Events

Seminar on "Programmable Photonic Circuits for Artificial Neural Networks and Quantum Machine Learning" by Dr Hui Zhang

Date: 13 October 2023, Friday Time: 4:30pm Venue: CD634, The Hong Kong Polytechnic University Zoom Meeting ID: 270 294 7238 Password: 123456 Speaker: Dr Hui Zhang, Hong Kong Polytechnic University   Programmable Photonic Circuits for Artificial Neural Networks and Quantum Machine Learning Dr Hui Zhang, Hong Kong Polytechnic University   Abstract: Photonic neural networks is a revolutionary paradigm in the field of artificial intelligence, harnessing the power of photonics to accelerate the computational speed and energy efficiency of neural networks to previously unimaginable levels. This talk introduces the chip design, algorithm innovation, and applications of silicon-based on-chip photonic neural networks. To give full play to the advantages of photonic computing, a complex-valued neural network architecture with high prediction accuracy, fast convergence and nonlinear decision boundaries, as well as an online training method, was proposed. Subsequently, a photonic quantum autoencoder with single-shot online training is demonstrated, which achieves successful application in the high-dimensional quantum subspace teleportation. On-chip photonic neural networks promise to unlock new frontiers in high-performance computing, enabling advancements in fields such as autonomous driving, medical diagnosis, and quantum computing.   Speaker’s Bio: Dr Hui Zhang is currently a research assistant professor with The Hong Kong Polytechnic University. She received the B.Eng. degree from Xi’an Jiaotong University, China and the Ph.D. degree from Nanyang Technological university, Singapore. Her research interests include photonic neural networks, quantum machine learning and AI in science. In recent years, she has developed the complex-valued photonic neural network architecture, gradient-free on-chip training method, and the quantum autoencoder facilitated teleportation protocol. Her works are published on Science Advances, Nature Communications, Laser & Photonics Reviews, PRX Quantum etc.

Seminar on "Mitigating Cascading Failure in Power Networks Based on Deep Reinforcement Learning" by Dr Xi Zhang

Date: 6 October 2023, Friday Time: 4:30pm Venue: YEUNG Building Y5-303, City University of Hong Kong Zoom Meeting ID: 967 1968 0176 Password: 123456 Speaker: Dr Xi Zhang, Beijing Institute of Technology, Beijing   Mitigating Cascading Failure in Power Networks Based on Deep Reinforcement Learning Dr Xi Zhang, Beijing Institute of Technology, Beijing   Abstract: Power grids are susceptible to cascading failures, which can have catastrophic consequences for modern society. Remedial actions, such as proactive islanding, generator tripping, and load shedding, offer a viable solution to mitigate cascading failures in power grids. The key to the success of this solution lies in generating timely and appropriate remedial actions during the rapid propagation process of cascading failure. This presentation introduces an intelligent method that leverages deep reinforcement learning to generate remedial actions. A simulation model of cascading failure is presented, considering power flow distributions and the probabilistic failure mechanisms of components to accurately depict the dynamic cascading failure process. Based on this model, a Markov decision process is formulated to address the problem of selecting appropriate remedial actions during the cascading failure process. The Proximal Policy Optimization algorithm is utilized to generate effective remedial actions. Experiments are conducted on representative power test cases, and the results demonstrate the proposed method’s ability to significantly reduce the affected area in the event of the same initial failure, thereby showcasing its advantages in mitigating cascading failure in power grids.   Speaker’s Bio: Dr Xi Zhang received the BEng degree in Automation from Beijing Jiaotong University, Beijing, China, in 2013, and the PhD degree in Electronic and Information Engineering from The Hong Kong Polytechnic University, Hong Kong SAR, China, in 2017. He is currently an assistant professor at the School of Automation, Beijing Institute of Technology, Beijing, China. His research interests focus on power system resilience, application of AI technologies to the analysis and decision-making in power networks, etc. He is an Associate Editor for IEEE Transactions on Circuits and Systems II: Express Briefs. He served as a Guest Editor for IEEE Journal on Emerging and Selected Topics in Circuits and Systems in 2023 and the secretary of the IEEE PES Beijing Chapter from 2018 to 2019.

Seminar on "UAV-Enabled Intelligent Transportation Systems (ITS) for Smart Cities" by Prof. Simon Pun

Date: 22 September 2023, Friday Time: 10:30am Venue: CD634, The Hong Kong Polytechnic University Zoom Meeting ID: 383 735 6917 Password: 274320 Speaker: Prof. Simon Pun, The Chinese University of Hong Kong (Shenzhen Campus)   UAV-Enabled Intelligent Transportation Systems (ITS) for Smart Cities Prof. Simon Pun, The Chinese University of Hong Kong (Shenzhen Campus)   Abstract: Intelligent transportation systems (ITS) is one of the defining components for smart cities. Thanks to their flexible mobility, autonomous operation, and outstanding communication capabilities, Unmanned Aerial Vehicles (UAV) have been well envisaged as one of the most promising technologies for future ITS. In this talk, the latest development of UAV-enabled ITS will be first reviewed with emphasis on the effective integration of UAV and Unmanned Ground Vehicles (UGV) in ITS. After that, some of our recent works on three research topics, namely the UAV-enabled UGV path planning, UAV-enabled aerial-ground communication networks and autonomous UAV landing, will be elaborated in depth. In this talk, we will extensively use videos to demonstrate the systems we have designed and implemented.   Speaker’s Bio: Dr. Simon Pun received his Bachelor degree from the Chinese University of Hong Kong (CUHK), Master degree from the University of Tsukuba, Japan and Ph.D. degree from University of Southern California (USC) in Los Angeles, U.S.A., respectively. He was a post-doctoral research associate at Princeton University, U.S.A from 2006 to 2008. Before he joined the Chinese University of Hong Kong, Shenzhen (CUHKSZ) in 2015, he held research positions at Huawei in NJ, U.S.A, the Mitsubishi Electric Research Labs (MERL) in Boston, U.S.A, and Sony in Tokyo, Japan. He is currently an Associate Professor with the School of Science and Engineering, CUHKSZ. His research interests include artificial intelligence (AI) Internet of Things (AIoT) and applications of machine learning in communications and satellite remote sensing. Prof. Pun has received best paper awards from the IEEE Vehicular Technology Conference 2006 Fall, the IEEE International Conference on Communication 2008, and the IEEE Infocom’09. He is the Founding Chair of the IEEE Joint Signal Processing Society-Communications Society Chapter, Shenzhen. He served as an Associate Editor for the IEEE Transactions on Wireless Communications from 2010 to 2014.

Seminar on "Resilience of Multimodal Public Transportation Infrastructure systems in Hong Kong" by Prof. Shauhrat Chopra

Date: 15 September 2023, Friday Time: 4:30 PM Venue: YEUNG Building Y5-303, City University of Hong Kong Zoom Meeting ID: 967 1968 0176 Password: 123456 Speaker: Prof. Shauhrat Chopra, City University of Hong Kong   Resilience of Multimodal Public Transportation Infrastructure systems in Hong Kong Prof. Shauhrat Chopra, City University of Hong Kong   Abstract: Cities worldwide are striving to enhance their resilience in order to withstand extreme weather events such as floods, storms, and extreme temperatures associated with climate change. Concurrently, these cities are aggressively pursuing their sustainability and climate change mitigation targets. This necessitates a rethinking of urban planning and a re-engineering of infrastructure systems in the urban built environment to ensure long-term sustainability and short-term resilience. In this talk, I will discuss resilience and sustainability of interconnected and interdependent infrastructure systems in the context of the multimodal urban public transport system. As it is one of the critical infrastructure systems, any disruption on the public transportation system, even if partial, could have a debilitating impact on a city’s health, security, and economy. The public transport system in Hong Kong, for example, accounts for 90% of passenger trips in a day, and any disruption could have catastrophic cascading impacts on the city’s economic and social interactions. This is particularly concerning considering the increasing instances of climate change-induced extreme weather events. A multi-dimensional quantitative resilience assessment framework, called the ‘Resilience Cycle’ will be discussed in the context of the Hong Kong public transport system. The utility of this framework in offering insights for improving the resilience of Hong Kong’s transportation system to disruptions will be discussed, while also considering its adaptability towards sustainable infrastructure. The talk will discuss the use of systems modelling approaches, such on network analysis and GIS, and tools like life cycle assessment (LCA) to address the potential trade-offs between resilience and sustainability in decision-making. Finally, I will stress the applicability of our findings from the study of resilient multimodal public transport systems in Hong Kong to the broader realm of real-world critical infrastructure systems.   Speaker’s Bio: Dr. Shauhrat S. Chopra is an Assistant Professor at the School of Energy and Environment (SEE), City University of Hong Kong (CityU). He received his PhD in Civil and Environmental Engineering from the Swanson School of Engineering at the University of Pittsburgh, USA, in 2015. Chopra’s doctoral dissertation was focused on resilience of complex systems, including economic systems, industrial symbiosis, and critical infrastructure systems at urban and national levels. Before joining SEE, he also worked as a Postdoctoral Researcher at the Institute for Environmental Science and Policy, University of Illinois at Chicago, on the U.S. EPA-funded project focused on sustainable design of future transformative emerging technologies. At CityU, his data-driven research is focused on designing indicators for sustainability and resilience of the built environment in support of environmental decision-making.  

Seminar by Prof Quek on "A Look at Federated Learning from Different Perspectives"

Date:  14 September 2023, Thursday   Time:  10:30 am   Venue: CD634 Speaker: Prof. Tony Q.S. Quek (Cheng Tsang Man Chair Professor, Singapore University of Technology and Design (SUTD))

The Technical Committee 2 of the International Association for Pattern Recognition

The Technical Committee 2 of the International Association for Pattern Recognition is offering a PhD Summer School on Deep Learning on Graphs to be held at The Hong Kong Polytechnic University on the 31st of August 2023. The event will cover topics ranging from graph neural networks for computer vision to quantum-inspired architectures and will include a hands-on session with PyTorch Geometric.   Event website: https://sites.google.com/view/tc2-dlg

The 1st TC2 Summer School on Deep Learning on Graphs concluded on PolyU campus on the last day of August.

The 1st TC2 Summer School on Deep Learning on Graphs concluded on PolyU campus on the last day of August. Organised by Technical Committee 2 (TC2) of the International Association for Pattern Recognition (IAPR), the event saw a team of five international speakers delivering six seminars to nine local and seven non-local students, selected out of a total of 25 applicants, evenly split between MSc and PhD levels.   Prof. Andrea Torsello from Ca' Foscari University of Venice kickstarted the event with a seminar covering the history of graph (deep) learning, highlighting the challenging issues faced by researchers in this area. Prof. Xiao Bai from Beihang University showed how graph neural networks can be applied to the processing of 3D computer vision data. Dr. Luca Rossi from PolyU drew the students’ attention to the importance of creating AI models whose predictions are interpretable and explainable. Finally, Prof. Edwin Hancock from University of York talked about the exciting opportunities lying at the intersection of quantum machine learning and graph analysis. The event also included two PyTorch coding sessions delivered by Dr Luca Cosmo from Ca' Foscari University of Venice where the students had an opportunity to put the concepts they had just learned into practice.   The event was free for the attendees and financially sponsored by the PolyU Department of Electrical and Electronic Engineering. The Summer School was the first of a series of similar events that TC2 plans to organise in alternate years with the Joint IAPR International Workshops on Statistical Techniques in Pattern Recognition and Structural and Syntactic Pattern Recognition (S+SSPR), a bi-annual event now in its 20th iteration jointly organised by TC1 and the TC2, respectively the first and second oldest technical committees of the IAPR.

Seminar by Prof. Angela Yingjun Zhang on "Task-oriented communications by maximum coding rate reduction"

Presenter: Prof. Angela Yingjun Zhang Date:  17 August 2023, Thursday   Time:  4:00 pm   Venue: Room CD634, Department of Electrical and Electronic Engineering, PolyU 

High-performance Perovskite Optoelectronic Devices via Defect Passivation & Potential Career Development Opportunities at Huaqiao University

Presenter: Prof. Zhanhua Wei Date: 16 August 2023 (Wednesday) Time: 9:30 a.m. Venue: Room CD634, EEE, PolyU

Optimally Scheduling Public Safety Power Shutoffs

Date: August 3, 2023 (THU) Time: 2p.m. to 5p.m. Venue: CF617 Speaker: Antoine LESAGE-LANDRY (Assistant Professor in the Department of Electrical Engineering at Polytechnique Montréal, QC, Canada)   Abstract In an effort to reduce power system-caused wildfires, utilities carry out public safety power shutoffs (PSPS) in which portions of the grid are de-energized to mitigate the risk of ignition. The decision to call a PSPS must balance reducing ignition risks and the negative impact of service interruptions. In this talk, we consider three PSPS scheduling scenarios, which we model as dynamic programs. In the first two scenarios, we assume that N PSPSs are budgeted as part of the investment strategy. In the first scenario, a penalty is incurred for each PSPS declared past the Nth event. In the second, we assume that some costs can be recovered if the number of PSPSs is below N while still being subject to a penalty if above N. In the third, the system operator wants to minimize the number of PSPS such that the total expected cost is below a threshold. We provide optimal or asymptotically optimal policies for each case, the first two of which have closed-form expressions. Lastly, we establish the applicability of the first PSPS model’s policy to critical-peak pricing, and obtain an optimal scheduling policy to reduce the peak demand based on weather observations.   Biography Antoine Lesage-Landry is an Assistant Professor in the Department of Electrical Engineering at Polytechnique Montréal, QC, Canada. He received the B.Eng. degree in Engineering Physics from Polytechnique Montréal, QC, Canada, in 2015, and the Ph.D. degree in Electrical Engineering from the University of Toronto, ON, Canada, in 2019. From 2019 to 2020, he was a Postdoctoral Scholar in the Energy & Resources Group at the University of California, Berkeley, CA, USA. His research interests include optimization, online learning, machine learning, and their application to power systems with renewable generation.   ALL ARE WELCOME For further information, please contact Dr. KOCAR Ilhan at ilhan.kocar@polyu.edu.hk , Department of Electrical and Electronic Engineering, The Hong Kong Polytechnic University