a broader area of artificial intelligenc e\, existing adversarial examples have shown that even sophisticated model s can be misled\, causing classifiers or learned policies to make incorrec t decisions. This raises important questions for cyber defense: could simi lar adversarial strategies undermine MARL-based defenders\, and how resili ent are they when facing intelligent and coordinated deception? \n\nIn our work\, we investigate the design of AI-powered adversaries that challenge MARL defense policies in distributed network environments. By modeling de fenders’ interactions as a decentralized decision-making process under u ncertainty\, we develop new classes of adversarial strategies that intelli gently manipulate feedback\, disrupt information flow\, and interfere with coordination—operating under both resource and stealth constraints to s ystematically degrade groups of AI decision makers. Our findings reveal cr itical gaps in MARL defense mechanisms and motivate next-generation securi ty frameworks that explicitly account for deception-aware adversaries\, ad vancing the robustness of AI-driven cyber defense in dynamic and adversari al environments.\n LOCATION:https://researchseminars.org/talk/IEEECSS_TCSP/2/ END:VEVENT BEGIN:VEVENT SUMMARY:Sibasis Padhi (Walmart Global Tech) DTSTART:20251121T140000Z DTEND:20251121T150000Z DTSTAMP:20260404T083746Z UID:IEEECSS_TCSP/3 DESCRIPTION:Title: Agentic AI for Secure and Resilient FinTech Microservices: An Industr y Perspective\nby Sibasis Padhi (Walmart Global Tech) as part of Risin g Star Symposium on Cyber-Physical Systems Security\, Resilience\, and Pri vacy\n\n\nAbstract\nThis talk explores how principles of cyber-physical sy stem (CPS) security and resilience can be effectively translated into the design of secure\, autonomous\, and self-healing microservices within the FinTech sector. Drawing from real-world industry experience\, I will demon strate how agentic AI\, performance-aware microservices\, and zero-trust p rinciples can be used to safeguard high-volume financial transactions from disruptions\, anomalies\, and threats. The session will cover practical a rchitectures for self-tuning systems\, discuss failure domains in distribu ted transaction systems\, and present secure-by-design strategies for clou d-native applications. By aligning industry-grade systems with CPS-like re silience models\, this talk aims to foster discussion between theoretical control systems security and its real-world application in modern digital finance infrastructure.\n LOCATION:https://researchseminars.org/talk/IEEECSS_TCSP/3/ END:VEVENT BEGIN:VEVENT SUMMARY:Ya-Ting Yang (New York University) DTSTART:20251125T140000Z DTEND:20251125T150000Z DTSTAMP:20260404T083746Z UID:IEEECSS_TCSP/4 DESCRIPTION:Title: Cross-Layered Design for Security and Resilience in AI-Driven Cyber P hysical Human Systems\nby Ya-Ting Yang (New York University) as part o f Rising Star Symposium on Cyber-Physical Systems Security\, Resilience\, and Privacy\n\n\nAbstract\nModern societies increasingly rely on AI-driven cyber-physical-human systems (CPHSs)\, such as intelligent transportation \, industrial automation\, and other critical infrastructure. While these systems promise efficiency and intelligence\, they also introduce new vuln erabilities where security\, privacy\, and resilience are tightly coupled with human trust. A central question arises: how can we design socio-techn ical systems that remain trustworthy and resilient even in the presence of adversarial manipulation and the cognitive biases of human decision-maker s? In this talk\, we will present a research agenda that develops princip led and computationally tractable frameworks for understanding trust in CP HSs. We will walk through four key perspectives: assessing trust via meta- game analysis of human–CPS interactions\, building trust in AI through c rowd auditing and accountability mechanisms\, exploiting trust in adversar ies through defensive deception\, and maintaining user trust under misinfo rmation with information design strategies. These frameworks will be illus trated through case studies in critical CPHS domains. We will conclude by outlining future directions toward resilient\, cognitive-aware CPHSs.\n LOCATION:https://researchseminars.org/talk/IEEECSS_TCSP/4/ END:VEVENT BEGIN:VEVENT SUMMARY:Peng Wu (Northeastern University) DTSTART:20251217T140000Z DTEND:20251217T150000Z DTSTAMP:20260404T083746Z UID:IEEECSS_TCSP/5 DESCRIPTION:Title: Bayesian Data Fusion for Distributed Learning\nby Peng Wu (Northe astern University) as part of Rising Star Symposium on Cyber-Physical Syst ems Security\, Resilience\, and Privacy\n\n\nAbstract\nBayesian data fusio n offers a principled route to distributed learning under privacy and unce rtainty. This talk develops a unifying framework that clarifies how local beliefs should be combined when priors are shared. We analyze the Conditio nally Independent Likelihood (CIL) and Conditionally Independent Posterior (CIP) rules\, identify the prior double-counting pitfall in naïve poster ior multiplication\, and derive corrections that preserve coherence while characterizing accuracy as a function of client count and prior informativ eness\, beyond Gaussian models. Building on this foundation\, we introduce federated posterior sharing for multi-agent systems\, in which agents exc hange posteriors rather than data to construct a global belief and act. Th e method supports single-shot or periodic synchronization\, avoids prior r euse\, and improves reward and sample efficiency under uncertainty and het erogeneity. Finally\, we present a Bayesian formulation of clustered feder ated learning that treats client–cluster assignment as latent data assoc iation\, yielding practical approximations that handle non-IID feature and label skew and outperform standard clustered FL. Together\, these results provide a coherent recipe—fuse beliefs\, correct for shared priors\, an d quantify uncertainty—for privacy-preserving learning and decision maki ng at scale.\n LOCATION:https://researchseminars.org/talk/IEEECSS_TCSP/5/ END:VEVENT BEGIN:VEVENT SUMMARY:Hanjiang Hu (Carnegie Mellon University) DTSTART:20260113T140000Z DTEND:20260113T150000Z DTSTAMP:20260404T083746Z 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 pa rt of Rising Star Symposium on Cyber-Physical Systems Security\, Resilienc e\, and Privacy\n\nAbstract: TBA\n LOCATION:https://researchseminars.org/talk/IEEECSS_TCSP/6/ END:VEVENT BEGIN:VEVENT SUMMARY:Joowon Lee (Seoul National University) DTSTART:20260203T140000Z DTEND:20260203T150000Z DTSTAMP:20260404T083746Z UID:IEEECSS_TCSP/7 DESCRIPTION:Title: Design of Controllers Having Integer Coefficients for Encrypted Contr ol\nby Joowon Lee (Seoul National University) as part of Rising Star S ymposium on Cyber-Physical Systems Security\, Resilience\, and Privacy\n\n Abstract: TBA\n LOCATION:https://researchseminars.org/talk/IEEECSS_TCSP/7/ END:VEVENT BEGIN:VEVENT SUMMARY:Helena Calatrava (Northeastern University) DTSTART:20260317T130000Z DTEND:20260317T140000Z DTSTAMP:20260404T083746Z UID:IEEECSS_TCSP/8 DESCRIPTION:Title: Statistical Signal Processing for Resilient Positioning and Tracking< /a>\nby Helena Calatrava (Northeastern University) as part of Rising Star Symposium on Cyber-Physical Systems Security\, Resilience\, and Privacy\n\ nAbstract: TBA\n LOCATION:https://researchseminars.org/talk/IEEECSS_TCSP/8/ END:VEVENT BEGIN:VEVENT SUMMARY:Kushal Chakrabarti (Tata Consultancy Services Private Ltd) DTSTART:20260217T140000Z DTEND:20260217T150000Z DTSTAMP:20260404T083746Z UID:IEEECSS_TCSP/9 DESCRIPTION:Title: Who Owns the Model? Protecting Model Confidentiality in Federated Lea rning Against Eavesdroppers\nby Kushal Chakrabarti (Tata Consultancy S ervices Private Ltd) as part of Rising Star Symposium on Cyber-Physical Sy stems Security\, Resilience\, and Privacy\n\nAbstract: TBA\n LOCATION:https://researchseminars.org/talk/IEEECSS_TCSP/9/ END:VEVENT BEGIN:VEVENT SUMMARY:Krishna Muvva (University of Nebraska - Lincoln) DTSTART:20260303T160000Z DTEND:20260303T170000Z DTSTAMP:20260404T083746Z UID:IEEECSS_TCSP/10 DESCRIPTION:Title: Learn to Fly: Enabling Deep Learning based Perception & Control in A erial Robotics\nby Krishna Muvva (University of Nebraska - Lincoln) as part of Rising Star Symposium on Cyber-Physical Systems Security\, Resili ence\, and Privacy\n\nAbstract: TBA\n LOCATION:https://researchseminars.org/talk/IEEECSS_TCSP/10/ END:VEVENT BEGIN:VEVENT SUMMARY:Abbas Yazdinejad (University of Regina) DTSTART:20260331T140000Z DTEND:20260331T150000Z DTSTAMP:20260404T083746Z UID:IEEECSS_TCSP/11 DESCRIPTION:Title: Toward Human-Aware Autonomous Cyber Defense: Cognitive–Physiologic al Intelligence for Adaptive Security Operations\nby Abbas Yazdinejad (University of Regina) as part of Rising Star Symposium on Cyber-Physical Systems Security\, Resilience\, and Privacy\n\nAbstract: TBA\n LOCATION:https://researchseminars.org/talk/IEEECSS_TCSP/11/ END:VEVENT END:VCALENDAR