CO-CHAIRS:
Chulsoon Hwang, Missouri University of Science and Technology
Ling Zhang, Zhejiang University

Achieving power integrity (PI) across the PCB, package, and silicon over multiple power domains of differing logic levels and functionalities remains a challenging and unsolved problem. Mature and proven commercial analysis tools for post-layout assessment are accurate and well-established. However, PI design and layout remain a trail-and-error process that requires considerable experience, effort, and engineering time to achieve successfully on modern high-speed designs. Locating the power net on a given layer(s) within the stackup, location of decoupling capacitors, values of capacitors, number, and achievable frequency range, are typical design parameters over the PCB and package. Mitigating undesired resonances resulting from interconnects and capacitances is critical. Currently, the physics of charge delivery across a power distribution network is well known, and optimization approaches for achieving a target impedance have been reported in the literature. However, PI design for modern systems that might include a sub-milli ohm target impedance for high-current ICs requiring hundreds of decoupling capacitors that must accommodate hundreds of routing channels, or mobile designs with numerous different power domains and limited space for decoupling, as well as other challenges remain unresolved with a proven design methodology or guidelines

AI/ML algorithms are proving to be a powerful tool for challenging PI problems. In recent years, AI/ML algorithms have achieved unprecedented success in various complex tasks due to their outstanding ability to fit complex functions. PI design with hundreds of decoupling capacitors required in modern electronic circuits is a time-consuming and tedious process that cannot be well addressed by conventional optimization approaches or commercial tools. AI/ML algorithms, such as deep learning and deep reinforcement learning, have recently demonstrated great potential in solving complex PI problems without human intervention, such as decoupling capacitor optimization and pre-layout power distribution network design. The incorporation of AI/ML in PI design has the potential to significantly reduce the PI design cycle and promote the advancement of artificial intelligence-assisted electronic design automation (EDA) tools. However, the reusability and generalization performance of the trained AI/ML algorithms in different scenarios is one of the biggest challenges in this direction. Moreover, the reliability of AI/ML algorithms in finding the optimal solution when optimizing a tremendous search space still demands continuous research. Therefore, a special session about AI/ML algorithms in PI design, which may contribute to new ideas and solutions to the above challenges, is essential and meaningful.

CO-CHAIRS:
Ji Chen, University of Houston       
Ananda Kumar, US Food and Drug Administration

The increasing complexity of medical devices and their widespread use in electromagnetically challenging clinical environments—such as MRI suites, operating rooms, and intensive care units—raise critical electromagnetic compatibility (EMC) and safety concerns. This Special Session focuses on emerging EMC challenges for medical devices, including susceptibility to electromagnetic fields, unintended coupling mechanisms, device–tissue interactions, and risk mitigation strategies. The session aims to bring together researchers, clinicians, device manufacturers, and regulatory experts to discuss recent advances in modeling, testing, and standards development related to medical device EMC. Topics will include both experimental and computational approaches, real-world incident analyses, and regulatory perspectives to ensure safe and reliable device operation in clinical settings.

CO-CHAIRS:
Yuichi Hayashi, Nara Institute of Science and Technology      
Michael McInerney, Consultant

As electronic devices become integral to critical infrastructure and daily life, ensuring physical layer security is paramount. This special session addresses the dual challenges of electromagnetic information security, specifically focusing on threats from passive information leakage and active electromagnetic interference. While advances in signal processing have refined leakage analysis, they have also enabled more precise injection and interference attacks. This session aims to bridge the gap between conventional EMC engineering and cybersecurity by exploring how electromagnetic emissions and interference characteristics can be exploited for attacks or utilized for defense. We will discuss cutting-edge research including advanced techniques for information reconstruction, sophisticated active electromagnetic injection attacks, novel methods for hardware assurance, and the integration of AI technology in electromagnetic security. The session brings together experts to define the forefront of evaluating risks and developing countermeasures against these evolving electromagnetic threats.

CO-CHAIRS:
Karol Niewiadomski, University of Twente 
Hanzhi Ma, Zhejiang University
Ling Zhang, Zhejiang University

Electromagnetic Compatibility (EMC) and Signal/Power Integrity (SIPI) engineering face increasingly complex challenges as modern electronic systems operate at higher frequencies, integrate greater component densities, and demand enhanced performance within stringent regulatory frameworks. Traditional design methodologies and problem-solving approaches, while foundational, often require extensive computational resources, specialized expertise, and iterative processes that can significantly extend development cycles.

The emergence of artificial intelligence agents and generative AI tools presents unprecedented opportunities to revolutionize EMC/SIPI engineering practices. AI agents capable of autonomous decision-making and adaptive learning can streamline complex electromagnetic modeling, optimize design parameters, and predict interference patterns with remarkable efficiency. Generative AI technologies, including large language models and machine learning-based design tools, offer novel approaches to circuit layout optimization, filter design, shielding strategies, and regulatory compliance verification.

This special session will explore the transformative potential of AI-driven methodologies in addressing critical EMC/SIPI challenges. Topics of discussion will encompass the application of AI agents for automated electromagnetic simulation workflows, generative algorithms for optimal PCB routing and component placement, machine learning approaches to EMI/EMC prediction and mitigation, and intelligent design tools for signal/power integrity optimization. Additionally, the session will examine practical implementation strategies, validation methodologies, and the integration of AI tools within existing EMC/SIPI design environments.