"Cybersecurity and Reliability in the Era of Agentic AI" "
Distinguished speaker: Dr. Angelos Stavrou
Bio:
" Dr. Angelos Stavrou is a Virginia Tech Innovation Campus founding Professor and the Entrepreneurship activities lead. He is also a member of the Bradley Department of Electrical & Computer Engineering at Virginia Tech. Dr. Stavrou is a serial entrepreneur and the founder of Quokka, Kryptowire Labs, Aether Argus, and Impedyme Inc. Quokka is a VC-baked Mobile Security company with a more than 200M valuation. In addition, Dr. Stavrou has served as a principal investigator on research awards from NSF, DARPA, IARPA, DHS, AFOSR, ARO, ARL, and ONR. He has written more than 150 peer-reviewed conference and journal articles. Stavrou received his M.Sc. in Electrical Engineering, M.Phil., and Ph.D. (with distinction) in Computer Science, all from Columbia University. Stavrou is an Associate Editor of IEEE Transactions on Computers, IEEE Security & Privacy, and IEEE Internet Computing magazine, and part of the governing board of the IEEE Blockchain initiative. He is a senior member of the ACM, USENIX, and IEEE. In 2013, he received the IEEE Reliability Society Engineer of the Year award. His team was awarded the DHS Cyber Security Division’s "Significant Government Impact Award" in 2017 and the “Bang for the Buck Award” in 2019. "
Abstract:
" The rise of agentic AI—autonomous systems capable of setting goals, planning actions, and adapting to dynamic environments—marks a transformative shift in how artificial intelligence integrates into critical infrastructure, decision-making, and cyber operations. However, as these systems gain more autonomy and operational latitude, they introduce novel challenges to cybersecurity and reliability. Unlike traditional automated systems, agentic AI can independently interpret context, revise objectives, and execute long-term strategies across interconnected platforms. This behavioral complexity expands the attack surface, introduces uncertainty in system responses, and makes predicting or auditing decisions difficult. Ensuring cybersecurity in this new paradigm requires moving beyond static threat models toward dynamic, context-aware defenses capable of adapting in real-time—much like the agents they protect. .Reliability also becomes a dual concern: hese systems must maintain technical uptime and functional correctness and uphold alignment with human intent and ethical boundaries, even under adversarial manipulation or environmental ambiguity. Techniques such as explainable AI, robust planning under uncertainty, adversarial resilience, and secure model deployment are critical to building trustworthy agentic systems. "
Tutorials
"Introduction to PHM Theory and Practice"
Tutorial speaker: Dr. Stephen Johnson
Bio:
" Dr. Stephen Johnson is the President of Dependable System Technologies, LLC, and the general editor for System Health Management: with Aerospace Applications (2011). His PHM experience includes being the control system fault protection engineer on the Magellan deep space probe to Venus in the 1980s, the head of Martin Marietta Astronautics Vehicle Health Management research in the early 1990s, a co-founder and head of engineering for a small PHM business in the 1990s, a faculty member of the Space Studies Department at the University of North Dakota from 1997 to 2005, the Analysis Lead for Mission and Fault Management on NASA Marshall Space Flight Center’s Space Launch System program and its precursors from 2005 to 2023, and numerous small PHM and systems engineering R&D contracts with government and industry from 2005 to the present. He has authored and co-authored numerous articles and books on system health management and fault management theory and practice, and other topics including systems engineering, space history and economics, and the philosophy of technology.
"
Abstract:
" This one-day short course introduces the core concepts and practices of Prognostics and Health Management / System Health Management. The class provides an introductory overview of the following: history of and motivation for PHM; core concepts and terminology; fault management functions; goals and requirements; architecture and design issues and strategies; technical performance metrics; and analysis methods. This course can be taken stand-alone or as part of the PHM Standards conference track.
"
"SPC methodologies for monitoring and optimizing HFC networks"
Tutorial speakers: Dr. Maher Harb and Nader Foroughi
Bio:
" Maher Harb is a Distinguished Engineer at Comcast working on problems at the intersection of Data Science, Machine Learning, and Telecommunications Networks. His research interests include applying Reinforcement Learning to optimize network performance, developing graph-based algorithms for network design & management, building deep neural network models for detection of RF impairments, and developing statistical methods for anomaly detection & root case identification. Prior to Comcast, Maher was an Assistant professor of Physics at Drexel University where he led an experimental Condensed Matter Physics laboratory investigating laser-matter interactions. Maher has a PhD in Physics from the University of Toronto and he held a postdoctoral fellowship at the Swedish National Synchrotron (Max-lab)."
Bio:
" Nader is a Distinguished Engineer at Comcast, where he is responsible for access network evolution, artificial intelligence and automation. Prior to joining Comcast, Nader was the Chief Technology Officer of Americas at Technetix, responsible for technology strategy and AI. His career also includes significant contributions at Shaw Communications, where he was responsible for access architecture and technology in conjunction with data sciences, making key advancements in proactive network maintenance, profile management application, and DOCSIS 4.0 development. Nader has a background in mathematics, engineering and systems architecture, with numerous white papers published spanning from DOCSIS 4.0 to applications of deep reinforcement learning in telecommunications."
Abstract:
" Statistical Process Control (SPC) and control charts, long-established quality control methodologies in manufacturing and process industries, have recently emerged as powerful tools for monitoring and optimizing Hybrid Fiber-Coaxial (HFC) networks, particularly in Full Duplex DOCSIS environments. Building upon decades of successful implementations in sectors ranging from semiconductor fabrication to pharmaceutical production, this paper presents an adaptation of SPC methodologies to cable network monitoring. Traditional network monitoring approaches relied on global thresholds for network parameters, limiting the ability to detect device-specific anomalies and performance degradation. Our implementation leverages individualized control charts for network devices, enabling dynamic threshold computation based on historical device behavior rather than system-wide metrics. By analyzing device-specific patterns and variations, the methodology demonstrates improved anomaly detection precision compared to global threshold approaches. The individualized control limits facilitate more accurate correlation between network events and performance degradation, particularly in identifying upstream channel impairments and RF interference patterns. Furthermore, the refined granularity of device-level control charts provides higher quality training data for backend machine learning algorithms, reducing false positives and enhancing predictive maintenance capabilities. This approach demonstrates significant potential for improving network reliability and maintenance efficiency in next-generation cable networks through data-driven, device-specific monitoring strategies.
"
"Hybrid Diagnostic & Prognostic Frameworks for Aerospace Health Management: Integrating Model-Based and Data-Driven Approaches"
Tutorial speakers: Dr. Afshin Rahimi
Bio:
" Dr. Afshin Rahimi is an Associate Professor in the Department of Mechanical, Automotive & Materials Engineering at the University of Windsor. He earned his PhD in Aerospace Engineering from Toronto Metropolitan University in 2017, after completing his MASc and BSc in the same field. His research centers on developing practical model-based and data-driven techniques for fault detection, diagnostics, and prognosis in complex systems, with a particular focus on aerospace applications. Dr. Rahimi has contributed to numerous peer-reviewed journals and conference papers and worked in the industry at Pratt & Whitney Canada. As a Professional Engineer (PEng) and a Senior Member of IEEE, he remains dedicated to continuous learning, sharing his insights and working collaboratively to advance the field of Prognostics and Health Management."
Abstract:
" This tutorial discusses hybrid frameworks for aerospace health management by integrating classical model-based techniques (e.g., Kalman filtering) with modern data-driven methods. Attendees will explore how these complementary approaches enable effective fault diagnosis and prognosis in complex aerospace systems, with some case studies. The session provides actionable insights to enhance system reliability and safety through robust, integrated PHM strategies.
"
"Design and reliability of LED based UVC light disinfection systems"
Tutorial speakers: Dr. Nicola Trivellin
Bio:
" Nicola Trivellin is an Associate Professor at the Department of Industrial Engineering at the University of Padua, where he was born in 1983. He obtained his bachelor's degree in Information Engineering in 2005 and his master's degree in Electronic Engineering in 2007, followed by a PhD in Information Science and Technology in 2010 at the same university. From 2011 to 2019, he served as a research fellow at the Department of Information Engineering and was the General Director of the spin-off company LightCube SRL. He has published over 60 scientific articles and holds 10 patents. Trivellin is a Guest Editor for the journals MDPI Materials and MDPI International Journal of Molecular Sciences. He teaches the courses "Microcontrollers & DSP" and "Automotive and Domotics" for the master's degree in Electronic Engineering, and "Photovoltaic Science and Technologies" for the master's degree in Energy Engineering. His research spans from the characterization of optoelectronic devices to innovative lighting for biomedical, industrial, and space applications."
Abstract:
" This tutorial provides a comprehensive review of recent advancements in the design and reliability of ultraviolet (UVC, between 100 nm and 280 nm) disinfection systems, specifically examining how design parameters and reliability factors impact overall system performance. UVC disinfection has become increasingly important due to the global COVID-19 pandemic, driving research into effective methods of pathogen inactivation, particularly against viruses such as SARS-CoV-2. This tutorial identifies and explains key factors essential to developing efficient and reliable disinfection technologies, from the optical design point of view to durability. The tutorial begins by thoroughly exploring innovative system design strategies aimed at maximizing UVC irradiation uniformity and efficacy. A particular focus is placed on a spherical irradiation system employing 275 nm UV-C LEDs. This innovative design exemplifies rapid and uniform pathogen inactivation, demonstrating the ability to achieve a 99.9% reduction of SARS-CoV-2 viral particles within just one minute at a dose of 83.1 J/m². The discussion includes an in-depth examination of optical analyses and simulation studies conducted to optimize the system’s configuration. Important considerations such as the spatial arrangement of LEDs, reflective material properties, and the geometric design of the irradiation chamber are critically assessed to illustrate their direct impact on the effectiveness and consistency of disinfection outcomes, for different target shape and surfaces area. Following the exploration of system design, the tutorial addresses critical reliability concerns associated with the use of UVC LEDs. The reliability of LEDs is identified as a fundamental constraint limiting broader commercial and clinical adoption. Stress testing and detailed reliability studies of commercially available UVC LEDs are reviewed, highlighting prevalent issues such as reductions in optical output power, efficiency degradation, and the emergence of parasitic luminescence pathways over the operational lifetime of the devices. These degradation phenomena are primarily attributed to increased non-radiative recombination through defect states within the LED’s active regions, resulting in diminished optical performance and compromised disinfection efficacy over time. Further, the tutorial synthesizes the interconnectedness between effective design principles and device reliability, emphasizing the critical importance of considering both factors concurrently during the development of UVC disinfection systems. Recommendations and insights are provided for improving the robustness and commercial viability of these systems. Special attention is directed toward understanding the underlying physical mechanisms responsible for LED degradation and integrating design strategies aimed at mitigating these reliability challenges. By comprehensively reviewing both the current state-of-the-art in system design and the inherent reliability issues affecting UVC LED technologies, the tutorial aims to provide a balanced perspective necessary for the informed development of advanced UVC-based disinfection solutions. The objective is to equip researchers, engineers, and practitioners with the knowledge required to enhance design methodologies, optimize disinfection efficacy, and overcome reliability hurdles, ultimately contributing to the wider adoption of this promising technology in various public health and clinical applications."
