Events

Seminar on "Optimal adaptive output regulation for discrete-time nonlinear stochastic systems" by Dr Zhaobo Liu

Date: 18 October 2024, Friday Time: 4:30 pm Venue: YEUNG-B5311, City University of Hong Kong Zoom Meeting ID: 859 2865 1236 Password: 123456 Speaker: Dr Zhaobo Liu, Shenzhen University   Optimal adaptive output regulation for discrete-time nonlinear stochastic systems Dr Zhaobo Liu, Shenzhen University     Abstract: In this presentation, we address the output regulation problem associated with a basic class of discrete-time nonlinear stochastic systems with unknown parameters. We allow the controlled plant to exhibit highly nonlinear growth, as well as nonlinearly parameterized structures in the exosystem. Our main purpose is to design an adaptive regulator to achieve the optimal performance of the regulated output. The design of the regulator is an integration of the recursive least squares estimator and a modified nonlinear least squares (NLS) algorithm. Under certain mild conditions, it is shown that with stable exosystems, the square of the norm of the closed-loop regulated output is asymptotically optimal in the average sense almost surely. Specifically, we propose a new formula for the strong convergence rate of NLS and derive an almost optimal strong convergence rate of the proposed modified NLS algorithm. Additionally, we employ novel closed-loop analysis techniques to overcome the Bayesian assumptions required in existing research on the stabilizability of discrete-time nonlinear stochastic systems.   Speaker’s Bio:Zhaobo Liu is currently an Assistant Professor at the Institute for Advanced Study, Shenzhen University. He received the B.Sc. degree from the School of Mathematical Sciences, Peking University, in 2015, and the Ph.D. degree from the Academy of Mathematics and Systems Science, Chinese Academy of Sciences, in 2020. His current research interests include adaptive control, data science, and computational intelligence. WEBINAR WEBSITE: http://cccn.ee.cityu.edu.hk/webinar/    

Seminar on "Can Graph Neural Networks Become More Interpretable?" by Dr Giorgia Minello

Date: 16 October 2024, Wednesday Time: 11:00am Venue: CD634, Department of Electrical and Electronic Engineering, PolyU Speaker: Dr Giorgia Minello, Shenzhen University

Seminar on “Encapsulation of perovskite solar cells” conducted by Prof. Aleksandra B. Djurišić.

Date: 7 October 2024 (Mon) Time: 3:30pm-5:00pm Venue: CD634, Department of Electrical and Electronic Engineering, PolyU Speaker: Prof. Aleksandra B. Djurišić, Department of Physics, The University of Hong Kong

Seminar on "Differentially Private Graph Neural Networks for Link Prediction" by Mr Xun Ran

Date: 4 October 2024, Friday Time: 4:30 pm Venue: CD634, The Hong Kong Polytechnic University Zoom Meeting ID: 383 735 6917 Password: 270831 Speaker: Mr Xun Ran, The Hong Kong Polytechnic University Differentially Private Graph Neural Networks for Link Prediction Mr Xun Ran, The Hong Kong Polytechnic University   Abstract:Graph Neural Networks (GNNs) have proven to be highly effective in addressing the link prediction problem. However, the need for large amounts of user data to learn representations of user interactions raises concerns about data privacy. While differential privacy (DP) techniques have been widely used for node-level tasks in graphs, incorporating DP into GNNs for link prediction is challenging due to data dependency. To this end, in this work we propose a differentially private link prediction (DPLP) framework, building upon subgraph-based GNNs. DPLP includes a DP-compliant subgraph extraction module as its core component. We first propose a neighborhood subgraph extraction method, and carefully analyze its data dependency level. To reduce this dependency, we optimize DPLP by integrating a novel path subgraph extraction method, which alleviates the utility loss in GNNs by reducing the noise sensitivity. Theoretical analysis demonstrates that our approaches achieve a good balance between privacy protection and prediction accuracy, even when using GNNs with few layers. We extensively evaluate our approaches on benchmark datasets and show that they can learn accurate privacy-preserving GNNs and outperforms the existing methods for link prediction.   Speaker’s Bio:Mr Xun RAN is currently pursuing his Ph.D. degree in the Department of Electrical and Electronic Engineering, The Hong Kong Polytechnic University, Hong Kong. His research interests include data privacy and neural networks. WEBINAR WEBSITE: http://cccn.ee.cityu.edu.hk/webinar/

Seminar on "Mitigating the Structural Bias in Graph Adversarial Defenses" by Mr Junyuan Fang

Date: 27 September 2024, Friday Time: 4:30 pm Venue: YEUNG-B5311, City University of Hong Kong Zoom Meeting ID: 859 2865 1236 Password: 123456 Speaker: Mr Junyuan Fang, City University of Hong Kong Mitigating the Structural Bias in Graph Adversarial Defenses Mr Junyuan Fang, City University of Hong Kong   Abstract: In recent years, graph neural networks (GNNs) have shown great potential in addressing various graph structure-related downstream tasks. However, recent studies have found that current GNNs are susceptible to malicious adversarial attacks. Given the inevitable presence of adversarial attacks in the real world, a variety of defense methods have been proposed to counter these attacks and enhance the robustness of GNNs. Despite the commendable performance of these defense methods, we have observed that they tend to exhibit a structural bias in terms of their defense capability on nodes with low degree (i.e., tail nodes), which is similar to the structural bias of traditional GNNs on nodes with low degree in the clean graph. Therefore, in this talk, we propose a defense strategy by including hetero-homo augmented graph construction, GNN augmented graph construction, and multi-view node-wise attention modules to mitigate the structural bias of GNNs against adversarial attacks. Notably, the hetero-homo augmented graph consists of removing heterophilic links (i.e., links connecting nodes with dissimilar features) globally and adding homophilic links (i.e., links connecting nodes with similar features) for nodes with low degree. To further enhance the defense capability, an attention mechanism is adopted to adaptively combine the representations from the above two kinds of graph views. We conduct extensive experiments to demonstrate the defense and debiasing effect of the proposed strategy on benchmark datasets.   Speaker’s Bio:Junyuan Fang received the B.Eng. degree in software engineering from the Guangdong University of Technology in 2018, the M.Eng. degree in software engineering from Sun Yat-sen University in 2020, and the Ph.D. degree in electrical engineering from the City University of Hong Kong in 2024. His current research interests include robustness analysis and optimization, applications of network science, and graph mining.   WEBINAR WEBSITE: http://cccn.ee.cityu.edu.hk/webinar/

Seminar on "Novel hybrid advanced traction power supply system and high-quality and efficient energy management technology" by Dr Xin Wang

Date: 13 September 2024, Friday Time: 4:30 pm Venue: YEUNG-B5311, City University of Hong Kong Zoom Meeting ID: 859 2865 1236 Password: 123456 Speaker: Dr Xin Wang, City University of Hong Kong   Novel hybrid advanced traction power supply system and high-quality and efficient energy management technology Dr Xin Wang, City University of Hong Kong   Abstract: At present, the problems of neutral section, power quality, and low energy utilization rate are the main constraints that hinder the development of traditional traction power supply systems. As a better power supply mode, the advanced traction power supply system (ATPSS) integrating with photovoltaics (PV) and energy storage (ES) based on power electronics is proposed, which provides new opportunities for simultaneously solving the above issues. Therefore, a novel hybrid-ATPSS and high-quality and efficient energy management technology is discussed in this seminar, and it will be introduced from four aspects: advanced traction power supply device, control technology, fault-tolerant operation, multi-timescale efficient energy management.   Speaker’s Bio: Xin Wang is currently an Associate Researcher at City University of Hong Kong. He was born in Henan, China, 1997. He received the B.S. degree in electrical engineering and automation from Northeast Agricultural University, Harbin, China, in 2019 and the Ph.D. degree in electrical engineering from Hunan University, Changsha, China, in 2024. His current research interests include flexible interconnection device of distribution network, advanced traction power supply system. WEBINAR WEBSITE: http://cccn.ee.cityu.edu.hk/webinar/

Seminar on "Differentially Private Graph Neural Networks for Link Prediction" by Mr Xun Ran

Date: 6 September 2024, Friday Time: 4:30 pm Venue: CD634, The Hong Kong Polytechnic University Zoom Meeting ID: 270 294 7238 Password: 123456 Speaker: Mr Xun Ran, The Hong Kong Polytechnic University   Differentially Private Graph Neural Networks for Link Prediction Mr Xun Ran, The Hong Kong Polytechnic University   Abstract: Graph Neural Networks (GNNs) have proven to be highly effective in addressing the link prediction problem. However, the need for large amounts of user data to learn representations of user interactions raises concerns about data privacy. While differential privacy (DP) techniques have been widely used for node-level tasks in graphs, incorporating DP into GNNs for link prediction is challenging due to data dependency. To this end, in this work we propose a differentially private link prediction (DPLP) framework, building upon subgraph-based GNNs. DPLP includes a DP-compliant subgraph extraction module as its core component. We first propose a neighborhood subgraph extraction method, and carefully analyze its data dependency level. To reduce this dependency, we optimize DPLP by integrating a novel path subgraph extraction method, which alleviates the utility loss in GNNs by reducing the noise sensitivity. Theoretical analysis demonstrates that our approaches achieve a good balance between privacy protection and prediction accuracy, even when using GNNs with few layers. We extensively evaluate our approaches on benchmark datasets and show that they can learn accurate privacy-preserving GNNs and outperforms the existing methods for link prediction.   Speaker’s Bio: Mr Xun RAN is currently pursuing his Ph.D. degree in the Department of Electrical and Electronic Engineering, The Hong Kong Polytechnic University, Hong Kong. His research interests include data privacy and neural networks. WEBINAR WEBSITE: http://cccn.ee.cityu.edu.hk/webinar/

EEE Inaugural Banquet

Celebrating the successful merger of the Department of Electrical Engineering and the Department of Electronic and Information Engineering, the Inaugural Banquet of the newly formed Department of Electrical and Electronic Engineering (EEE) was a grand affair, graced by the distinguished presence of PolyU Council Members, Senior Management, Chair Professors, and key executives from leading engineering firms in Hong Kong. The event not only commemorated the rich heritage of both departments but also marked the beginning of a new chapter filled with exciting possibilities and collaborations. Under the guidance of Professor C.Y. Chung, the Head of the EEE Department, attendees were treated to a comprehensive overview of the Department’s educational, research, and knowledge transfer initiatives. The banquet served as a platform to honour former Heads of Department and alumni, acknowledging their contributions to the Department’s legacy. Throughout the evening, guests were captivated by a series of engaging presentations, including a custom-made video depicting the Department’s history, a splendid Chinese painting symbolizing unity and progress, story-sharing between an alumnus and a current student and a delightful musical performance by talented students and a staff member.  The event showcased students’ talents and promising futures as they embarked on their academic and professional paths.

Quantum Hong Kong 2024

Seminar by Prof. Chi-Man Pun on "Image Manipulation Localization with Deep Neural Networks"

Date: 18 July 2024, Thursday Time: 11:00 am Venue: Z505, The Hong Kong Polytechnic University Speaker: Prof. Chi-Man Pun, Department of Computer and Information Science, University of Macau