Seminar

Abstract Most power electronic equipment, including power supplies for Artificial Intelligence (AI) and high-performance computing systems, has traditionally been designed using discrete active and passive components. However, the power electronics industry has reached a stage where improving one performance attribute often comes at the expense of others. As AI and high-performance computing systems demand increasingly […]

Abstract Advances in fiber-optic systems are crucial for the internet. To cope with this continued exponential growth as Keck´s law, multimode fibers (MMF) with spatial division multiplexing (SDM) are proposed. However, there are challenging mode scattering effects in MMF. Physics-informed deep learning enables to correct the scattering, resulting in advancing security and data rate. We […]

Abstract Magnetic Resonance Imaging (MRI) is a versatile medical imaging modality. Despite its crucial role in modern healthcare, MRI remains largely inaccessible to the general population. The development of ultra-low-field (ULF) MRI offers opportunities for enabling low-cost, low-power, and potentially portable clinical applications. However, the imaging performance of these emerging ULF MRI scanners remains poor […]

Abstract Recent years have witnessed the remarkable success of deep neural networks in healthcare, particularly in the analysis of medical images and signals. However, their performance is often constrained by the scarcity of labelled data, driven by high labelling costs and challenges related to data privacy and sharing. In this talk, I will explore how […]

Abstract The evolution of generative artificial intelligence (GAI) has driven revolutionary applications like ChatGPT. The proliferation of these applications is underpinned by the mixture of experts (MoE), which contains multiple experts and selectively engages them for each task to lower operation costs while maintaining performance. Despite MoE's efficiencies, GAI still faces challenges in resource utilization […]