Proposed TopicSpeaker
DeepVariant and DeepConsensusDr. Pi-Chuan CHANG, Staff Software Engineer, Genomics team in Google Research
Urbantech and Social InnovationMr. Mingles Tsoi, ParticleX
Smart HomeMr. Mike SO, Votion
Social impact of life and death education (LDE) in quality of life (QoL)Ms. Faye Chan, Society for Life and Death Education
Reshaping physical world with digital technologiesMr. Weilan Zhang, China Design Digital Technology, Director of AEC Digital Solution
Flexible Ultrasound Sensor for Nonocclusive and Continuous Blood Pressure MonitoringDr. Chang Peng, School of Biomedical Engineering, ShanghaiTech University
Building Sustainable Power Systems: Innovative Ideas from Female Researchers
Field Ambient signal based load modeling and its application in power systemsDr. Ying Wang, Postdoc, the University of Hong Kong
Renewable Energy Planning Under Decision-Dependent UncertaintiesDr. Wenqian Yin, Postdoc, the University of Hong Kong
Distributed Control and Operation of Decarbonized and Informatized Low-
Voltage Networks with Multi-Energy Resources
Dr. Qian Hu, Postdoc, Hong Kong Polytechnic University
A Distributionally Robust Resilience Enhancement Strategy for Distribution
Grids Considering Decision-Dependent Contingencies
Dr. Yujia Li, Postdoc, the University of Hong Kong
Topic:
DeepVariant and DeepConsensus
Presenter:
Pi-Chuan Chang, Staff Software Engineer, Genomics team in Google Research

Abstract:

Our team has open sourced two deep learning based bioinformatics tools – DeepVariant and DeepConsensus.

DeepVariant is a germline variant caller on diploid organisms. In 2016 and 2020 Precision FDA Truth Challenge v1 and v2, DeepVariant has won several awards for highest accuracy on several different sequencing instruments. DeepVariant has also been used on large-scale genetic studies such as UK Biobank, and its high accuracy has been essential in clinical research studies.

DeepConsensus (published on Nature Biotechnology in Sept 2022) is a new method for correcting sequencing errors on Pacific BioScience (PacBio) circular consensus sequencing (CCS) data. We demonstrate that using reads from DeepConsensus improves the quality of genome assembly, and improves accuracy of variant calling.

By formulating these bioinformatics problems into machine learning tasks, we are able to achieve high accuracy. In addition, for DeepVariant, machine learning also allows us to quickly adapt DeepVariant models to new sequencing instruments.

This talk will describe these bioinformatics problems, the machine learning approaches we used, and the impact of the tools we’ve built.

Biography:

Pi-Chuan Chang is a Staff Software Engineer at the Genomics team in Google Research, leading the open source projects DeepVariant and DeepConsensus.

Pi-Chuan received the B.S. and M.S. degrees from National Taiwan University, and the Computer Science Ph.D. degree from Stanford, specializing in natural language processing and machine learning. After Stanford, she worked at several tech companies (Google, LinkedIn, AltSchool), applying machine learning to various domains, including YouTube video personalization, features in search, semantic parsing for voice actions, and improved sequence models in LinkedIn search. In the past 5 years, Pi-Chuan has focused her work on the intersection of applied machine learning and genomics.


Topic:
Urbantech and Social Innovation
Presenter:
Mingles Tsoi , ParticleX

Abstract:

During the outbreak of COVID-19 in 2020, ParticleX was aware of the challenges faced by both big corporates and startups, therefore initiating the ParticleX Proptech Global Challenge in order to connect both parties with a strategic matching approach for collaboration. Somehow this approach is proven to be sustainable in nurturing innovation for smart city initiatives, and we explore the wider scope of urbantech which should create more social impact and even innovation to the cities. With his extensive background and hands-on experience in social entrepreneurial research, education, consultancy, and currently as an angel investor, Mingles Tsoi, CXO of ParticleX will share insights on how technologies can generate social innovation, especially on urbantech covering areas from energy, building, food, transportation to education.

Biography:

MinglesMingles Tsoi is currently the CXO (Chief eXploration Officer) of ParticleX, an early stage to Pre-A round institutional investment fund aiming at startups with hands-on technology initiative, proven customer-problem-fit and problem-solution-fit achievements. His leadership in providing unique post-investment support services has gained much reputation from the portfolio companies. Mingles was an entrepreneur owning his own consultancy company since 1997. He is also a pioneer in startup education. He had been working with the CUHK Center for Entrepreneurship as Project Director since 2007. Some of his major accomplishments include the “Hong Kong Social Enterprise Challenge”, the first CU Alumni Entrepreneur Census to investigate the success factors of leading entrepreneurs from local university education and the Google’s Empowering Young Entrepreneur (EYE) Program. He was invited to join KPMG China as the Director since 2015, to set up the Innovation and Startup Centre in Beijing. Mingles is also enthusiastic in youth development. He contributes his experience and knowledge in providing voluntary mentoring and business training services to various non-profit organizations. He was the Co-opted Members of the Working Group of Youth Development Fund of the Commission on Youth, and as Non-official Member of the Social Enterprise Advisory Committee respectively. He is currently serving on several boards and committees, including Hong Kong Business Angel Network (HKBAN), Cyberport Investors Network (CIN), DIT Committee of the Hong Kong General Chamber of Commerce, Vice Chairman of the Startup Council, Executive Committee Members of HKITIC and HKPSC of the Federation of Hong Kong Industries (FHKI), Member of Product & Technology Development Panel of the Hong Kong Medical and Healthcare Device Industries Association Limited (HKMHDIA), Founding and Council Member of the Lions Club HK IFC, and the Chairman of the Academic and Accreditation Advisory Committee of the Institute of Financial Technologists of Asia (IFTA). He is a CFT® holder, a GCDF, and fellow members of accountancy qualifications of IFA, IPA and CMA.


Topic:
Social Impact of life and death education (LDE) in quality of life (QoL)
Presenter:
Faye Chan, Society for Life and Death Education

Abstract:

Life and Death Education (LDE) is educational programs and activities that address topics associated with dying and death with the intention of identifying and interrupting negative emotions, assumptions, and beliefs. Topics may include taboos, silences, bereavement, grief, life stages, illness, as well as end-of-life matters, such as advance care planning (ACP) and survivor accommodations. In this talk, the focus is on ACP to illustrate how it can contribute to people’s QoL. ACP is the process of communication among a patient with advance progressive diseases, his/her health care providers, and his/her family members regarding the kind of care that will be considered appropriate when the patient becomes mentally incompetent. In Hong Kong, ACP is relatively new to the public. However, older adults and patients with a progressive and life limiting illness are beginning to embrace the concept. As such, frontline healthcare professionals, family members and public need to have a good understanding of the concepts as well as the skills to initiate such conversation. Early discussions on end-of-life care can facilitate positive outcomes in improving the quality of life of the person. Early engagement of patients with a life limiting condition in advance care planning can improve patient and family satisfaction. A real-life case will be used to illustrate how ACP can maintain and enhance the QoL of the stakeholders involved.

Biography:

MSocSc in Behavioral Health, The University of Hong Kong
BSc in Nursing, University of Saskatchewan, Canada
Fellow Hong Kong Academy of Nursing (Education & Research)
Fellow in Thanatology: Death, Dying and Bereavement, USA
Registered Nurse (HK, US, and Canada)

Mrs. Chan is one of the pioneers in promoting life and death education in Hong Kong. She is Chairperson of the Society for Life and Death Education. She is a Fellow in Thanatology: Death, Dying & Bereavement, and is also a member of the International Work Group on Death, Dying & Bereavement. Mrs. Chan is a seasoned educator in palliative care/end-of-Life care training for the healthcare and social welfare sectors. She is Nurse Trainer of the Capacity Building and Education Program in End-of-Life Care, CUHK Institute of Ageing. She is an academic supervisor of Master in Clinical Gerontology and End-of-Life Care, CUHK.


Topic:
Reshaping physical world with digital technologies
Presenter:
Weilan Zhang, China Design Digital Technology, Director of AEC Digital solution 

Abstract:

Construction is the largest global industry that accounts for 13% of global GDP. It encompasses infrastructure, industrial structures, and buildings. The industry has major impact on environmental protection, carbon emission and energy consumption. During the last decade, the construction industry is embracing digital technology progression, such as building information modeling (BIM),virtual design and construction(VDC), off-site construction, city information modeling(CIM) and so on, the physical world in future will be reshaped by digital technologies.

This talk will discuss the how the digital technologies will transform the design, construction and operation of our buildings and cities in near future. Both opportunities and challenges are ahead of us to explore and resolve. The goal is to reach a sustainable development between economy, society and environment.

Biography:

Weilan ZhangWeilan Zhang is a registered architect of US, a member of American Institute of Architect (AIA) and Accredited Professional of Leadership in Energy and Environment Design (LEED AP). He received his Bachelor of Architecture degree from Tianjin University China and Master of Architecture degree and Certificate in Business Administration from University of Washington Seattle USA. Mr. Weilan Zhang had been working on number of prominent landmark projects across America and China, including Expedia Headquarters, UW Bioengineering center, China National Tennis Center, and Zhengzhou Baoneng Center. He is currently the director of BIM solution department at China Digital Design Technology. His expertises are on Building Information Modeling (BIM), sustainable building design, prefabrication in architecture and digital transformation in design and construction.

 

 


Topic:
Flexible Ultrasound Sensor for Nonocclusive and Continuous Blood Pressure Monitoring
Presenter:
Chang Peng, School of Biomedical Engineering, ShanghaiTech University, China

Abstract:

High blood pressure has become a prevalent and growing public health problem worldwide and is currently the most common cause of death at the global level. Continuous blood pressure monitoring in everyday life is important and necessary to detect and control high blood pressure in advance. While the existing blood pressure monitoring techniques are well suited for application in current clinical settings, they are inadequate for next-generation wearable long-term monitoring of blood pressure on a daily basis. In this talk, a flexible piezo-composite ultrasonic sensor will be introduced first for continuous blood pressure measurement through ultrasonic blood vessel wall motion tracking that eliminates an inflatable cuff. Continuous blood pressure measurement will then be conducted in the ulnar artery on a volunteer’s right arm, which is based on tracking the arterial diameter variations during cardiac cycles by transmitting and receiving ultrasonic signals from the sensor. Finally, I will provide an overview of future research directions and a range of applications of flexible ultrasound sensors.

Biography:

Chang PengDr. Chang Peng is an Assistant Professor in the School of Biomedical Engineering at ShanghaiTech University. He received his Ph.D. degree from the University of Florida, USA (2014-2018). After that, he was a Postdoc Research Scholar at the North Carolina State University, USA (2018-2021). His research interests mainly focus on the development of ultrasonic transducers for medical applications, flexible ultrasonic devices, and intravascular ultrasound imaging. Dr. Peng has been published more than 30 papers, including 14 first-author peer reviewed papers in top-level international journals. He is also the author of 1 book chapter, 4 invited review publications, 3 applied US Patents, and 2 issued Chinese Patents.

 

 


Building Sustainable Power Systems:
Innovative Ideas from Female Researchers

 

The wide deployment of distributed energy resources (DERs), such as rooftop PV panels and electric vehicles, helps pave our way towards a more sustainable, affordable, and flexible electric grid. But meanwhile, the fluctuating renewable sources and unpredictable demands pose a great challenge to the power system security. Many female researchers have put forth innovative ideas and solutions to address these challenges. In this panel, we invite four female researchers from different universities in Hong Kong to share their latest findings. The topics span power system modeling, operation, resilience, and planning. We believe that it can provide useful insights into sustainable power systems development. The potential presenters are as follows:

Chair: Dr. Yue Chen, Assistant Professor, the Chinese University of Hong Kong
Co-Chair: Dr. Rui Xie, Postdoc, the Chinese University of Hong Kong


Topic:
Field Ambient Signal Based Load Modeling and its Application in Power Systems
Presenter:
Dr. Ying Wang, Postdoc, the University of Hong Kong

Abstract:

Load modeling is an important but difficult task in power system modeling and analysis. On the one hand, applying an improper load model may result in inaccurate results. On the other hand, power loads are complex, time-varying, and distributive. The characteristics of power loads for different substations at different times are various. Thus, establishing accurate load models, especially describing the time-varying characteristics of power loads, is both necessary and challenging. With the rapid development of synchrophasor measurement techniques, the measurement data of phasor measurement units (PMUs) can accurately capture the dynamics of power systems in real-time, which can be applied as an important data source for load model parameter identification. Besides, the previous research indicates that the ambient signals, which mean the measurements under normal operating conditions, can be used for electromechanical dynamic analysis. On this basis, the ambient signal based load modeling method was proposed, which indicates that load model parameters can be identified from ambient signals. Then lots of research has been carried out in the ambient signal based load model parameter identification.


Topic:
Renewable Energy Planning Under Decision-Dependent Uncertainties
Presenter:
Dr. Wenqian Yin, Postdoc, the University of Hong Kong

Abstract:

Driven by energy transition, renewable energy is expected to replace fossil energy as the dominant power source in the future. The growing integration of weather-reliant renewable resources will inevitably increase uncertainties in renewable-dominated power systems and aggravate the inverse impacts of planning/operation decisions on uncertainty features. Nevertheless, incorporating such decision-dependent uncertainty (DDU) would alter the traditional optimization paradigm that dedicates to only exogenous uncertainties, necessitating a recreation of modeling and solution methodologies. In this presentation, we will focus on methodologies of stochastic renewable energy planning under DDU. We will identify how renewable energy planning decisions would alter some uncertainty features and distinguish such DDU from exogenous uncertainties. Stochastic expansion planning methods for renewable energy and flexibility resources considering such DDU will be presented. The mathematical structural features of the established optimization models with DDU will be exploited, and efficient solution methods tackling the coupling relation between decisions and DDU will be developed accordingly. The value of such DDU in coordinated planning of renewable energy and flexibility resources will be presented. This work will provide insights for constructing renewable-dominated systems in the presence of DDU.

Topic:
Distributed Control and Operation of Decarbonized and Informatized Low-Voltage Networks with Multi-Energy Resources
Presenter:
Dr. Qian Hu, Postdoc, Hong Kong Polytechnic University

Abstract:

Low-voltage networks are undergoing a significant transformation of decarbonization with the growing integration of renewable energy resources. Meanwhile, the deployment of information and communication technology empowers the informatization of low-voltage networks to support intelligent functionalities. However, despite benefits achieved under transformations of decarbonization and informatization, new challenges are posed which cannot be tackled efficiently and effectively by using centralized control and operational schemes. To address arising challenges, the research focuses on the reconstruction of control strategies and operational schemes for future decarbonized and informatized low-voltage networks.


Topic:
A Distributionally Robust Resilience Enhancement Strategy for Distribution Grids Considering Decision-Dependent Contingencies
Presenter:
Dr. Yujia Li, Postdoc, the University of Hong Kong

Abstract:

When performing resilience enhancement for distribution grids, there are two obstacles for reliably modeling uncertain contingencies: 1) decision-dependent uncertainty (DDU) resulting from different line hardening decisions, and 2) distributional ambiguity due to limited outage information under extreme weather events (EWEs). To address these two challenges, scenario-wise decision-dependent ambiguity sets (SWDD-ASs) are proposed, where the DDU and distributional ambiguity inherent in EWE-induced contingencies are simultaneously captured. Then, based on SWDD-ASs, a two-stage trilevel decision-dependent distributionally robust resilient enhancement (DD-DRRE) model is formulated. Subsequently, by equivalent reformulating the DD-DRRE model, a customized column-and-constraint generation (C&CG) algorithm is utilized to derive its optimal solutions. Finally, numerical tests demonstrate a remarkable improvement in the out-of-sample performance of our model, compared to its prevailing stochastic and robust counterparts.