TechTalk

HKAES TechTalk – Integration of AI and Wireless Networks for 6G Systems

May 21 2024 (Tuesday) 4:00-5:00pm
AI will be tightly integrated with wireless networks in future 6G wireless systems. On one hand, more and more AI applications will be running directly on mobile terminals for inferencing and training because dataset exists locally at mobile or IoT devices but large neural network models will be in the cloud. As such, wireless communication networks will be “in-the-loop” between mobile devices and the neural network models. it is important for future wireless networks to be optimised to support more efficient AI deployment. On the other hand, AI as a powerful tool can have huge potential to solve challenging problems in wireless communications design, resulting in better and more efficient wireless networks. There are lots of potential to deploy domain-specific AI design to implement challenging modules such as the physical layer and system designs in wireless systems. In this talk, we shall explore the integration of AI and future 6G wireless networks.

TechTalk – Greening the Slope: Revolutionizing Urban Water Management in Hilly Environments

May 14, 2024 (Tuesday) 3:30-4:30pm
Urban landscapes on slopes present unique challenges for managing stormwater. Traditional drainage systems often exacerbate other environmental problems, but sustainable drainage systems (SuDS) offer a nature-inspired solution. By mimicking natural processes, SuDS absorb, clean, and slowly release rainwater, effectively reducing stormwater and enhancing urban resilience. This talk explores innovative SuDS designs tailored for hilly environments. Findings highlight the effectiveness of integrating SuDS with existing infrastructure, particularly in collection-storage-release forms or terraced configurations. These showcase not only nearly complete water quantity control but also the potential of SuDS to transform urban water dynamics on slopes and emphasize their role in fostering sustainable cities.

TechTalk – Healthcare Decision-Making: From Data to Solutions and Insights

May 09 2024 (Thursday) 4:30-5:30pm
Every day, numerous decisions must be made in healthcare systems, addressing a wide range of issues, including patient admissions, physician assignments, operating room scheduling, pharmaceutical inventory control, and ambulance dispatch. Conventionally, these decisions are made by experienced experts. However, human decision-makers may face challenges such as the unavailability of complete information, rapidly changing healthcare environments, personal biases, and prompt responses to medical emergence. Additionally, decision-makers often struggle to evaluate the effectiveness and possible consequences of their decisions.
This talk will introduce a set of scientific and systematic tools that leverage operations research techniques (originally developed to solve decision-making problems during World War II) and data collected in healthcare systems. The talk will begin with an overview of traditional healthcare decision-making approaches before delving into data-driven methodologies that have emerged in recent years. Subsequently, the speaker will share his experience over the years in tackling decision-making problems in healthcare systems in Hong Kong and the region, including emergency department patient flow modelling and optimization, medical appointment scheduling, vaccination strategies, and healthcare logistics. Finally, potential future directions for addressing healthcare decision-making challenges will be explored.
The supports by HKRGC (Refs: 27200419; 17204823; F-HKU704/22) and HMRF of the Health Bureau (Refs: 14202115; 21222881) for the projects to be presented in this talk are greatly acknowledged.

Young Scholar TechTalk – On Causality and the Upper Limit of Sound Absorption

April 30 2024 (Tuesday) 4:30-5:30pm
Sound absorption across a wide range of frequencies is a focus in contemporary acoustics. Recently, integral bounds of absorption or reflection coefficients were introduced as a guide of design optimization following the footsteps of electromagnetics, where integral relations were derived based on system causality considerations. This talk carefully examines the proper formulation of physical causality and its implications on the scattering properties of the system. Taking into consideration the effects of different physical boundary conditions and the bulk absorber material, a more generalized integral bound is derived. It becomes evident that, while the bound exists, it is governed by system stiffness rather than the causality constraint. By studying the effects of various approximations made during mathematical derivations, the physics of the bound is thoroughly discussed, and the limitations in utilizing integral bounds as reference for design optimization are highlighted. The findings are expected to have significant implications for the development of effective noise reduction strategies and the advancement of smart acoustic design.

TechTalk – Optimal of the Service Reliability of Self-Service Systems

April 11 2024 (Thursday) 4:30-5:30pm
Self-service systems, such as electrical vehicle charging piles, are often deployed in a fleet at a designated location. These systems may fail either while idle or during customer use. If the system initially chosen by a customer is nonfunctional or fails during service, the customer may switch to another available system for service. This phenomenon is called failure-induced demand switching (FDS). To maximize the long-run profit of the fleet, the operator must balance revenue with operation and maintenance costs, by conducting appropriate maintenance actions. However, it is challenging to develop an optimal maintenance policy due to FDS. This talk defines and assesses the service level of such systems in terms of the proportion of demands being completed in the long run, and proposes a three-dimensional maintenance policy to maximize the fleet’s long-run profit rate. The insights are expected to guide the operators of such systems in enhancing their profitability.

TechTalk – Heterogeneous Silicon Photonic Integrated Circuits: Progress and Prospects

April 18 2024 (Thursday) 5:00-6:00pm
Photonics has played a vital role in shaping information technology infrastructure. Now it is considered crucial for addressing the emerging bandwidth constraints and power consumption challenges in hyperscale datacenters and high-performance computing systems. Central to these photonic solutions are photonic integrated circuits (PICs) which have seen dramatic progress in the past decades. A heterogeneous silicon photonics platform that integrates materials with superior properties is projected to fully unlock the application potentials of PICs. Examples include combining novel light sources (e.g., optical frequency combs and narrow-linewidth lasers) with other functional components on a single chip, to construct PICs with complete optical functionalities and advantages in terms of integration density, energy consumption and performance. In this talk, I will introduce the recent progress of heterogeneous silicon photonic integrated circuits and outline the prospects.

HKAES TechTalk – Creating New Radio Frequency Wave Technology for 6G

April 23 2024 (Tuesday) 4:00-5:00pm
Radio frequency (RF) waves are a fundamental phenomenon that can carry electromagnetic signals and energy through space and interact with it. Their use in wireless communication has revolutionized our lives and created a mobile information society and new industries. However, RF wave technology can do much more and in this talk I explore new RF wave technologies that can be further exploited for 6G. I broadly classify the new RF technologies into wave shaping and wave sensing and in this talk I focus on the development of RF Imaging, RF energy harvesting and reconfigurable intelligent surfaces (RIS). In particular both the concepts and experimental results obtained from our prototypes are presented for each of these technologies. Furthermore, while each of these technologies is promising, significant further research is needed to exploit the enormous potential of new RF wave technology for 6G and this is also detailed.

TechTalk – Machine Learning and Animal Behavior: Interdisciplinary Innovations in Neuroscience and Engineering

April 12 2024 (Friday) 3:30-4:30pm
This presentation explores machine learning (ML) integration with animal behavior studies and its transformative applications across neuroscience, human mobility analysis, and engineering. At the core of our research is the affinity between the complex behaviors observed in animal society and the predictive capabilities of ML algorithms.
In neuroscience, we introduce the development of robotic microscopes and feedback projections, offering insights into animal behavior at macroscopic and microscopic levels. This foundation supports our further applications in diverse fields such as retail analytics, human relationships estimation, and mobility service design. In the engineering domain, our work extends to preventive maintenance (condition-based maintenance: CBM) in manufacturing and transportation, drawing from the predictive nature of ML to foresee and mitigate equipment failures. Moreover, we introduce ML for digital twinning to create dynamic virtual models of physical systems.
This array of applications highlights the critical role of integrating computer vision and ML into problem-solving workflows across various production plants. This presentation emphasizes the essential need for interdisciplinary collaboration, bridging the gap between biologists, data scientists, and engineers.

HKAES TechTalk – Collaborative Edge Computing for Ubiquitous AI

March 26 2024 (Tuesday) 4:00-5:00pm
Edge computing provides a powerful way to rapidly analyze data and process tasks at the edge of the network, closer to the end-user. Edge AI extends edge computing to enable AI on edge devices to make instantaneous intelligent decisions. In this talk, I will describe the vision of future Edge AI, called Ubiquitous AI, empowered by collaborative edge computing (CEC), which is a new-generation scalable edge computing infrastructure. With the computing power shared by collaborative edge nodes, edge AI models can be trained and seamlessly deployed jointly by the edge nodes that can be geographically distributed across a wide area. I will highlight the architecture and framework of CEC and how it enables ubiquitous AI demanded by advanced smart city applications like autonomous driving, industry 4.0 and metaverse.

TechTalk – UAV Navigation and Mapping with Light Detection and Ranging (LiDAR) Sensors

March 28 2024 (Thursday) 4:30-5:30pm
Over the last decades, small-size multi-copter unmanned aerial vehicles (UAVs) have received intensive research interests. These UAVs have shown promising potential for various applications, including aerial photography, farming, delivery, mapping, and surveying. However, for these applications to be successful, autonomous flights in unknown environments are necessary. In this talk, we will discuss our work on developing autonomous UAVs using lidar navigation. Specifically, we will explore recent advancements in lidar technologies and focus on navigation algorithms, including localization, mapping, planning, and control. We will showcase how lidar sensors can be utilized on small UAVs to enable complex navigation tasks, such as high-speed flight navigation, environment exploration, and estimation of agile UAV motion.