Department of Electrical and Computer Engineering, HKU

🧑🏼‍💻 從繪製立體的世界，到重建它所發出的光🔮✨
👩🏼‍💻 𝐅𝐫𝐨𝐦 𝐑𝐞𝐧𝐝𝐞𝐫𝐢𝐧𝐠 𝐚 𝟑𝐃 𝐖𝐨𝐫𝐥𝐝 𝐭𝐨 𝐑𝐞𝐜𝐨𝐧𝐬𝐭𝐫𝐮𝐜𝐭𝐢𝐧𝐠 𝐭𝐡𝐞 𝐋𝐢𝐠𝐡𝐭 𝐈𝐭 𝐄𝐦𝐢𝐭𝐬🔮✨

💡想像一下如果全息顯示不只是讓影像「看起來立體」，而是能重現光在場景中的傳播方式，會是怎樣的體驗？香港大學電機與計算機工程系的研究團隊近期發表以下兩項成果，進一步推動電腦生成全息技術由「繪製立體影像」邁向「重建光的傳播」。雖然仍屬研究原型，但已為未來虛擬實境（VR）及擴增實境（AR）全息顯示提供值得期待的發展路線。

📍刊於 𝘕𝘢𝘵𝘶𝘳𝘦 𝘊𝘰𝘮𝘮𝘶𝘯𝘪𝘤𝘢𝘵𝘪𝘰𝘯𝘴 的 𝗠𝗲𝘀𝗵 𝗛𝗼𝗹𝗼𝗴𝗿𝗮𝗽𝗵𝘆，直接利用多邊形網格生成全息圖。透過保留連續幾何、可見性、紋理、振幅與相位，系統能夠呈現自然的焦點轉換、視差，以及隨觀看位置改變的遮擋關係。

📍刊於 𝘈𝘊𝘔 𝘛𝘳𝘢𝘯𝘴𝘢𝘤𝘵𝘪𝘰𝘯𝘴 𝘰𝘯 𝘎𝘳𝘢𝘱𝘩𝘪𝘤𝘴，並於 SIGGRAPH 2026 發表的 𝗛𝗼𝗹𝗼𝗣𝗮𝘁𝗵𝗧𝗿𝗮𝗰𝗲𝗿，則把路徑追蹤帶進波動光學。它能將反射、折射、光澤材質與全局光照直接編碼到複數光波場之中。

💡What if a holographic display could do more than make images look 3D, and instead recreate how light behaves in the real world? Two recent works from the Department of Electrical and Computer Engineering at The University of Hong Kong (HKU ECE) explore this vision through a unified graphics-to-wave pipeline. Together, the two projects move computer-generated holography beyond rendering images of 3D scenes towards reconstructing how light propagates through them. These are still research prototypes, but they offer a promising route toward more physically faithful holographic displays for future Virtual Reality (VR) and Augmented Reality (AR) systems.

📍𝗠𝗲𝘀𝗵 𝗛𝗼𝗹𝗼𝗴𝗿𝗮𝗽𝗵𝘆, published in 𝘕𝘢𝘵𝘶𝘳𝘦 𝘊𝘰𝘮𝘮𝘶𝘯𝘪𝘤𝘢𝘵𝘪𝘰𝘯𝘴, generates holograms directly from polygon meshes. By preserving continuous geometry, visibility, texture, amplitude, and phase, it enables natural focus transitions, parallax, and view-dependent occlusion.

📍𝗛𝗼𝗹𝗼𝗣𝗮𝘁𝗵𝗧𝗿𝗮𝗰𝗲𝗿, published in 𝘈𝘊𝘔 𝘛𝘳𝘢𝘯𝘴𝘢𝘤𝘵𝘪𝘰𝘯𝘴 𝘰𝘯 𝘎𝘳𝘢𝘱𝘩𝘪𝘤𝘴 and presented at SIGGRAPH 2026, brings path tracing into wave optics. It encodes reflections, refractions, glossy materials, and global illumination directly into complex wave fields.

🔗 𝗠𝗲𝘀𝗵 𝗛𝗼𝗹𝗼𝗴𝗿𝗮𝗽𝗵𝘆: https://urimoo.github.io/Meshholography
🔗 𝗛𝗼𝗹𝗼𝗣𝗮𝘁𝗵𝗧𝗿𝗮𝗰𝗲𝗿: https://zhou-wb.github.io/holopathtracer

🔗 Related News: https://ece.hku.hk/20260731-1
🔗 Read more on WeLight@HKU Website: https://hku.welight.fun/publications

#全息顯示 #計算成像

💬 𝐒𝐞𝐦𝐢𝐧𝐚𝐫 𝐨𝐧 𝐁𝐚𝐲𝐞𝐬𝐢𝐚𝐧 𝐀𝐬𝐲𝐦𝐩𝐭𝐨𝐭𝐢𝐜𝐬: 𝐅𝐫𝐨𝐦 𝐔𝐧𝐢𝐯𝐞𝐫𝐬𝐚𝐥 𝐒𝐨𝐮𝐫𝐜𝐞 𝐂𝐨𝐝𝐢𝐧𝐠 𝐭𝐨 𝐆𝐞𝐧𝐞𝐫𝐚𝐥 𝐋𝐚𝐫𝐠𝐞-𝐬𝐜𝐚𝐥𝐞 𝐌𝐈𝐌𝐎 𝐂𝐨𝐦𝐦𝐮𝐧𝐢𝐜𝐚𝐭𝐢𝐨𝐧

In this talk, the speaker will first provide a brief overview of the role of Bayesian asymptotics in universal source coding, specifically through mixture codes such as the well-known Krichevsky-Trofimov coding distribution for i.i.d. sources and the context-tree weighting algorithm for variable-order Markov sources. The speaker will then transition to the large-scale MIMO communication problem by connecting the minimum Bayesian redundancy in source coding to the maximum mutual information in channel coding. Specifically, the speaker will demonstrate that the asymptotic capacity of a general MIMO channel—under both peak and average power constraints—admits a simple and elegant expression through what we term the Jeffreys factor, with the optimal input distribution being the tilted Jeffreys prior. This Jeffreys factor generalises the role of the signal-to-noise ratio (SNR) found in conventional MIMO systems, capturing all the statistical nuances of general MIMO channels. Finally, the speaker will provide a systematic, three-step recipe to straightforwardly compute this asymptotic capacity, applying it to diverse scenarios such as channels with clipping, phase noise, or output quantisation, fading channels with imperfect channel knowledge, and even the optical Poisson channel. If time permits, the speaker will conclude with a discussion on the more practical problem of constellation design for large-scale MIMO channels. This talk is based on the paper below: S. Yang and R. Combes, “From Bayesian Asymptotics to General Large-Scale MIMO Capacity,” IEEE Transactions on Information Theory, vol. 72, no. 5, pp. 3014-3029, May 2026.

📅 𝗗𝗮𝘁𝗲: August 19, 2026 (Wednesday)
⏰ 𝗧𝗶𝗺𝗲: 4:30 pm – 5:30 pm
📍 𝗩𝗲𝗻𝘂𝗲: CB-603, 6/F, Chow Yei Ching Building, The University of Hong Kong
🔗 𝗗𝗲𝘁𝗮𝗶𝗹𝘀: https://ece.hku.hk/events/20260819-1

𝑭𝒐𝒍𝒍𝒐𝒘 𝑰𝑮/𝑭𝑩: 𝑯𝑲𝑼𝑬𝑪𝑬

💬 𝐒𝐞𝐦𝐢𝐧𝐚𝐫 𝐨𝐧 𝐎𝐩𝐭𝐢𝐜𝐚𝐥 𝐍𝐞𝐮𝐫𝐚𝐥 𝐏𝐫𝐨𝐜𝐞𝐬𝐬𝐢𝐧𝐠 𝐟𝐨𝐫 𝐒𝐮𝐩𝐞𝐫 𝐂𝐚𝐦𝐞𝐫𝐚𝐬

The light field incident through camera aperture typically consists of more than 10^16 photons per second, suggesting that cameras capturing megapixel scale images at 30 frames per second vastly under represent the optical data cube. Here we briefly review the history of gigapixel array cameras and optical neural processing. We then discuss recent results suggesting a path combine these technologies in low power optical processing for terapixel per second super cameras.

📅 𝗗𝗮𝘁𝗲: August 11, 2026 (Tuesday)
⏰ 𝗧𝗶𝗺𝗲: 9:30 am – 10:30 am
📍 𝗩𝗲𝗻𝘂𝗲: CB-603, 6/F, Chow Yei Ching Building, The University of Hong Kong

🔗 𝗗𝗲𝘁𝗮𝗶𝗹𝘀: https://ece.hku.hk/events/20260811-1

𝑭𝒐𝒍𝒍𝒐𝒘 𝑰𝑮/𝑭𝑩: 𝑯𝑲𝑼𝑬𝑪𝑬

💬 𝐒𝐞𝐦𝐢𝐧𝐚𝐫 𝐨𝐧 𝐂𝐨𝐧𝐭𝐚𝐜𝐭𝐥𝐞𝐬𝐬 𝐌𝐚𝐠𝐧𝐞𝐭𝐢𝐜 𝐒𝐞𝐧𝐬𝐢𝐧𝐠 𝐢𝐧 𝐂𝐨𝐧𝐝𝐢𝐭𝐢𝐨𝐧 𝐌𝐨𝐧𝐢𝐭𝐨𝐫𝐢𝐧𝐠 𝐚𝐧𝐝 𝐀𝐧𝐨𝐦𝐚𝐥𝐲 𝐃𝐞𝐭𝐞𝐜𝐭𝐢𝐨𝐧 𝐟𝐨𝐫 𝐒𝐦𝐚𝐫𝐭 𝐆𝐫𝐢𝐝: 𝐍𝐞𝐰 𝐏𝐨𝐬𝐬𝐢𝐛𝐢𝐥𝐢𝐭𝐢𝐞𝐬 𝐚𝐧𝐝 𝐀𝐥𝐭𝐞𝐫𝐧𝐚𝐭𝐢𝐯𝐞𝐬

Our physical and cyber environments are becoming increasingly intertwined with smarter sensing, communication, and data analytics. Our daily livings are indeed surrounded by a wide variety of sensors, IoT connectivity, and edge computing devices, constituting smart grid, smart city, smart transportation, and so on. The availability of sensing devices with measurement, communication, and processing capabilities is providing fine-grained data. Together with multimodal sensory data collection and sensor fusion can result in actionable insights and decisions. This synergy can lead to improved ways and quality of life in what we call smart living.
Magnetism is one of the six energy forms of measurands in sensing. Magnetic sensing plays a critical role in smart living due to various sources of magnetic fields such as magnetic fields from current-carrying wires and permanent magnets which are geometrically determined by Biot-Savart Law and Ampere's Law respectively. These magnetic fields can range from DC to AC, from low frequency to high frequency...

📅 𝗗𝗮𝘁𝗲: August 3, 2026 (Monday)
⏰ 𝗧𝗶𝗺𝗲: 3:00 pm – 4:00 pm
📍 𝗩𝗲𝗻𝘂𝗲: CB-603, 6/F, Chow Yei Ching Building, The University of Hong Kong
🔗 𝗗𝗲𝘁𝗮𝗶𝗹𝘀: https://ece.hku.hk/events/20260803-1

𝑭𝒐𝒍𝒍𝒐𝒘 𝑰𝑮/𝑭𝑩: 𝑯𝑲𝑼𝑬𝑪𝑬

💬 𝐒𝐞𝐦𝐢𝐧𝐚𝐫 𝐨𝐧 𝐔𝐬𝐢𝐧𝐠 𝐀𝐈, 𝐈𝐨𝐓𝐬, 𝐚𝐧𝐝 𝐌𝐞𝐭𝐚𝐯𝐞𝐫𝐬𝐞 𝐢𝐧 𝐇𝐊𝐔 𝐂𝐚𝐦𝐩𝐮𝐬𝐥𝐚𝐧𝐝 𝐟𝐨𝐫 𝐒𝐓𝐄𝐀𝐌+𝐈𝐏 𝐄𝐝𝐮𝐜𝐚𝐭𝐢𝐨𝐧

Current education focuses on instructional-based learning, problem-based learning, online learning, blended Learning, and experiential learning. It emphasises structured memorisation, repetition, teacher-centred, problem analysis, self-engagement, internal mental processes, and transfer experience. It does NOT develop students’ innovation, industrial knowledge, real-world practical experience, and integrated knowledge. Although existing STEAM education provides interdisciplinary training for developing students’ critical thinking, complex problem-solving, adaptability, and creativity. However, teachers have limited knowledge in creating practical learning content. By using artificial intelligence, IoTs and metaverse together can provide an intelligent, interactive, collaborative and virtual environment can be provided for teachers to generate practical learning content without prior industrial knowledge and experience, and allow learners to learn through a virtual, collaborative, practical and interactive environment in a virtual world. This talk will discuss the current and future AI technologies for education, and how we use AI, IoTs and metaverse together to create the STEAM+IP learning content.

📅 𝗗𝗮𝘁𝗲: July 23, 2026 (Thursday)
⏰ 𝗧𝗶𝗺𝗲: 2:00 pm – 3:00 pm
📍 𝗩𝗲𝗻𝘂𝗲: CB-603, 6/F, Chow Yei Ching Building, The University of Hong Kong

🔗 𝗗𝗲𝘁𝗮𝗶𝗹𝘀: https://ece.hku.hk/events/20260723-2

𝑭𝒐𝒍𝒍𝒐𝒘 𝑰𝑮/𝑭𝑩: 𝑯𝑲𝑼𝑬𝑪𝑬

💬 𝐒𝐞𝐦𝐢𝐧𝐚𝐫 𝐨𝐧 𝐒𝐢𝐧𝐠𝐥𝐞-𝐈𝐧𝐝𝐮𝐜𝐭𝐨𝐫 𝐌𝐮𝐥𝐭𝐢𝐩𝐥𝐞-𝐎𝐮𝐭𝐩𝐮𝐭 (𝐒𝐈𝐌𝐎) 𝐈𝐧𝐯𝐞𝐫𝐭𝐞𝐫𝐬-𝐃𝐞𝐬𝐢𝐠𝐧 𝐚𝐧𝐝 𝐀𝐩𝐩𝐥𝐢𝐜𝐚𝐭𝐢𝐨𝐧𝐬

Single-inductor multiple-output (SIMO) inverters offer a compact, well-proven, and scalable solution for generating multiple AC or hybrid DC-AC outputs using only a single magnetic component in the power stage. This approach delivers key advantages for low- to medium-power, space-constrained applications, including lower cost, smaller size, reduced weight, higher efficiency, and improved scalability. This seminar introduces the system architecture and circuit topologies of patented SIMO inverters, and presents a scalability model for determining the maximum achievable number of outputs. In practical applications, SIMO inverters can be used as multi-coil wireless chargers in Qi-compliant multi-coil wireless power transfer (MC-WPT) systems, enabling faster simultaneous charging of multiple devices as well as greater positional freedom for charging a single device.

📅 𝗗𝗮𝘁𝗲: July 16, 2026 (Thursday)
⏰ 𝗧𝗶𝗺𝗲: 2:00 pm – 3:00 pm
📍 𝗩𝗲𝗻𝘂𝗲: CB-603, 6/F, Chow Yei Ching Building, The University of Hong Kong

🔗 𝗗𝗲𝘁𝗮𝗶𝗹𝘀: https://ece.hku.hk/events/20260716-1

𝑭𝒐𝒍𝒍𝒐𝒘 𝑰𝑮/𝑭𝑩: 𝑯𝑲𝑼𝑬𝑪𝑬

💬 𝐒𝐞𝐦𝐢𝐧𝐚𝐫 𝐨𝐧 𝐀𝐜𝐜𝐞𝐬𝐬𝐢𝐛𝐥𝐞 𝐅𝐞𝐭𝐚𝐥 𝐍𝐞𝐮𝐫𝐨𝐢𝐦𝐚𝐠𝐢𝐧𝐠 𝐓𝐡𝐫𝐨𝐮𝐠𝐡 𝐔𝐥𝐭𝐫𝐚𝐬𝐨𝐮𝐧𝐝 𝐚𝐧𝐝 𝐃𝐞𝐞𝐩 𝐋𝐞𝐚𝐫𝐧𝐢𝐧𝐠

Two-dimensional (2D) ultrasound imaging is the standard routine modality for monitoring fetal brain development, given its cost-effectiveness, safety and real-time acquisition capabilities. However, its clinical utilisation is often limited by a lack of 3D anatomical context and high operator dependency, requiring expensive and less accessible magnetic resonance imaging (MRI) for more advanced diagnosis and analysis. This talk explores how deep learning can bridge this gap by transforming routine 2D bedside ultrasound, even with limited data, into a sophisticated and accessible tool for 3D fetal neurodevelopmental assessment.

📅 𝗗𝗮𝘁𝗲: July 15, 2026 (Wednesday)
⏰ 𝗧𝗶𝗺𝗲: 4:00 pm – 5:00 pm
📍 𝗩𝗲𝗻𝘂𝗲: CB-603, 6/F, Chow Yei Ching Building, The University of Hong Kong

🔗 𝗗𝗲𝘁𝗮𝗶𝗹𝘀: https://ece.hku.hk/events/20260715-2

𝑭𝒐𝒍𝒍𝒐𝒘 𝑰𝑮/𝑭𝑩: 𝑯𝑲𝑼𝑬𝑪𝑬

💬 𝐒𝐞𝐦𝐢𝐧𝐚𝐫 𝐨𝐧 𝐒𝐩𝐚𝐫𝐬𝐞 𝐂𝐨𝐧𝐯 𝐚𝐧𝐝 𝐆𝐫𝐚𝐩𝐡 𝐅𝐨𝐮𝐫𝐢𝐞𝐫 𝐑𝐞𝐩𝐫𝐞𝐬𝐞𝐧𝐭𝐚𝐭𝐢𝐨𝐧 𝐟𝐨𝐫 𝐏𝐨𝐢𝐧𝐭 𝐂𝐥𝐨𝐮𝐝 𝐒𝐮𝐩𝐞𝐫-𝐑𝐞𝐬𝐨𝐥𝐮𝐭𝐢𝐨𝐧, 𝐏𝐫𝐞𝐝𝐢𝐜𝐭𝐢𝐨𝐧, 𝐚𝐧𝐝 𝐂𝐨𝐦𝐩𝐫𝐞𝐬𝐬𝐢𝐨𝐧

Due to the increased popularity of 3D immersive visual communication in augmented and virtual reality applications, as well as 3D sensing for digital tunes and auto-driving, the interest in capturing high resolution real-world point clouds has grown significantly in recent years. Point cloud is a new class of signal that is non-uniform and sparse and this present unique challenges to the signal processing, compression and learning problems. In this talk, we present our multi-scale sparse convolutional learning and Graph Frourier Transform (GFT) based framework for large scale point cloud processing, with applications to the geometry and attributes super-resolution, and dynamic point cloud compression with latent space compensation. The architecture is memory efficient and can learn a deep networks to handle large scale point cloud in real world applications. Initial results demonstrated that this framework achieved new state of the art results in geometry super-resolution, attributes deblocking and super-resolving, and dynamic point cloud sequence compression, and is adopted in the MPEG AI based point cloud coding framework.

📅 𝗗𝗮𝘁𝗲: July 27, 2026 (Monday)
⏰ 𝗧𝗶𝗺𝗲: 4:00 pm – 5:30 pm
📍 𝗩𝗲𝗻𝘂𝗲: CB-603, 6/F, Chow Yei Ching Building, The University of Hong Kong

🔗 𝗗𝗲𝘁𝗮𝗶𝗹𝘀: https://ece.hku.hk/events/20260727-1

𝑭𝒐𝒍𝒍𝒐𝒘 𝑰𝑮/𝑭𝑩: 𝑯𝑲𝑼𝑬𝑪𝑬

💬 𝐋𝐞𝐜𝐭𝐮𝐫𝐞 & 𝐖𝐨𝐫𝐤𝐬𝐡𝐨𝐩 𝐨𝐧 𝐓𝐨𝐰𝐚𝐫𝐝 𝐀𝐮𝐭𝐨𝐧𝐨𝐦𝐨𝐮𝐬 𝐙𝐞𝐫𝐨-𝐓𝐨𝐮𝐜𝐡 𝟔𝐆 𝐍𝐞𝐭𝐰𝐨𝐫𝐤𝐬: 𝐀𝐈-𝐍𝐚𝐭𝐢𝐯𝐞 𝐌𝐚𝐧𝐚𝐠𝐞𝐦𝐞𝐧𝐭 𝐰𝐢𝐭𝐡 𝐋𝐚𝐫𝐠𝐞 𝐋𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐌𝐨𝐝𝐞𝐥𝐬

Beyond delivering unprecedented performance, the sixth-generation (6G) mobile network is expected to simultaneously support sub-millisecond latency, terabit-per-second data rates, integrated communication and sensing, and massive service heterogeneity. These stringent requirements create an unprecedented management complexity, where the configuration and optimisation space grows exponentially, rendering conventional manual provisioning and static rule-based management ineffective. Moreover, optimisation must occur across multiple temporal scales, from real-time physical-layer adaptation to long-term service orchestration, creating a hierarchical control problem beyond human operational capabilities. To address these challenges, 6G adopts an AI-native architecture in which intelligence is embedded directly within network functions rather than deployed as an external optimisation layer. Such a paradigm enables continuous network adaptation and provides support for Zero-Touch Service Management (ZSM), which aims to automate the entire network lifecycle with minimal human intervention. While autonomous closed-loop controllers can optimise localised network functions, coordinating distributed intelligence requires higher-level reasoning. Large Language Models (LLMs) have emerged as promising cognitive orchestrators capable of...

📅 𝗗𝗮𝘁𝗲: July 23, 2026 (Thursday)
⏰ 𝗧𝗶𝗺𝗲: 9:30 am – 10:15 am (Lecture) & 10:15 am – 11:00 am (Workshop)
📍 𝗩𝗲𝗻𝘂𝗲: CB-603, 6/F, Chow Yei Ching Building, The University of Hong Kong
🔗 𝗗𝗲𝘁𝗮𝗶𝗹𝘀: https://ece.hku.hk/events/20260723-1

𝑭𝒐𝒍𝒍𝒐𝒘 𝑰𝑮/𝑭𝑩: 𝑯𝑲𝑼𝑬𝑪𝑬

💬 𝐒𝐞𝐦𝐢𝐧𝐚𝐫 𝐨𝐧 𝐑𝐞𝐜𝐨𝐧𝐟𝐢𝐠𝐮𝐫𝐚𝐛𝐥𝐞 𝐃𝐢𝐬𝐭𝐫𝐢𝐛𝐮𝐭𝐞𝐝 𝐀𝐧𝐭𝐞𝐧𝐧𝐚𝐬 𝐚𝐧𝐝 𝐑𝐞𝐟𝐥𝐞𝐜𝐭𝐢𝐧𝐠 𝐒𝐮𝐫𝐟𝐚𝐜𝐞 (𝐑𝐃𝐀𝐑𝐒): 𝐀 𝐍𝐞𝐰 𝐀𝐫𝐜𝐡𝐢𝐭𝐞𝐜𝐭𝐮𝐫𝐞 𝐟𝐨𝐫 𝐖𝐢𝐫𝐞𝐥𝐞𝐬𝐬 𝐂𝐨𝐦𝐦𝐮𝐧𝐢𝐜𝐚𝐭𝐢𝐨𝐧𝐬 𝐚𝐧𝐝 𝐒𝐞𝐧𝐬𝐢𝐧𝐠

**Due to unforeseen circumstances, this seminar has been rescheduled to July 17, 2026 (Friday), from 3:00 pm to 4:00 pm.**

A new architecture, “Reconfigurable Distributed Antennas and Reflecting Surface (RDARS)”, will be introduced for future 6G wireless communications and sensing in this talk. Specifically, RDARS is a flexible combination of Distributed Antenna System (DAS) and Reconfigurable Intelligent Surface (RIS). It inherits the low-cost and low-energy-consumption benefits of fully-passive RISs by default configuring all the elements as passive to perform the reflection mode. On the other hand, based on the design of the additional direct-through state, any element of the RDARS can be dynamically programmed to connect with the base station (BS) via fibers/wires and perform the connected mode as remote distributed antennas of the BS to transmit/receive signals. As such, this novel architecture exploits the benefits of both RIS and DAS with a controllable trade-off between the “reflection gain” and the “distribution gain” achieved via RDARS at the BS. Moreover, additional "selection gain" can be achieved from the reconfigurability of the operation mode for each element. Outage probability and ergodic achievable rate under the maximum ratio combining (MRC) scheme at BS are analysed in closed-forms to characterise the system behaviour of the RDARS-aided system...

📅 𝗗𝗮𝘁𝗲: July 17, 2026 (Friday)
⏰ 𝗧𝗶𝗺𝗲: 3:00 pm – 4:00 pm
📍 𝗩𝗲𝗻𝘂𝗲: CB-603, 6/F, Chow Yei Ching Building, The University of Hong Kong
🔗 𝗗𝗲𝘁𝗮𝗶𝗹𝘀: https://ece.hku.hk/events/20260715-1

𝑭𝒐𝒍𝒍𝒐𝒘 𝑰𝑮/𝑭𝑩: 𝑯𝑲𝑼𝑬𝑪𝑬