Project No.: R1012-21
Project Title: Secure and Incentivized Data Sharing with Enhanced Ownership for Decentralized Networks
Project Coordinator: Prof Xiaohua JIA (CityU)

Abstract

We are in a society that healthcare, business decisions, and government operations all heavily rely on the availability of data for accurate decision making. However, data is often fragmented and the fragments are stored locally, because data is usually collected and administrated by different institutions. To unleash the power of big data, we need to break the barriers for data owners to share their data out. The primary goal of project is to develop a secure and incentivised data sharing platform, so that data owners are encouraged to share their data out without worrying about data security and reap the benefits generated from their data.

The project consists of three major tasks as the following:

(1) Develop a secure and verifiable platform for data sharing authorization without a trusted central authority: (a) design a privacy-preserving and verifiable authorization framework; (b) design new techniques for the support joint authorization of multiple owners.

(2) Design secure and verifiable data analytics schemes with the enforcement of the data authorization: (a) design new techniques and algorithms for the secure computation of aggregate statistics; (b) design new schemes and algorithms for secure collaborative machine learning; (c) design new privacy-enhancing techniques for releasing the analytics results.

(3) Design quality-aware data valuation and privacy-preserving reward distribution schemes: (a) design new algorithms and frameworks for quality-aware and fair data valuation; (b) devise new techniques for privacy-preserving and verifiable reward distribution over decentralized ledger.

A prototype system will be developed to validate the proposed data sharing platform, security algorithms, and schemes.

The outcomes of this project include: (1) new models and frameworks for decentralized data sharing authorization and its enforcement; (2) new theories and algorithms for secure and verifiable analytics applications over encryption-protected data; and (3) new schemes on data valuation and scalable applications over decentralized ledger. These findings will greatly push forward the frontier of hot fields today, including: data ownership management, data valuation, applied cryptography, confidential computation, differential privacy, and decentralized ledger technologies.

Project No.: R4015-21
Project Title: A microrobotic platform with swarming magnetic microgels for endovascular embolization
Project Coordinator: Prof Li ZHANG (CUHK)

Abstract

Aneurysm refers to an abnormal, outward bulging or ballooning in the blood vessel wall, usually caused by a weak spot on blood vessels. As its size increases, internal bleeding may occur due to the rupture of the vessel, which can be lethal. Minimally invasive endovascular procedures are standard treatments for aneurysms, including embolization, stent placement and aortic aneurysm endograft repair. However, their long-term outcome is not satisfying due to the occurrence of complications after the procedures. To tackle the aneurysm, as well as the complications developed after the endovascular procedures, a new embolization strategy with a better clinical performance and long-term stability is desired.

Micro-/nanorobots are promising for minimally invasive medicine, such as active and targeted therapy. In this collaborative project, which exploits synergies between researchers and medical doctors from CUHK, HKU, CityU, ETH Zurich, Prince of Wales Hospital, Synapse Therapeutic Ltd., Cornerstone Robotics Ltd. and Vascular and Interventional Radiology Foundation, the team aims to address these key challenges and study functionalized magnetic swarming microgels for the image-guided endovascular intervention of embolization treatment. With the participation of clinical partners and industrial partners, the potential and preparation of a clinical translation of the proposed microrobotic platform will be considered in the early stage of the project. This interdisciplinary project by a research team with a global and industrial presence will generate outputs that provide fundamentally critical new data and references for the field. The proposed intervention will move toward a more efficient and stable embolization treatment with reduced side effects, significantly impacting the healthcare system and benefiting numerous patients with different aneurysm diseases.

Project No.: R4024-21
Project Title: Physical Activity Routines, Education, Assessment, Literacy, and Information Technology application in Young children (PA REALITY): A social cognitive theory-based movement education programme for preschool students
Project Coordinator: Prof Sau Ching Amy HA (CUHK)

Abstract

Engaging in appropriate volumes of physical activity is important to health. Being able to develop habits for physical activity during early years of life hence may have long lasting benefits throughout individuals’ life course. For preschool-aged children, the World Health Organization (WHO; 2019) recommends engagement in a range of activity of varying intensities (from light to vigorous intensities) for 180 minutes every day. However, studies showed that only 12.8% of children from Hong Kong met these recommendations. One key contributing factor to physical activity is one’s fundamental movement skills (FMS), such as running, jumping, throwing, and catching, which are considered the “building blocks” of many physical activities and sports. Having better FMS is related to more active lifestyles, better cardiovascular fitness and healthier weight statuses. Building children’s fundamental movement skills proficiency from a young age is thus important.

“Physical Development” is a development objective under the Hong Kong Kindergarten Education Curriculum. Hence all local preschools provide some forms of education to develop children’s FMS. However, most preschool teachers have not received formal training in physical education, and standardized curricula in the physical domain is currently unavailable. As such, preschool teachers’ ability to instruct and assess children’s development in these skills may be suboptimal.

In this study, we aim to (1) provide preschool teachers with professional development training in pedagogical knowledge and methods in the physical domain, (2) create a set of teaching materials (in printed and digital formats) that could be incorporated into the overall and preschool-based curricula, (3) create a parental toolkit to enhance parents’ knowledge and promote family-school collaboration, and (4) develop a smartphone-based artificial intelligence rating system to help preschool teachers assess children’s motor skills and provide appropriate instructional feedback. We expect these intervention contents will enhance knowledge, confidence, motivation, and activity behaviours (i.e., physical literacy) of all key stakeholders.

To evaluate the effectiveness of these intervention components, a two-armed clustered randomised controlled trial will be conducted. Changes in the primary outcomes of physical activity behaviours and FMS of preschool children randomly allocated to an experimental group or a wait-list control group will be compared. We hypothesise that children in the experimental group will demonstrate larger improvements in the measured outcomes after receiving the intervention, while the effects can be sustained. Process evaluation, using focus group interviews with teachers and parents, will be conducted regularly to enhance intervention contents.

Project No.: R4032-21
Project Title: Developing probiotics into high-efficacy and safe prophylactics and adjuvant therapies for colorectal cancer
Project Coordinator: Prof Jun YU (CUHK)

Abstract

Colorectal cancer (CRC) is the third most common cancer in the world and the commonest cancer in Hong Kong. To reduce CRC burden, it is important to identify agents to prevent progression at early cancer stage or reduce recurrence following disease remission. Development of agents that can enhance efficacies of chemotherapy which are currently used for treating patients with CRC is also highly warranted. Our project team has done extensive works on gut microbiota in CRC and their translational applications. In this translational project, we will (1) evaluate the anti-cancer activities of potentially beneficial probiotic species identified from our preliminary studies, conduct experiments in established animal models with molecular and immunological evaluation; as well as identify and validate microbiota findings in clinical samples; and (2) develop bacterial predictive markers and efficacy-enhancing probiotics for chemotherapy in CRC. In both parts, we expect this project to produce important deliverables, including (i) recognition of how these species can protect against CRC or improve chemotherapeutic response, (ii) acquisition of efficacy, toxicity, dosage, formulation and other data required for an Investigational New Drug (IND) application for pre-market manufacture and subsequent clinical trials, and (iii) generation of intellectual property rights that are ready for commercial out-licensing, corporate spin-out or industrial co-development. The ultimate aim of this translational project is to develop probiotics into marketable products to reduce CRC burden.

Project No.: R5006-21
Project Title: Deeper understanding of color matching mechanism for developing high-quality lighting and imaging systems
Project Coordinator: Dr Minchen Tommy WEI (PolyU)

Abstract

Color characterization is critically important to lighting and imaging (e.g., cameras and displays) systems, which have a total global market size of more than US$352 billion in 2020. These systems aim to accurately emit, render, record, and reproduce colors for human beings. Numerical values (e.g., tristimulus values and chromaticity coordinates) are commonly used to specify colors in design, production, and calibration processes, with the same numerical values representing the colors having matched appearance. These values are derived using a set of functions — Color Matching Functions (CMFs), which characterizes the color matching mechanisms of the human visual system. Though the most widely used CMFs (i.e., the CIE 1931 2° CMFs) were developed based on the color matching experiments performed by only 17 Caucasian observers, they worked generally well in the past 70 years. Recent studies have clearly suggested the failure of using this set of functions to characterize the color matching phenomenon, with matched colors having different values and same values resulting in unmatched colors.

More importantly, though color perception is pretty complicated, involving eyes and brains, which are of great interests to color scientists, vision scientists, and neuroscientists, these three groups of scientists generally work individually, carrying out investigations from their own perspectives and only focusing on individual stages. This has limited our understanding about how the human visual system perceives color matches.

In this project, the color scientists, vision scientists, and neuroscientists will work together to comprehensively investigate how the color match is perceived by the human beings. Color matching apparatuses will be designed and built for carrying out color matching experiments. Electroretinogram (ERG), Electroencephalograhy (EEG), and functional magnetic resonance imaging (fMRI), together with psychophysical experiments, will be used to collect data to reveal the underlying color matching mechanism. The developed model and collected database will be used to guide the development of high quality lighting and imaging systems.

Project No.: R5009-21
Project Title: Reliable multiagent collaborative global navigation satellite system positioning for intelligent transportation systems
Project Coordinator: Dr Li-Ta HSU (PolyU)

Abstract

Traffic jams are one of the most frustrating experiences in daily life. Although intelligent transportation systems (ITSs) have helped alleviate traffic jams, the traffic situation remains fraught; thus, there is a need for further ITS development, especially for road transportation frameworks involving buses and minibuses. ITSs apply state-of-the-art sensors, technologies, and algorithms to improve the efficiency of overcrowded road networks. Most public buses are tracked using global navigation satellite systems (GNSSs). Based on the GNSS-determined location, an ITS can derive the estimated time of arrival of buses to improve user experience and increase public transport utilization. Additionally, the GNSS trace data of buses play a role in revealing the location and time of jams. Thus, GNSS data are used for the holistic design of traffic signaling plans, aiming to reduce traffic jams. However, traffic engineers may design ineffective algorithms if they have erroneous data. Thus, the accuracy and integrity of the GNSS data are important considerations.

This project aims to develop a collaborative GNSS positioning system to solve the problem of data inaccuracy and unreliability by effectively leveraging Hong Kong’s existing smart city digital infrastructure. The 3D city model developed by the Lands Department and the Internet of Things (IoT) connectivity of the buses developed by the Department of Transport are keys to developing the proposed collaborative GNSS positioning system. The 3D model can reveal how the GNSS signals are deflected and propagate inside the city, which can help mitigate the unfavorable multipath effects of the signals. The IoT connectivity introduces GNSS big data, which can enable us to develop an artificial intelligence (AI) algorithm to intelligently (that is, mathematically and statistically) improve the accuracy and reliability of the GNSS traces of all buses and minibuses.

In summary, we will develop a collaborative GNSS positioning algorithm that uses data from a centralized cloud server and that is based on 3D city models, AI, and IoT. The proposed system can play an irreplaceable role to provide accurate, reliable, and absolute position tags in the coming digital era.

Project No.: R6003-21
Project Title: Spatiotemporal graph neural networks for trend forecasting: theory and applications for forecasting weather and electric power load
Project Coordinator: Prof Dit-Yan YEUNG (HKUST)

Abstract

Many real-world applications require simultaneous modeling of the spatial and temporal dimensions of a system. For example, spatiotemporal modeling is necessary to model the change of traffic patterns in a highway network or the development of epidemic transmission patterns in a country over a certain period of time. In addition to modeling the behavior of events which have already occurred, spatiotemporal trend forecasting goes a step further by making predictions. Due to the prevalence of spatiotemporal trend forecasting in a wide variety of real-world applications, this project aims to advance both the theoretical and application aspects of the state of the art by conducting high impact research that can lead to real-world effects in the short to medium term.

In this project, we will develop spatiotemporal graph neural networks (GNNs) that use graphs to model the interactions between entities in the spatial domain. In particular, we will explore graph-based extensions to recurrent neural networks and transformer networks that can increase the accuracy and certainty of the models when making longer-term forecasts. Specifically, we will incorporate recent ideas from adversarial learning, an exciting new topic in machine learning, to the design of these models.

Based on the spatiotemporal GNNs proposed, we will develop specific trend forecasting models for weather forecasting applications by collaborating with the Hong Kong Observatory, the government department responsible for providing weather forecasting services to the city. In addition, based on weather forecasts and other types of data relevant to the applications, we will extend the application scope of our trend forecasting models to cover other forecasting tasks in smart city applications. In particular, we will apply them to problems experienced in forecasting for the electric power industry by collaborating with HK Electric, the older of the two power companies in Hong Kong.

For both the weather forecasting and the electric power load forecasting applications, pilot studies will be launched to extensively and comprehensively evaluate and compare our models with existing methods when run in the operational systems of our organizational partners. We expect that findings from the pilot studies will guide subsequent technology roadmaps for the organizational partners, enabling them to incorporate the technologies developed into their operational systems to create a real-world impact.

Project No.: R7003-21
Project Title: Optical and computational technologies to combat microplastics and nanoplastics pollution
Project Coordinator: Prof Edmund Yin Mun LAM (HKU)

Abstract

“Lost at sea: Where is all the plastic?” This 2004 Science article coined the term microplastic (MP), referring to plastic fragments less than 5mm in size, and alerted the world about these largely invisible pollutants in our environment. Nevertheless, only in the past few years did it begin to take hold in public awareness and research publications. A major hindrance has been the difficulty of quantifying the pollution scientifically. Furthermore, experts argue that smaller plastics – called nanoplastic (NP) for those smaller than 1μm – are more pervasive and should cause greater concerns. Unfortunately, even less is known about their origin, pathway, concentration, and impact.

In this project, with an interdisciplinary team consisting of domain and technology experts, we are at the forefront of advancing knowledge of MPs and NPs through developing imaging, spectroscopy, and computational tools. We seek to extract their morphology, dimension, composition, and other physical parameters, overcoming various imaging challenges. We seek to understand the microbial colonization on the plastic surfaces, and delineate its role on plastic breakdown mechanism and as vectors for potential pathogens. We seek to advance spectroscopic analysis, in order to screen for plastic types and exposures in environmental media with a high throughput. We then seek to integrate such physical, biological, and chemical information in a multimodal computational system, facilitating efficient query and toxicity analysis. These scientific studies together will bring us full circle to the ultimate goal of this project: informing us effective approaches to combat MPs and NPs.

Project No.: R7028-21
Project Title: The longer-term impact of parental labour migration: Well-being, indebtedness and family sustainability in Southeast Asia
Project Coordinator: Dr Lucy Porter JORDAN (HKU)

Abstract

The well-being of labour migrants and their families is important to migrants and to policymakers concerned with migrant welfare. This study investigates the financial, physical, and psychological health of migrants’ families in Indonesia and the Philippines, the two major origin countries for more than 350,000 domestic workers in Hong Kong. While evidence to date suggests that the short-term impact of parental migration varies according to which parent migrates and other family circumstances, very little is known about the longer-term costs and benefits for children and other family members who remain in migrant sending countries. This study aims to extend understanding and provide the sound evidence needed for the development of effective policies and practices around labour migration both within origin countries and within host cities such as Hong Kong.

The study builds on and extends an established mixed-methods programme of research on Child Health and Migrant Parents in South-East Asia (CHAMPSEA) to investigate relationships among well-being, migration, and indebtedness over time. It follows Indonesian and Filipino families, already interviewed in 2008 and 2016, as well as collecting new data. It recognises that the current global COVID-19 pandemic has resulted in unplanned repatriations and income reductions, which may have impacted negatively on transnational families.

Collaboration with the UN International Organisation for Migration global and national offices combined with ongoing engagement with different governments and stakeholders, including community-based organizations will ensure the impact of the research is wide-reaching for diverse beneficiaries. The findings will not only be disseminated among families of migrants ‘left-behind’ and NGOs in sending countries but also to policymakers and service providers for migrant domestic workers in Hong Kong.

Project No.: R7031-21
Project Title: Pathways to harm reduction drug policy in Hong Kong and East & Southeast Asia: principles, process and practices
Project Coordinator: Prof Karen Ann LAIDLER (HKU)

Abstract

Over the past 25 years, Hong Kong has witnessed a dramatic change in illicit drug use, from the predominant use of heroin to a broader range of psychoactive drugs. These shifts are similar to elsewhere in the world, including East and Southeast Asia. At the same time, policies and rehabilitation around drug use locally and regionally are based on the abstinence principle – that one must cease all use. Although this may be an end goal, drug policy must also respond to the risks and needs of people who use drugs. This project seeks to develop and implement research informed policies and program strategies that are based on the principles of harm reduction and wellbeing.

This three-year project builds on the team’s impact-oriented research and includes two components. Component one focuses on Hong Kong and seeks to develop and implement research informed policy and program strategies for frontline workers. The team will first develop a baseline of drug user trends and associated problems based on interviews with users and frontline workers. This information will form the basis for developing and delivering a harm reduction training program for frontline workers. The team will revisit users and frontline workers to assess whether and how practices have changed having had the training and knowledge gained about harm reduction, and to revise the training package for others working in the sector.

Component two is regional in focus and builds on our harm reduction and drug policy trainings over the past seven years. From these trainings, we have established a network of policymakers and practitioners in East and Southeast Asia who have the knowledge and experience of working on drug policy issues and are well situated in working collectively. This component’s aim is to develop a regional research impact hub on drug policy. We will first host professional training events, drawing on our established networks to reach out to others in the region. These events will focus on key priority issues in regional drug policy and assess what is known (or unknown) and begin capacity building to promote change. These events will then form the basis for a research innovation hub with members collaborating to develop a research agenda, collect data from their respective countries, and collectively establish an overall regional profile and suitable action plans.

Project No.: R7035-21
Project Title: Trustworthy quantum gadgets for secure online communication
Project Coordinator: Prof Giulio CHIRIBELLA (HKU)

Abstract

Quantum cryptography promises a revolution in our communication technologies, offering enhanced security based on the fundamental laws of Nature. This level of security is essential to protect today’s sensitive information, such as census data, medical data, and industrial plans, from being decrypted in the future when more powerful computational resources will become available. In the short term, quantum cryptography can already benefit a vast range of applications, such as e-commerce, secure internet browsing, online messaging, online games, and online polls. Quantum cryptosystems have already been successfully employed to secure bank transactions, ballots, and, more recently, phone calls and video-conferences. If broadly implemented, they could benefit society at large, bringing unprecedented security and opening up a new market of quantum-enhanced products.

Still, the full potential of quantum security has not been realized yet. Commercial implementations are secure, but only as long as the devices behave according to their specifications. A new generation of quantum protocols, called device-independent (DI) and semi-device-independent (SDI) guarantee security even in the extreme scenario where the devices have been manufactured by the hackers themselves. However, DI/SDI protocols are still far from being practical. The goal of this project is to circumvent the roadblocks to DI/SDI cryptography, and to bring it closer to businesses and ordinary users. To achieve this goal, we have formed a world-class team including leading experts from the International Centre for the Theory of Quantum Technologies (Gdansk), Stockholm University, and Université Libre de Bruxelles. Building on a unique combination of expertise covering both theory and experiment, we will develop secure protocols for key distribution and random number generation that can be implemented on devices of moderate size while tolerating noise, imperfect initial seeds, and defects in the measurement devices.

This project will build up a local expertise in Hong Kong in the field of quantum cryptography, and is likely to spawn new start-up ventures in the near future. In the longer term, the development of quantum-enhanced communication systems will contribute to create a secure digital future and to strengthen Hong Kong’s leadership in the technological and financial sectors.

Project No.: R7036-21
Project Title: Therapeutic application of induced pluripotent stem cells derived mesenchymal stromal cells for treatment of heart failure
Project Coordinator: Prof Hung Fat TSE (HKU)

Abstract

Human mesenchymal stromal cells (MSC) are one of the oldest progenitor cells that have been introduced for “regenerative medicine”, as well as in modulation of a range of immune and inflammatory-mediated conditions. Nonetheless, adult somatic tissue-derived MSC, such as bone marrow and adipose tissue are susceptible to senescence in culture, have great batch-to-batch variability in quality (gene expression, differentiation and proliferation), and repeated often costly recruitment and qualification of donors that hinder their clinical application. To overcome these limitations, our research team in Hong Kong University (HKU) was one of the first in the world to successfully derive different MSC lines from human pluripotent stem cells, including human induced pluripotent stem cells (hiPSC) for pre-clinical investigation in different cardiovascular diseases. In this project, we propose to generate clinical grade hiPSC-MSC under Good Manufacturing Practice in HKU for preclinical and clinical investigation for treatment of heart failure.