Project Reference No.: UGC/FDS24/E09/23
Project Title: An Innovative Hybrid Nanofluid Spray Cooling Based Thermal Management System for Efficient Cooling of Electric Vehicle High Power Electronics
Principal Investigator: Dr ASIM Muhammad (PolyU SPEED)
Abstract
In recent years, as electric vehicles gained popularity and countries introduced policies for clean energy and clean environment, thermal management of electric vehicle (EV) high power electronics became a research focus due to high power density, chip miniaturization, dense packaging and increased vehicle speeds and mileages. Heat dissipation in high power electronics of current electric vehicles (EVs) can reach up to 500 W/cm2, and it may exceed 1000 W/cm2 in future EVs. Such a high heat flux may not be removed by even efficient cooling technologies (for instance, spray cooling) due to the limited heat removal capacity of existing thermal fluids, such as water and dielectric fluids. To address this issue, the spray cooling potential of the next generation thermal fluid, called the hybrid nanofluid, is proposed in this research.
Research suggests that the hybrid nanofluid (dispersion of two different nanoparticle types in the base fluid) possesses synergistic thermal properties, where its thermal conductivity is higher than that of the base fluid and single-particle nanofluids. In this proposed research, the hybrid nanofluid will be used in spray cooling applications that may address heat dissipation challenges in high power electronics of current and future electric vehicles. Each spray droplet will comprise hybrid nanoparticles that can extract a large amount of heat due to its synergistic thermal behaviour. The preliminary results of this research show that the hybrid nanofluid spray cooling can remove the heat dissipation flux of EV high power electronics modules while keeping them within safe temperature levels. On the other hand, spray cooling using traditional fluids (such as water and dielectric fluids) cannot remove the heat dissipation flux of EV high power electronics resulting in temperatures higher than the device failure temperature. This is because existing thermal fluids exhibit low heat transfer coefficients due to their poor thermophysical properties. This project will investigate both one-sided and two-sided cooling approaches for the hybrid nanofluid spray cooling of EV high power electronics. Furthermore, the package thermal resistance of EV high power electronics will be reduced using a thickness optimization approach for different layers of EV power electronics module. The preliminary results of this proposed research indicate that a ceramic material used in a power electronics module contributes a great deal of total package thermal resistance. In this proposed research, the thickness of the ceramic material will be optimized such that it can withstand high voltages while adding a minimum thermal resistance to power electronics packaging. Also, the temperature effects will be considered in the thickness optimization study, as power electronics modules operate at high temperatures that may affect the dielectric strength of a ceramic material used in it. Moreover, new materials with even better dielectric and thermal properties than currently used ceramic materials (such as alumina and aluminium nitride) will also be investigated in this proposed research. Furthermore, the cleaning and re-suspension mechanisms for hybrid nanofluid spray residue will be investigated in this project. This research will reveal the spray cooling potential of the hybrid nanofluid for effective thermal management of high power electronics of both current and future electric vehicles. Although hybrid nanofluids proposed in this research may pose potential threat to aquatic and terrestrial environments, it will be used in the closed loop confined space once integrated in EVs and will not be exposed to the environment.
Project Reference No.: UGC/FDS16/H29/23
Project Title: The role of active and passive attentional capture in the attentional boost effect
Principal Investigator: Dr AU Ricky Kwok-cheong (HKMU)
Abstract
The human senses are receiving massive amount of information all the time from the surrounding environment. The processing capacity of the mind has its limits. The attention system is a mechanism of the mind to select and focus on processing goal-relevant information. Attention can be understood as a mechanism that the mind distributes mental resources to process different concurrent streams of information (e.g., talking to someone while driving a car). The conventional point of view in cognitive psychology would suggest that such scenario of “divided attention” would lead to impaired performance in both of the tasks involved, due to limited amount of mental resources being divided to process the two tasks simultaneously. This “interference hypothesis” has been in place in the psychology field for long time, for explaining various experimental findings of impaired performance during dual-tasks or multitasking.
While it is reasonable for one to expect such an interference effect when people are performing dual-tasks, recent research has discovered the possibility that engaging in a dual-task sometimes can result in facilitation in task performance. In particular, the “attentional boost effect” (ABE) initially reported by Swallow and Jiang (2010) has demonstrated that when people are memorising materials shown to them and performing a concurrent target-detection task (e.g., pressing a designated key when seeing a target square) on the same computer screen, they would remember the materials better, as indicated by their better performance in a subsequent memory test. Such an “enhancement” effect in dual-task is surprising, as one would expect dividing attention should normally impair mental processing of information. This ABE has been demonstrated with memorisation of various verbal and pictorial materials, and across the senses of vision and hearing. Theorists have proposed that detection of target in the dual-task leads to enhancement of mental processing for a very brief duration (of about 0.2 seconds), leading to enhanced encoding of the concurrently received information into the memory system.
Most previous studies of ABE have focused on properties of the to-be-memorised materials, while very few have examined the attention-capturing target itself. This is undoubtedly an important research gap to be filled, since any phenomenon involving capture of visual attention ultimately depends on the attention-capturing object. Regarding this topic of ABE, the PI has conducted a study previously (Au & Cheung, 2020) and found that the effect is dependent on the salience of the attention-capturing object (including physical appearance and frequency of occurrence), but independent of the spatial position of the object to the to-be-memorised materials on the screen. To extend the investigation along this line of important yet under-researched topic, the proposed project here aims to systematically examine the properties of the attention-capturing target in the ABE—specifically, the active and passive attentional capture.
The two proposed behavioural experiments will involve human participants memorising English words presented on the computer screen, while detecting a target object by pressing computer keys. After this memorisation part, the experiment will proceed to the recognition part in which the participants will indicate whether they have seen each presented word during the memorisation part. In Experiment 1, each target detection will be assigned with a reward score, and participants will gain the score if they correctly detect the target. The aim is to test whether a higher value attached to the target detection would lead to stronger attentional capture (active engagement due to higher motivation), which in turn leads to better memorisation (stronger ABE). In Experiment 2, before the memorisation part, participants will see a series of images which all suggest a colour that is consistent (or inconsistent) to the colour of the target (e.g., red target: tomato, red light, fire hydrant). This would lead to “priming” (a kind of implicit learning) towards that colour in the participants. When they see a target in a consistent colour in the subsequent target-detection task, their attention would be passively drawn to the target to a stronger extent. By comparing memorisation of materials paired with a target in a colour consistent vs. inconsistent with the implicitly-learned colour, we can determine whether passive engagement of attention plays a significant role in the ABE.
Project Reference No.: UGC/FDS16/E04/23
Project Title: Behaviour and design of high performance steel bolted connections at ambient and high temperatures
Principal Investigator: Dr CAI Yancheng (HKMU)
Abstract
The use of high performance steel (HPS) is consistent with the public policy of the Government of the Hong Kong Special Administrative Region in terms of construction waste reduction, energy efficiency and sustainable development. It has been advocated that HPS could be used for steel modular units in Modular Integrated Construction (MiC) in Hong Kong, e.g., using high strength steel (HSS) for larger load-carrying capacities (Shan et al. 2019). HPS with superior performance in characteristics such as high strength, cold formability, corrosion resistance and ductility, has been developed and produced in China and elsewhere. These superior characteristics are favoured by engineers in the design of steel structures and steel-concrete composite structures, as well as steel modular units for MiC projects in coastal areas like Hong Kong. The recently developed HPS, namely the cold-formed HSS with a nominal 0.2% proof stress (fn,0.2) of 900 MPa and the relatively new stainless steel called lean duplex stainless steel (LDSS) with nominal fn,0.2 of 530 MPa, are available in the market. However, structural design using HPS materials, such as cold-formed HSS and LDSS, at ambient and high temperatures (in fire) are not covered by the existing international design specifications, except for the design of LDSS at ambient temperature in Eurocode (EC 3-1.4, 2015). This situation restricts the use of such high performance materials in construction.
Bolted connections are one of the most commonly used connection types in steel structures. The structural behaviour of bolted connections plays one of the most critical roles in the safety and stability of steel structures at ambient and high temperatures. While there is a great interest and need in adopting high performance steel materials (i.e. cold-formed HSS and LDSS in this project) in the construction industry, there is no existing design rule for HSS and LDSS bolted connections at ambient and high temperatures. It is proposed in this research project to conduct experimental and numerical investigations on cold-formed HSS and LDSS bolted connections at ambient and high temperatures. The investigations will be firstly conducted at the component level and then at the system level, such as HPS frame tests. The main objectives of this project are to generate experimental and numerical data on the resistances, deflections and failure modes of HSS and LDSS bolted connections at ambient and high temperatures, and to derive new design rules for different failure modes that facilitate the adoption of HSS and LDSS in the construction industry. These newly proposed design rules can be adopted by international structural steel design codes and used by practising engineers.
Project Reference No.: UGC/FDS51/H02/23
Project Title: A Study of the Corpus included in Samuel Williams’ A Tonic Dictionary of the Chinese Language in the Canton Dialect
Principal Investigator: Dr CHAN Abraham Yee-shun (UOWCHK)
Abstract
Printed at the Office of the Chinese Repository in 1856, A Tonic Dictionary of the Chinese Language in the Canton Dialect by the American missionary, official, linguist and sinologist Samuel Wells Williams (1812–1884) was the first dictionary written specifically for Cantonese. Its value lies not only in its age, but also in its preservation of early 19th-century Cantonese features that were precipitately lost by the end of the century. While the general characteristic of the dictionary has been described by a few scholars, as of today there is still no dedicated study on the Cantonese pronunciation, vocabulary and expressions described therein, no doubt due to the complexity involved—the phonological system it depicts is distinct from that found in S.L. Wong’s A Chinese Syllabary Pronounced According to the Dialect of Canton published just 85 years later, and similarly many 19th-century expressions have fallen out of use. The situation is worsened by Williams’ decision to leave out Chinese characters in his examples due to the unavailability of small Chinese fonts in his office. Even Samuel Cheung’s comprehensive database of early Cantonese texts does not include this important work.
The present study attempts to fill in this significant gap by building a database of this treasured dictionary. The database will form the backbone of an online enquiry system from which users can easily search for particular words and phrases. But in order to achieve that, we first have to fill in the most essential missing piece of the puzzle—the Chinese characters left out in all the examples. Doing so requires a thorough understanding of 19th-century Cantonese phonology, and the ability to accurately infer a Cantonese term from its English explanation offered in the dictionary. The database will allow searches based on English, Chinese characters, and/or modern Cantonese romanizations such as Jyutping.
The present project will benefit all Cantonese speakers who have an interest in the history and development of this important Chinese dialect, which for an extended period of time served as the main vehicle of East-West dialogue. A comprehensive study of the corpus contained therein will shed light on the social, economic, and political exchange between China and the West amid the Opium Wars.
Project Reference No.: UGC/FDS14/H03/23
Project Title: News Consumption in Hong Kong: News Avoidance, Movement Abeyance, and Critical Events
Principal Investigator: Dr CHAN Chi-kit (HSUHK)
Abstract
News avoidance is drawing both scholarly and public attention in recent years. Scholars want to understand why some people intentionally reject news information of public affairs. News avoidance has substantial implication on political communication and assessment of public opinion. Whether and to what extent news is no longer a preferred source of social information to people thus are significant academic and policy questions. This study aims to examine 1) are there any demographic, psychographic or media usage patterns behind the practice of news avoidance? 2) would news avoidance fundamentally challenge our current understanding of public opinion and citizenship? 3) under the context of Hong Kong, would news avoidance relate to the aftermaths of critical events, such as the 2019 protests and COVID crisis?
Project Reference No.: UGC/FDS51/H(B)01/23
Project Title: Housing preferences of young adults in Hong Kong: How should shared housing be designed and priced?
Principal Investigator: Prof CHAN Hing-lin (UOWCHK)
Abstract
Hong Kong’s housing prices are among the highest in the world, and few of its residents can afford to buy private property. Therefore, many low-income families rely on housing aid to secure accommodation. This problem is particularly acute for young people, who have little work experience and typically work in low-paying jobs. Many surveys have found that young people are the most unhappy demographic group in Hong Kong. Although the government has repeatedly sought to expedite the construction of public housing, the number of completed units lags far behind the target. Clearly, Hong Kong’s housing problem cannot be solved by traditional public housing plans. The government must instead develop innovative, multi-pronged approaches to address this issue.
We argue that the emergence of shared housing is a viable policy option for young people in Hong Kong under the Youth Hostel Scheme (YHS) introduced by the Hong Kong government. Shared housing has several advantages over other options. For example, unlike traditional public housing, in which tenants are entitled to occupy a flat for their entire lifetime, tenants of shared housing move out when they reach a certain age. Therefore, the government subsidy provided for the YHS is considerably smaller than that provided under the traditional approach. The YHS aims to provide young people with housing at a level of rent that enables them to save and thus buy private property in the future. However, for the scheme to succeed, the rent must be sufficiently low, which can be achieved if the shared housing approach allows the designers to maximise the use of common areas.
Against this background, our proposed study will explore the housing preferences of young adults in Hong Kong, with a special focus on shared housing. Such a study is important for three reasons. First, various housing aid options have been proposed in recent years, including public housing, transitional housing, a cash allowance scheme and shared housing. These options differ in housing features, rental costs, queues, permitted occupation periods and income ceilings. We will thus determine how young adults rank these options and then use average marginal component effects to measure the impacts of individual factors on their choices. The results will reveal how people rank existing shared housing options, the market demand for these options and whether the options are perceived as close substitutes.
Second, any new shared housing units should be built in accordance with the demands of the young people who will live there. Therefore, the importance of various shared housing attributes should be evaluated. We will thus use the analytic hierarchy process (AHP) to quantify young people’s rankings of housing attributes. Unlike conventional survey methods, the AHP will enable us to assign a numerical value to each housing attribute that represents young people’s valuation of that attribute.
Third, our proposed study will ascertain young people’s willingness to pay (WTP) for this type of housing, an important factor for understanding their demand preference. Shared housing remains in the development stage, and pricing and attribute data are not available. Furthermore, shared housing is provided solely by non-governmental organisations, which means that there are no market signals from which to derive a price. Fortunately, WTP surveys are widely used to estimate the value of non-market goods. Our results will provide policymakers with valuable information on the rents they can collect, the potential costs of such projects and, therefore, the required subsidy. This information will enable the government to more accurately evaluate the feasibility of a shared housing scheme in Hong Kong.
Project Reference No.: UGC/FDS15/H19/23
Project Title: Hong Kong Children’s Literature: City and Visual-Textual Narratives
Principal Investigator: Dr CHAN Michelle Chi-ying (Shue Yan)
Abstract
Hong Kong children’s literature has been considerably sidelined in academic scholarship. When it is mentioned, it is mostly a brief entry in historical overviews up to the 1990s. It may seem that the academia value the genre when there are two anthologies in the series of Hong Kong Literature dedicated to children’s literature, published in 2014 and 2021, respectively. Still, the anthologies have only listed the works of a number of writers without in-depth discussion. Most academic analyses (Dong 1995; Zheng 1996; Zhou 1996; Liu 1997) see the genre as having petered out in the early 1990s. There is clearly a void of scholarship on the subject of the writers, publishers, intents and types of published materials published in the 1990s. Nonetheless, the rise of children’s picture books in the last decade (Wu 2019) has revived the genre significantly. Not only are the chain bookstores willing to separate an area for international and local children’s books now, but there is an increasing number of independent bookstores and publishers dedicating their business to picture books.
The proposed research aims to compensate for the dearth of academic studies of Hong Kong children's literature by addressing the following issues: 1) the use of visual-textual narrative in Hong Kong children's literature; 2) the rise of picture books; and 3) the manifestation of Hong Kong as a city in the selected works. The focus of this project will be on the inter-connectedness between the city culture of Hong Kong and the local visual-textual narrative. It has been raised by a number of critics (Lo 2009; Yesi 2012; Wong 2020) that Hong Kong literature is always related to its city culture. Hong Kong literature embraces its colonial history and politics, geopolitical role as an international city in Asia, and indispensable city culture. This raises the issue of how children’s literature will align with these characteristics of Hong Kong literary texts. Additionally, in the study of Hong Kong literature, the graphic narrative does not enjoy equal status with textual narrative. Even if the graphic narrative has been examined as a communication mode of picture books (Nodelman 1988; Nikolajeva and Scott 2004) and comic books (McCloud 1993; Kress and Van Leeuwen, 1996; Groensteen and Miller 2013), it is being underestimated by the local academia (Dong 1995; Liu 1997; A Nung 1997; Ho 1997) who do not see the contribution of visual narrative to Hong Kong literature. In order to provide a more comprehensive and expand the scope of reading literary texts, this project will include the study of the visual narration and its collaboration with textual narrative.
Contextualized in light of the contemporary urban context, this proposed project will examine the varying ways in which children’s literature is engaged with the landscape of Hong Kong as a city space, both in terms of realism and as an idea, such as the fantasy of the city as a space of interconnectedness and transverses, evolving dynamics, and culturally interlocking network. It will examine the visual-textual representation of Hong Kong in Children’s Paradise Magazine (1953-1995), an important children’s magazine series that manifests the evolution of visual-textual collaboration, and how the current trend of picture books has opened a new platform for Hong Kong children’s literature and, thence, shed new light on the understanding of the city in a brand new visual-textual narrative. The proposed study will investigate the aforementioned issues with qualitative content analyses of selected primary texts and a quantitative examination through a corpus machine, Google Vision API, so as to take an overview of the image of city in both textual and visual narrative. The investigation will be supplemented by the interviews of publishers, writers, and illustrators so to facilitate a more comprehensive understanding of the current trend.
Project Reference No.: UGC/FDS16/H10/23
Project Title: Fable of the Bees in Our Time: An Inquiry to Recover Assets and Values of Waste Pickers in Hong Kong
Principal Investigator: Dr CHAN Wai-yin (HKMU)
Abstract
Bees have been depicted in classical works from Mandeville to James Meade and Steven Cheung which gather nectar for survival (private interest) and unwittingly generate positive externalities (public benefits) through pollination. Waste pickers are like bees. In the affluent but economically polarized Hong Kong, waste picking is a common informal economic activity. It is mainly performed by older people who recover recyclable materials every day for extra income. While they are highly motivated by economic incentives, their work generates environmental and economic benefits. Amid the strong initiative of the HKSAR government to develop the circular economy for the city’s liveability, the work practices, assets, and contributions of waste pickers as specialized waste management service providers are invisible from public policies, academic research and public discussion. On the one hand, penalizing and banning policies against waste picking are dominant. On the other hand, the handful of local research and public discourses commonly see waste pickers as deprived and passive needy with a romantic view of their self-reliance.
This proposed research aims to examine the work practices, assets and values of waste pickers, and the values of informal waste picking activity to the society as well as waste pickers themselves. The study will be situated in the academic fields of waste studies and informal economy. Waste is not a fixed object with static meaning, but created, managed and circulated in intertwined economic, political and socio-cultural systems. As a frontier of the waste management system, waste pickers divert discarded objects from landfills so that these objects can undergo the process of value transformation to become valuable useable resources and tradeable commodities. They therefore work as enablers who initiate the social process of “unwasting”. They are nevertheless excluded from the formal waste management system and their work is highly devalued. Adopting asset-based sustainable livelihood approach, this study serves to uncover five assets of waste pickers, namely human, social, physical, financial, and public assets, and further identify vulnerabilities and structural factors that constraints their assets and thus their livelihood. It will employ multiple qualitative methods to generate a thick description and analysis of livelihood strategies of waste pickers in Hong Kong and the values generated when they work to transform the values of waste.
This research will be a timely response to the HKSAR government’s initiatives to build Hong Kong as a socially cohesive and livable city. It not only addresses the local concern on wealth disparities, old-age poverty, and social inclusion, but also the global concern on waste management and the call to add social and political dimensions to the currently ecological-driven model of circular economy. It also serves to fill in the longstanding research gap on waste pickers and waste management in Hong Kong with an asset-based approach and add the perspective of Hong Kong to the scholarly discussion on waste pickers and informal economy which is now predominately conducted in the Global South. Research findings will facilitate policymakers and service providers to have a fresh look at waste management and social inclusion, and to formulate new initiatives to facilitate waste picking activity to generate maximum social benefits for equality and environmental sustainability in Hong Kong.
Project Reference No.: UGC/FDS24/H22/23
Project Title: The Echoes of Air Raids: An Exploration of Wartime Soundscapes through Literature and Print Media
Principal Investigator: Dr CHEN Hazel Shu (PolyU SPEED)
Abstract
This project aims to examine the representation of wartime soundscapes in Republican China from 1937 to 1945 in relation to print media. It will explore the significance of wartime soundscapes as cultural memories and examine the representation and documentation of these sounds and the concept of modern acoustic experience through literature and other forms of textual evidence. The study will focus on newspapers and literary periodicals published in Chongqing and Kunming, to gain insight into the wartime soundscape and analyze the historical and literary expression of acoustic modernity during emergency and contingency.
The representation of soundscapes and modern aural experience has not been widely studied in the field of modern Chinese literature and culture. This comprehensive study of wartime soundscapes will encompass various aspects, including the noises and sounds of air raids, such as sirens, machine guns and bomb explosions in dangerous environments, and the literary techniques used to depict the wartime soundscape in news articles, fiction, and autobiographies.
The researcher will conduct extensive archival research, including online archives such as the Chinese Periodical Full-text Database and microfilm collections from university libraries, and onsite research in Chongqing and Kunming. By examining both the sonics and the mediated sonic experience, the project aims to reconstruct the historical soundscapes shaped by modern warfare and media technology and to use literature as a primary site to explore the emotional and aesthetic impact of sound.
Project Reference No.: UGC/FDS14/P03/23
Project Title: InsurTech: Risk Classification and Premium Calibration with Data Analytics
Principal Investigator: Dr CHEN Yongzhao (HSUHK)
Abstract
In the era of big data, with the enrichment of channels of data collection and the aid of high-tech device that gathers instantaneous data with high precision, the number of feature variables in the data collected and the size of the datasets are both rapidly increasing. As a result, data analytics has gained much popularity over the past decade, and its widespread applications range from the insurance and finance sectors to medical science and even social sciences.
Taking InsurTech for example, the traditional premium calculation can be further modernized by incorporating data analytics, e.g. with the finer tracking of the change of market interest rate against time, and a more elaborated record of the lapse behavior of the policyholders. As another powerful tool in premium calibration, the traditional credibility theory only focuses on the individual level, yet with the incorporation of data analytics, the company may adopt some subjective view on each of the policyholders based on the data collected, which may lead to different premiums collected for them.
On the other hand, risk classification is not only important in InsurTech and FinTech, but also inevitable in other fields including advertising, logistics and education. Indeed, the increasing number of feature variables naturally provides more information, but the underlying dependence structure is more complicated than ever; a natural problem is how we can extract the most useful information from all these feature variables and reveal the hidden dependence relationship among feature variables as much as possible.
In this proposed project, we aim to revisit these two important issues in InsurTech, i.e., premium calibration and risk classification, from the modern perspective of data analytics. For premium calculation of a specific type of insurance contract, called Guaranteed Minimum Benefit (GMxB), we shall adopt a more realistic model formulation to incorporate the developments in the data analytics context as mentioned above, based on which we further develop a numerical scheme that achieves a similar level of accuracy but with significantly reduced running time compared with the commonly adopted Monte Carlo method by insurance companies. We then proceed to risk classification, where we shall devise a mechanism to handle all feature variables in a unified manner, and all will be converted to take discrete numerical values, allowing us to reveal their dependence structures as much as possible; this will further facilitate the implementation of our proposed new Bayes classifier that also takes the dependence structures into consideration, and preliminary numerical studies has shown the competitiveness and robustness of our classifier against other commonly used machine learning algorithms. Finally, we return to premium calibration via credibility theory, and investigate the inclusion of the corporate-level viewpoint from the modelling perspective by augmenting an “aggregate” loss function to the objective function, which originally contains individual loss functions only. In this case, a nice closed-form solution for the premium, such as the one in the celebrated Bühlmann model, becomes infeasible due to the increased complexity of the objective function, hence we seek for an implicit solution via fixed-point expressions instead, which is practically useful and efficient, especially for the purpose of computer simulation.
Our overall research goal will be accomplished by addressing various topics, including but not limited to ordinary differential equations, numerical methods, machine learning and credibility theory. Our anticipated results should (1) facilitate a more efficient GMxB premium calibration for insurance companies, so as to save their time cost in their daily operation; (2) provide a competitive and efficient classifier based on the simple yet powerful notion of comonotonicity, together with a well-developed feature engineering mechanism that is also compatible for other machine learners; and (3) indicate an implicit approach to calculate the premium of each policyholder in the presence of corporate-level considerations, with valid economical interpretations on the formulas obtained, and set up a numerical scheme for numerical computation of the premiums.
Project Reference No.: UGC/FDS24/H10/23
Project Title: A Multimodal Discourse Analysis of Dynamic Handwritten Annotations as Visual Aids in Live Lecture Recording
Principal Investigator: Dr CHEUNG Eric Lok-ming (PolyU SPEED)
Abstract
Technological advancement makes educational technologies (EdTech) easily accessible. The Covid-19 pandemic has also significantly changed the educational landscape – synchronous online teaching and learning has been made possible with teleconferencing applications with screen- sharing functions. Aside from using visual aids such as PowerPoint presentations and video contents, it is also common for teachers to annotate the screen with handwriting in real time. Research studies have demonstrated the benefits of on-screen handwritten annotations in terms of comprehension, motivation and engagement; however, it is less known as to how such annotations are coherently linked with other visual contents on the screen, as well as what kind of annotations are effective in engaging students and facilitating learning.
Therefore, this 24-month project aims to examine the meaning-making practices of handwritten annotation through collecting and analysing live lecture recordings from four main disciplines (Business, Engineering, Language and Psychology) in a tertiary institution through multimodal discourse analysis informed by Systemic Functional Linguistics. The project also aims to investigate the learners’ perceptions and experience of using learning materials with dynamic on-screen handwritten annotations, and the observed instructors’ practices of dynamic on-screen handwritten annotations in live lectures through class observations, surveys and interviews.
Through the three studies, the proposed project seeks to make explicit the logical relations between handwritten annotations and static visual aids in live lecture screens, and to identify what meaning- making strategies can increase students’ engagement and reinforce their memory for the learnt content. Aside from informing teaching and developing teacher training materials, the proposed project also aims to advance the research field of multimodality and EdTech by filling the gap of the research literature on the synergies among different visual aids in live lecture recordings.
Project Reference No.: UGC/FDS17/M02/23
Project Title: Characterization of subtypes of Mild Cognitive Impairment with artificial intelligence: Multiomics approach
Principal Investigator: Dr CHEUNG Eva Yi-wah (TWC)
Abstract
Background:
Magnetic resonance imaging (MRI) sequences have been used in the diagnosis of Alzheimer’s disease (AD), mild cognitive impairment (MCI) and other forms of dementia. Several studies using structural MRI have reported that patients with AD have significant medial temporal atrophy compared with healthy controls. Resting-state functional MRI (rs-fMRI) has also been used to identify regional deficits in functional connectivity among patients with dementia. Positron emission tomography–computed tomography (PET/CT) images provide information on amyloid-beta plaque locations, which have been shown to be associated with AD progression. Such imaging techniques can significantly improve dementia diagnosis. However, the prodromal stage of dementia presents as mild cognitive impairment with findings of minimal structural changes and low amyloid load on imaging. Multiple imaging modalities are required to obtain a conclusive diagnosis. A comprehensive algorithm that incorporates various imaging modalities may improve the accuracy of imaging-based characterization of mild cognitive impairment, which could allow timely intervention (e.g., support and medication) and thus greatly improve patient outcomes.
Aim and Objectives:
The proposed project will aim to develop an artificial intelligence system to support early diagnosis of diseases presenting with various types of mild cognitive impairment. The first objective will be identifying relevant clinical and imaging features from that characterize subtypes of mild cognitive impairment. Second, extracting image features relevant to MCI from structural MRI, rs-fMRI and 18F-flutemetamol positron emission tomography–computed tomography (PET/CT). Third, building a diagnostic model based on clinical and imaging features using artificial neural network (ANN) modelling. The final objective will be implemented and evaluate the ANN model using existing preliminary and newly acquired data for validation.
Method:
The proposed project will comprise three parts. (1) Structural MRI, rs-fMRI, PET/CT images and clinical information will be downloaded from public databases for extraction of imaging features and analysis. (2) Discriminative clinical and imaging features will be identified for neural network model building. (3) The model and associated clinical and imaging features will be validated using preliminary clinical data collected from a local hospital. (4) Further validation will be done using prospective clinical data. A multiomics neural network model will then be developed to aid the accurate characterization and diagnosis of diseases presenting with mild cognitive impairment.
Expected outcome:
The multiomics artificial neural network model to be developed in the proposed project will be utilized for characterization of various subtypes of mild cognitive impairment. In the future, the identified clinical and imaging features may be used to monitor and predict disease progression and improve treatment planning for patients.
Project Reference No.: UGC/FDS24/H06/23
Project Title: Effectiveness of Early Support to At-risk Students Identified by using AI Prediction
Principal Investigator: Dr CHIU Hon-sun (PolyU SPEED)
Abstract
E-learning has been used in education for many years. Especially in recent years, the utilization of e-learning has come to a new high due to the COVID-19 pandemic. By using e-learning, students are allowed to learn at any time and any place, as long as they have access to the Internet (Crescente and Lee, 2011). Many online systems have been developed to support e-learning. Apart from allowing teachers to release teaching materials to students, it also provides a lot of additional features to assist students’ learning, such as forum discussion, interactive activities, voting and collaboration (Giannakas et al., 2021).
Different from face-to-face classes, all actions taken by students on e-learning systems can be recorded in log files. This is essentially the learning footprint of students, including the number of logins, duration of each login, number of accesses to learning materials, posts in forum discussions, grades of assignments, and a lot more (Bravo-Agapito et al., 2021). By analysing this huge amount of data, the learning behaviour of students can be observed, which is considered highly correlated to their academic performance (Murray et al., 2013).
With the above consideration, we are going to develop an artificial intelligence (AI) prediction model using a deep artificial neural network (ANN) approach. The prediction model aims at identifying the group of students who are at-risk of failure, dropping out or withdrawing from their studies in an early stage. Then necessary support can be provided to assist them in successfully completing their studies. Our objective is to evaluate the effectiveness of early support to at-risk students, in terms of minimising the failure rate, drop-out rate and withdrawal rate. Unlike most research works in the literature that focus on the prediction in a specific course, our design would consider a more generic view of at-risk students in tertiary education. By doing so, our prediction model can be reused in any course and is suitable for any background of students.
In order to achieve our research objective, there are three research questions to address. (1) Do the at-risk students share a set of common characteristics regardless of any specific subject? (2) What is the optimal time to conduct the performance prediction? And (3) Does early support to at-risk students help to minimise the failure rate, drop-out rate and withdrawal rate? To answer these questions, we will carry out three research tasks. In the first task, a data mining strategy will be designed, with the aim of determining the common characteristics of at-risk students regardless of any specific course. In the second task, we will design an early students’ performance prediction model using deep ANN. In the last task, we will evaluate the optimal time to perform the prediction and study the effectiveness of early support to at-risk students.
Project Reference No.: UGC/FDS14/H08/23
Project Title: Arts for well-being: The impact of arts engagement on parental stress and mental well-being
Principal Investigator: Dr CHOY Christine Hiu-ying (HSUHK)
Abstract
The pandemic has imposed stressors on many parents who faced economic burdens, physical burnout, leisure constraints, and inadequate social support. Despite these issues, more community arts resources have been developed for parents to engage in arts and cultural activities with their children to improve well-being. However, it is notable that unequal access to information and communication technology among parents of varying demographical backgrounds may limit the potential benefits.
The project aims to theorise a framework of arts for parental well-being. With more parents engaging in the on-/ and offline mixed modes of social leisure in a post-pandemic context, this project examines the impact of arts engagement on parental stress and mental well-being, considering the role of digital literacy. This study will adopt a mixed-method design that integrates surveys of parents, focus groups with creative practitioners, non-profit organisations and social workers, and participant observation of community arts programmes that promote parental well-being. The opportunities offered to parents and children for well-being will be described, along with the implications of art and mental health for artists, creative practitioners, social services organisations and policymakers.
Project Reference No.: UGC/FDS11/H10/23
Project Title: The building and operation of transnational elder care network: a case study of left-behind elderly in Hong Kong
Principal Investigator: Dr FUNG Ka-yi (Caritas)
Abstract
This project addresses the issue of elder care in Hong Kong. In the past few years, more than two hundred and twenty-two thousand people have migrated to other societies. Many emigrants were adults with young children, leaving behind their elderly parents. The parents were left behind because their adult children did not plan to bring them along or they refused to emigrate. Among these parents, some are staying in Hong Kong without physical and emotional support from their younger generation. Elder care for these left-behind parents provokes discussions in the academia, government, and social service providers. Studies of migration find that migrants would establish a transnational elder care network to look after left-behind parents. However, the existing literature lacks holistic evidence on how different parties deal with the situations they face. First, in cases that there is someone from the family to shoulder the responsibility of a primary caregiver, questions such as how family members negotiated with each other in the caring duties allocation process, how the designated caregiver responded to this new role, why would he or she shoulder these duties are rarely asked. Second, there were very few studies that targeted the left-behind older people who have inadequate or no kin support at home, thus providing limited evidence on what they and their migrant adult children would do, who could support them, and what their coping strategies would be in the face of their new living style and change of caregivers. Third, while giving emotional care to their left-behind parents, how would the migrant adult children abroad handle family crises at home, for example, in the face of parental health decline? What were the outcomes? They might have to work with different people to seek support for their parents and might harbor worry, anxiety, and stress. How had these emotions influence the crisis-solving process, as well as their transnational emotional caregiving to their left-behind parents?
To fill the above gaps, the research team attempts to establish how kin and non-kin ties constitute the transnational elder care network. Using “transnational kinscription” and “economies of recognition” as fundamental concepts underlying the research framework, this study focuses on the negotiation process among migrant children, siblings, other family members and the care recipients in the transnational context, the structure and the operation of the transnational elder care network and network members’ views and recognitions on members’ contributions in the caring process. Data will be collected by in-depth interviews. This project includes two groups of respondents. The first are migrant adult children who have migrated to the UK. Before they left Hong Kong, they were involved in the caring of their parents. The second group are the care recipients (parents of the migrant adult children). In the cases that there are other members in Hong Kong who are involved in the caring of the parents, like siblings of the migrant adult children, or close kin, they will be invited to be interviewed as well.
The findings of this study will add to the empirical evidence in migrant studies, ageing studies, and studies in social network analysis. In terms of contribution to practice, the findings can be a reference for social workers to identify which types of family will be more likely to have very limited family members to look after the left-behind elders. Furthermore, the findings may help the Hong Kong SAR Government and local NGOs design an evidence-based human service delivery model for the left-behind older people.
Project Reference No.: UGC/FDS13/B02/23
Project Title: From Broken to Better than Ever: The Impact of Divorce on Leadership Emergence
Principal Investigator: Dr GU Jingyang (Chu Hai)
Abstract
As society evolves, divorce is becoming increasingly prevalent, with the number of couples divorcing each year in Hong Kong more than tripling over the past three decades (Census and Statistics Department of HKSAR, 2022). Divorce is a significant life event that has a far-reaching psychological impact, often surpassing the effects of career-related events like job termination, retirement, or employment changes (Holmes & Rahe, 1967). However, existing research in organizational behavior seldom addresses this important event. The negative consequences of divorce have been outlined in family and economic domains, such as its interference with employees' pay levels and career mobility (Hutchison & Hutchison, 1979; Raz-Yurovich, 2013). Due to sampling challenges, the in-depth exploration of divorce's impact, particularly concerning underlying mechanisms, remains limited.
Against this research backdrop, our current proposal aims to explore the influence of divorce on leadership emergence through a more nuanced lens. In our pilot study, we analyzed a large-scale national survey database and discovered distinct subclasses within the leadership development trajectory among divorced individuals. This suggests that divorced employees don't respond uniformly to divorce; some may even experience an upswing in leadership following the event. Building upon these preliminary findings, we construct our theoretical model from a cognitive adaptation perspective. We propose that divorce hostility can alter divorcees' goal priorities through cognitive adaptation, being successful and prosocial as primary focuses. Additionally, we put forth a theory-driven boundary condition for this relationship, suggesting that the impact of divorce hostility on goal priority changes will be stronger among divorcees exhibiting a rational defensiveness style. These shifts in priorities, in turn, drive leadership emergence. To test our proposed model, we plan to combine survey and experimental methods. The project aims to offer a fresh theoretical perspective on post-divorce career development and provide insights into guiding divorcees' career adaptation more effectively.
Project Reference No.: UGC/FDS15/H07/23
Project Title: Counselling support for the social-emotional wellbeing of young people in Hong Kong: Developing a model of counsellor effectiveness
Principal Investigator: Dr HARRISON Mark Gregory (Shue Yan)
Abstract
The wellbeing of young people in Hong Kong is poor and has been negatively impacted by the COVID-19 pandemic. Secondary school students are particularly at risk of suffering from poor mental health, impacting their life satisfaction, academic work, and future contribution to society. School counselling is an evidence-based practice which can support the wellbeing of children and adolescents. Local schools typically provide wellbeing support by means of Comprehensive School Guidance Programmes, planned and delivered primarily by teachers and social workers. There are many qualified and experienced counsellors in Hong Kong possessing a counselling degree from UGC-funded universities. However, few of these counsellors are currently contributing effectively to the city’s wellbeing provision for children and adolescents. Hence, there exists an ‘untapped resource’ in Hong Kong, which could be effectively utilised to support the wellbeing of young people. Many factors influence the effectiveness of school counsellors, including their perceived professional identity and roles, factors related to the school climate such as principal support, the adequacy of available resources, and collaborative working practices among staff, and the wider sociocultural setting. The proposed study aims to investigate the factors contributing to the effectiveness of school counsellors working in secondary schools in Hong Kong. The final product of the research is a model of how counsellors’ effectiveness depends on their professional identity and role function, factors which are themselves influenced by school climate factors and the sociocultural context of Hong Kong. Given that so little research has been conducted into school counsellors’ effectiveness in Hong Kong, the study adopts an exploratory qualitative approach, collecting interview data on the perceptions of different stakeholders in local and international schools. The benefit of this approach is in developing a rich and nuanced understanding of the factors supporting and impeding school counsellors’ effectiveness through the triangulation of findings across a diverse range of school settings serving a varied demographic of the city’s young people. The qualitative approach makes possible the proposal of a theoretical model which can be tested quantitatively in subsequent research. Several categories of participant will be recruited from the researchers’ extensive professional networks, namely school counsellors, students, principals, and students. Individual semi-structured interviews will be conducted, and reflexive thematic analysis will be used to develop themes which, in turn, will form the basis of the model of school counsellor effectiveness. The theoretical novelty of this model is in bridging research on professional identity, the influence of school climate and cultural factors on counsellors’ practice, and school counsellor effectiveness. The model may also be applicable to other Confucian-heritage societies, given their similarities to Hong Kong in several sociocultural dimensions. In addition to the theoretical contribution of a testable model of school counsellor effectiveness, the study has several policy and practice impacts at the levels of individual counsellors, schools, and government. A set of recommendations derived from the study’s findings will facilitate the development and implementation of policy at the school and government levels to better utilise school counsellors in Hong Kong.
Project Reference No.: UGC/FDS15/H01/23
Project Title: Cinema and Cinemagoing in Early-twentieth-century Shanghai
Principal Investigator: Dr HE Qiliang (Shue Yan)
Abstract
The present research project investigates the introduction of film in Shanghai and its impact on city governance in the first three decades of the twentieth century. The arrival of cinema as a major pastime in the early twentieth century fascinated a wide spectrum of urban residents and thereby created a new type of crowd—filmgoers—despite the audiences’ diverse racial, cultural, and economic backgrounds. The blending of people of differing social standings in newly constructed movie theaters posed a challenge to the local authorities in Shanghai, who were impelled to respond to the rise of film—the most cutting-edge technology and a novel form of entertainment. In consequence, Shanghai’s colonial authorities—the Shanghai Municipal Council (SMC) and Shanghai Municipal Police (SMP)— enacted new architectural codes and fire prevention rules, tightened up anti-crime and anti-prostitution measures, and, finally, established a censorial system to make films containing obscene and subversive elements inaccessible to the viewing public.
While the existing scholarship on films and film exhibitions in early-twentieth-century China focuses mainly on how the motion picture managed to make inroads into the city dwellers’ everyday life, lent experiences of modernity, and elicited modern sensibilities, scholars have paid scant attention to cinema’s role in reshaping a city. Likewise, students of urban history in modern China usually view the popularization of the motion picture as mere evidence of the triumph of modernism and cosmopolitanism but fall short of understanding it as a key player that refashioned the physical, administrative, legal/political, and cultural aspects of Chinese cities. Thus, this proposed study is inherently interdisciplinary as it stitches together two formerly disparate academic traditions—film and urban studies—to explore the dyadic relationship between cinema and city in the early decades of the twentieth century.
By exploring the legislative and reform efforts made by the local authorities in Shanghai in response to the dominance of the motion picture in the first three decades of the twentieth century, the present research project attempts to demonstrate how cinema and city were mutually constitutive: On the other hand, the prevalence of cinemagoing as a new pastime prompted the political authorities in Shanghai to reformulate their agendas of city administration by devising new urban plans, maintaining public safety, diminishing racial segregation, resolving racial/national conflicts, and achieving political stability. On the other hand, moral anxiety and political expediency caused by pervasive fears for the display of scenes of crimes and revolutions on the silver screen cultivated a preference for movies about family, romantic love, and the destinies of individuals that masked intense racial, national, and class clashes in the external world, resulting in the market success of melodramatic films, particularly those of D. W. Griffith (1875-1948), throughout the 1920s.
Project Reference No.: UGC/FDS14/E03/23
Project Title: Sustainable and Resilient Automated E-Fulfilment Operations in the Era of Industrial 5.0
Principal Investigator: Dr HO George To-sum (HSUHK)
Abstract
The rapidly growing e-commerce sector has significantly transformed customer behaviour worldwide. In the European Union, the gross value of retail sales in April 2020 diminished by 17.9%, whereas sales via e-commerce orders increased by 30% (OECD, 2020). The burgeoning of e-commerce purchasing has highlighted the growing importance of e-commerce logistics. In the era of Industry 4.0, automated e-fulfilment centres are adopting technologies such as artificial intelligence (AI), Internet of things (IoT), and automated guided vehicles (AGVs) to enhance the capability and reliability of industries. In the modern business environment, apart from the use of technology, a more value-driven movement is a concern that drives the revolution of Industry 5.0.
Studies on Industry 5.0 have focused on sustainability and resilience (Ghobakhloo et al., 2022; Azadeh et al., 2019). Moreover, concerns about environmental, social, and governance (ESG) performance are growing across business research and industries around the world, including Hong Kong (Linnenluecke, M. K, 2022; The Standard, 2022). Considering the boost in the efforts to mitigate environmental pollution, China has committed to achieving carbon neutrality before 2060 and curbing carbon dioxide (CO2) emissions before 2030, as China is one of the leading contributors of CO2 emissions–accounting for 28% of global CO2 emissions in 2018. However, the lack of integral consideration of interactions between different operation processes in automated e-fulfilment centres would affect operational efficiency (Azadeh et al., 2019). Despite the growth of technology, which increases the operational efficiency of automated e-fulfilment centres, robustness and security remain important for the industry to cope with a changing environment within the supply chain and other factors that might affect everyday processes (van Geest, M., 2022). To meet the concept of Industry 5.0, automated e-fulfilment centres need to integrate additional facilities to reduce non-renewable energy consumption and develop a resilient operation model. However, the implementation of Industry 5.0 is challenging owing to the absence of a practical framework for Industry 5.0. As such, this research gap in the practical implementation framework is a major barrier to the development of Industry 5.0.
This project aims to design a digital twin (DT) to jointly optimize sustainability and resilience in automated e-fulfilment centres. Specifically, solar power is employed as a source of renewable energy for an automated e-fulfilment centre. The proposed model contributes to infrastructure design recommendations by developing a DT that transforms the actual automated e-fulfilment centres to editable virtual automated e-fulfilment centres. Based on the DT, a simulated study is performed to determine the optimal renewable energy model to achieve self-reliance. With the optimal renewable energy model, operational resilience can be achieved by considering the interactions between different operational processes and renewable energy consumption. From the perspective of automated e-fulfilment centres, the proposed model presents a framework to transition from non-renewable energy to renewable energy. In addition, automated e-fulfilment centres can handle the rapidly changing supply chain environment by accomplishing resilient operations. With the aid of the proposed model, sustainable and resilient automated e-fulfilment centres can be realized, resulting in better economic competitiveness and environmental benefits.
Project Reference No.: UGC/FDS14/E02/23
Project Title: Optimizing Model Freshness for Federated Transfer Learning with Extreme Cases in Autonomous Driving Networks
Principal Investigator: Dr HOU Aileen Yun (HSUHK)
Abstract
This project aims to tackle the training data shortage problem and optimize the freshness of the machine learning models for extreme cases in the C-V2X enabled autonomous driving networks. Autonomous vehicles rely on artificial intelligence (AI), visual computing, radar, monitoring equipment, and positioning system to work together so that the system can automatically and safely operate the automobile without active operation. However, an unsafe ‘decision’ of the autonomous driving system is likely to endanger human life and cause huge economic losses. To maintain the high safety requirement of autonomous driving, the accuracy and freshness of the AI networks used for perception and decision-making play a decisive role. However, decision reinforcement learning requires a lot of interaction with the real world to learn strategies and deal with various situations. This makes it difficult for reinforcement learning to learn scenarios rarely seen in reality, e.g., extreme cases. Therefore, in this project, we aim to build a Federated Transfer Learning Framework at roadside units with extreme cases generated and pre-trained in a Digital Twin with optimization for model freshness. To achieve this goal, we will (1) develop a digital twin for extreme case generation in autonomous driving; (2) build a federated transfer learning framework for autonomous driving to improve road safety using the extreme cases emulated in the digital twin; and (3) optimize the federated transfer learning framework in C-V2X networks depending on the real-time network status to dynamically decide how broad the model should be federated and how mature the model should be trained at edge nodes towards the optimal freshness of transferred models.
The research outcomes from this project will lay the foundation for the future optimizations of collaborative deep learning in vehicular edge networks toward the new metric “model freshness”. Given the rapid evolvement in high-level autonomous driving, the research outcomes are expected to open a new dimension in providing all-round cooperative sensing and decision planning to assist vehicles safely navigate through complex and extreme scenarios while ensuring safety in all aspects.
Project Reference No.: UGC/FDS16/E18/23
Project Title: Development of Novel Quaternion Signal Processing and Feature Extraction Methods for the Monitoring and Early Detection of Mental Disorders in a Head-based Mobile Health System
Principal Investigator: Dr HUNG Kevin King-fai (HKMU)
Abstract
The prevalence of mental disorders and dementia causes a heavy burden and unprecedented challenge for healthcare and social systems globally. Early detection of mental abnormalities is the key to solving this problem. It enables early intervention, prevents illness deterioration, and reduces overall costs. Unfortunately, it is difficult to implement on a large scale because it requires frequent in-person visits to a clinic and use of specialized equipment. A feasible alternative is mobile health (m-health), which refers to using wearable sensors to monitor health continuously and unobtrusively. Because certain patterns in the eye and head movements during daily activities have a strong association with mental disorders and dementia, some researchers have suggested using m-health systems in the form of smart eyewear and headset that specifically monitor mental health.
For accuracy and easier understanding by clinicians, the eye-gaze and head movement information captured by these systems must be expressed as rotational angles in three-dimensional (3D) space. However, the conventional manner of representing and calculating these rotations (in Euler angles) requires substantial computing power, and may cause gimbal lock that leads to data ambiguity. A continual search also exists for new methods of extracting mental condition-related features based on the mutual information of the eye and head signals. However, the process of generating these features is time-consuming, making them impossible to deploy in body-worn systems with limited computing power. Quaternion singular spectrum analysis (QSSA) has emerged as a signal processing technique with the potential to resolve the aforementioned problems. The quaternion system is a four-dimensional (4D) hypercomplex number system that provides an alternative manner of representing multiple channels of signals and describing an object’s rotations in 3D space without encountering the problem of gimbal lock. QSSA also inherits the benefits of singular spectrum analysis (SSA), which is superior for denoising and extracting features based on the common statistics of different channels of signals. Previously, quaternion-valued representations of eye and head rotations have been suggested, and the integration of more relevant signals such as pupil size variation and eye blinks into the quaternion for more meaningful data representation warrants investigation. Moreover, no research group has used QSSA to simultaneously process eye-head signals and extract features specific to mental health conditions. Furthermore, limited real-world data exist in this area.
This proposed project aims to fill these knowledge gaps by studying signal processing. The newly gained insights will become the foundation for the signal processing algorithms to be deployed in future m-health systems. First, a new quaternion-valued representation framework that utilizes all four dimensions will be developed. Comprising two quaternions, the framework will host information on eye and head rotations, pupil size fluctuations, and eye blinks. Second, on the basis of expert psychiatric advice and a literature review, the research team will perform computer simulations to synthesize eye dynamics and head movement signals of healthy individuals and those with mental health conditions. Third, an improved version of QSSA will be performed on the synthesized signals to generate a set of new feature signals that consider the joint statistical properties of the eye and head signals. Then using a quantitative approach, the team will shortlist the features that are most associated with mental disorders. Finally, the new features will be entered into a conventional machine learning classifier to examine its ability to differentiate between healthy individuals and those with mental health conditions. This project will provide a more efficient signal processing technique for processing and fusing eye dynamics and head movement signals. This will enable real-time monitoring of eye dynamics and head movements for the early detection of mental disorders and dementia using m-health systems. This accessible approach will enable the early detection of mental health conditions, thus enhancing mental wellness and quality of life in community settings.
Project Reference No.: UGC/FDS24/E07/23
Project Title: Investigation of Hydrogen-enriched Low Calorific Value Landfill Gas Combustion in a Porous Medium Based Burner
Principal Investigator: Dr KAHANGAMAGE Udaya Priyadarshana (PolyU SPEED)
Abstract
Methane (CH4) containing Landfill gas (LFG) is one of the promising renewable sources of energy which is somewhat underutilized due to various techno-economic issues. Currently, high calorific value LFG is utilised for electricity generation, production of synthetic natural gas and heating needs. Low calorific value LFG containing <40 vol% CH4 is mainly managed by on-site flaring or controlled venting which is a waste of renewable source of energy and also harmful to the environment. The low calorific value LFG is undesirable for industrial energy applications due to low energy content, low laminar burning velocity and difficulty in maintaining stable flame. The prior research has shown that fuel enrichment with high quality fuel such as hydrogen, can effectively be used to enhance the combustion performance of low calorific value LFG. Furthermore, porous medium burners are proven to be effective in combustion of low calorific value gases and results in low levels of harmful emissions. However, limited research investigations have been carried out to study the combustion performance of hydrogen-enriched low calorific value LFG in porous medium burners and hence lack the knowledge of optimum operating parameters for efficient combustion and reduced emissions. In this research study, the combustion and emission performance of hydrogen-enriched low calorific value LFG in a porous medium burner is investigated using numerical and experimental techniques with an expectation of developing necessary knowledge and data to develop efficient porous medium based combustion technology for potential industrial process heating.
According to prior research, Silicon Carbide (SiC) porous medium is more effective in combustion of low calorific value gases due to its high heat transfer properties and thermal shock resistance. Numerical model will be developed to investigate the combustion and emission performance of hydrogen-enriched low calorific value LFG (with <40 vo% CH4) in a SiC porous medium burner. ANSYS CHEMKIN will be used with GRI-Mech 3.0 chemical kinetics mechanisms for gas-phase chemical kinetics analysis. The effect of different LFG compositions, hydrogen enrichment percentage, inlet velocities and porous media characteristics on combustion and emissions will be investigated fully. The insight from the numerical analysis will be used to develop a suitable SiC porous medium burner and experimental investigations will be carried out to establish stable combustion region for different fuel gas mixtures and equivalence ratios. The combustion and emission performance data will then be obtained experimentally for different gas mixtures operating within the established stable flame regions. The experiments will be carried out for SiC porous media with different pore densities. It is expected that the numerical and experimental investigations results provide necessary knowledge and fundamental data to develop optimised porous medium based burner technologies to utilise hydrogen-enriched low calorific value LFG for industrial applications.
Project Reference No.: UGC/FDS16/B05/23
Project Title: Examining the role of Benefit Corporations (B Corps) in deriving shared value and achieving Sustainable Development Goals (SDGs)
Principal Investigator: Dr KHURSHID Hamid (HKMU)
Abstract
In recent years, businesses have been called by governments, non-governmental organizations (NGOs), civil society, regulators, the media and development organizations to add value to society. Consequently, new business models such as bottom of the pyramid, social entrepreneurship, benefit Corporation (B Corps) and creating shared value (CSV) etc. have been proposed in the domain of sustainability. The notion of CSV has been proposed as a business strategy for firms to derive economic and social value simultaneously, which in turn enables firms to gain competitive advantage. B Corp has emerged as a new kind of hybrid enterprise that embraces the CSV business model and creates social and economic value simultaneously through its business activities.
After the emergence of “B movement” in 2006, the idea of B Corp has gained popularity across the world, and multinational corporations (MNCs) and small and medium sized enterprises (SMEs) across the world have begun to join the “B movement” and espoused the B Corp business model in order to derive shared value for focal firms and respective communities concurrently. In recent years, the notion of B Corp also gained traction among the business firms in Hong Kong and numerous firms espoused the B Corp business model and deriving economic and social value for their respective stakeholders by integrating market and social logic. The proposed project will draw on institutional logic theory to examine how B Corps operating in the Asian context embrace financial, social and environmental objectives by espousing dual logics (market and social) in their business. B Corps sign up a declaration and commit to create value for all stakeholders; therefore, another key aims of the proposed project is to explore how B Corps collaborate and create ties with various stakeholders in order to derive shared value and address the legitimate needs of the latter. It will also analyze how the engagement of B Corps in the shared value process generates various social, economic and other outcomes for focal firms and their respective stakeholders. One of the other key objectives of the proposed project is to examine how B Corps embrace the agenda of sustainable development goals (SDGs) and contribute to meeting SDGs.
To address the main research questions of the proposed project, an interpretative qualitative research approach will be adopted. We will use the case study method and collect data through one-to-one semi-structured interviews with management officials and employees of 10 Hong Kong based B Corps. To get various stakeholders’ perspectives, interviews will also be conducted with representatives of stakeholder organizations of the focal B Corps.
The overall objective of the proposed project is to develop a grounded theoretical model of how B Corps create shared value by integrating market & social logic and contribute to meeting SDG. This model will illustrate and explain the underpinning processes and mechanisms through which B Corps derive shared value by integrating dual logics (market and social) in the context of Hong Kong. The proposed research will provide a rich set of practical organizational learning process examples. These will serve as exemplars for other hybrid enterprises and for-profit firms operating in Hong Kong and other similar contexts on how to derive shared value and contribute to meeting SDGs.
Project Reference No.: UGC/FDS14/P04/23
Project Title: Probabilistic Classification Error Modeling for Distance Metric Learning and Deep Metric Learning Theory and Applications
Principal Investigator: Dr LAM Benson Shu-yan (HSUHK)
Abstract
Feature extraction plays a crucial role in classification problems in the field of machine learning, especially big data problems. Its purpose is to extract discriminant information from data to enhance the accuracy of classification methods. It has numerous applications in various fields, including language detection, sentiment analysis, fraud detection, image recognition, disease classification, etc. One example is image recognition in computer vision. The problem is to recognize the types of objects, such as basketball, on the basis of a set of given images. This is a challenging problem because images of an object can be taken from different views. Furthermore, the images may have noise and be taken against various backgrounds. However, the noise and the backgrounds do not carry much discriminant information and can confuse many classification methods.
To cope with the noise and background problems, metric learning methods have been developed. The purpose of metric learning is to extract discriminant features from data by optimizing a mathematical objective function. The function converts the data features to a feature space so that samples from the same classes move closer to each other while samples from different classes move further away. Metric learning methods can be broadly categorized into distance metric learning and deep metric learning. Distance metric learning extracts the features from the data using the linear projection method, while deep metric learning extracts the features using multi-layer neural networks, such as convolutional neural networks. However, many objective functions were designed by intuition. It is hard to find the connections between these objective functions and the theoretical error rate. The theoretical error rate is a fundamental concept in machine learning that quantifies the best possible accuracy any classifier can achieve on a fixed probability distribution. Theoretically, extracting the features that can minimize the theoretical error achieves good performance. Moreover, the designs of intra-class and inter-class information are based on some heuristic rules. Intra-class and inter-class information refers to information for the samples drawn from the same classes and from different classes, respectively. However, some designs only include one type (such as intra-class) of information and ignore the other type (such as inter-class).
In this project, we aim to develop a new type of metric learning method based on the theoretical error rate. This new method can be applied to both distance metric learning and deep metric learning problems. The class information can be modelled by probability density functions. Based on the nature of probabilities, both intra-class and inter-class information are implicitly formulated. No heuristic rules are applied. We will also develop new optimization strategies. For distance metric learning, the linear projection can be found by a two-stage optimization method. The solutions obtained are unique. In other words, unlike the current body of work, which uses optimization methods that are sensitive to initial guesses, the methods we propose are not sensitive to the initial guess set. Different initial guesses lead to different solutions. For deep metric learning, we aim to develop a fast approach to find the deep learning parameters. Currently, stochastic gradient descent methods are mainly used to find a solution for deep learning models. The essential idea is to select a min-batch of samples from the data and update the parameters of the layers using gradient descent methods. However, if the min-batch of samples is drawn uniformly from the data, the convergence speed can be slow (take several days) because not all samples carry the same discriminant information. Given a big data problem, the sample size can be over a million. It is hard to draw a min-batch with lots of discriminant information. To alleviate this problem, we aim to develop a new sampling strategy based on probability simulation, namely importance sampling. The idea of importance sampling is to select samples on the basis of a proposal distribution that is like the target distribution. In this project, we will develop a new probability function that has a similar form to the deep learning function. With this new function, we will be able to draw more effective samples and enhance the convergence speed in the training of deep learning models.
Project Reference No.: UGC/FDS24/B11/23
Project Title: What Contributes to Hospitality Green Service Innovative Behaviour? A Multi-level Analysis of Adaptability, Technology Readiness and Organisational Cultures
Principal Investigator: Dr LAM Chin-wah (PolyU SPEED)
Abstract
Several phenomena appearing in recent years have aroused three key issues towards the hospitality industry. First, the onset of the global pandemic and its fallout have been making the ever-changing hospitality industry more volatile. Hospitality organisations have continued revising policies and work procedures or even implementing new sets of policies and procedures to protect guest and employee safety. Such flexible and adaptive working culture has become a new normal and will remain thereafter. Second, the emergence of sophisticated advanced technology, such as artificial intelligence, application platforms and robots, has greatly benefited service operations and customer experience and has been widely adopted in the hospitality industry. The use of smart technologies in hotel organisations has dramatically increased since the occurrence of the global pandemic. Third, hospitality industry has been recognised as the main source of natural resource consumption. Apart from implementing green policies and green services, hospitality organisations have attempted to use green technology, including programmes, tools and devices, to integrate into their operations and management for sustainability.
These issues have been widespread vertically, from operations to management, and horizontally, from the policies and procedures to routine work practices, the pace of innovation has been exponentially increasing. To cope with these challenges, hotel frontline employees, the critical mass for quality service experience, are required to be not only adaptable and versatile, but also to be able to embrace technology and to deliver green service innovatively in a technology- and sustainability-driven hospitality work setting.
The significances of this study aims to integrate the Ability-Motivation-Opportunity (AMO) theory and person-environment fit (P-E) theory, to make an in-depth assessment of the influence of hotel frontline employees’ adaptability on their technology readiness (TR) and green service innovative behaviour (GSIB), and to examine the moderating effect of organisational learning culture (OLC) and green innovation culture (GIC). A conceptual model with a set of hypotheses is developed based on strands of literature on employee adaptability, TR, GSIB, OLC and GIC. Two stages of quantitative and qualitative approaches will be conducted. Structural Equation Modelling will be used to examine the hypothesised relationships in the hospitality context. Following the statistical analysis, a semi-structured multi-level in-depth analysis will be conducted to translate the survey results to the frontline staff, junior/middle managers and senior hotel executives and their views will be collected for comprehensive and holistic interpretations of the findings.
The proposed study primarily contributes to hospitality career and green development theory by providing insights into the importance of hotel frontline employees’ competencies and attitudes towards their organisation and customer-desired GSIB. From the practical perspective, the research findings may provide directions for strategic employee development for hotel operators, thereby enhancing important employee competencies and attitudes, strengthening company’s competitive edge and contributing to the environmental sustainability. This study may offer hospitality education sector insight into equipping and shaping future hospitality professionals. Incorporating AMO theory and P-E fit theory throughout the proposed hypothesised model, the effects among ability, motivation, opportunity and performance may guide organisation area(s) of efforts/resources to be reinforced and to meet demands-abilities fit and supplies-values fit of P-E fit theory. In that sense, the proposed study would be able to provide theoretical insights into multi-disciplinary literatures, including sustainability, technology, attitude and behaviour, human resource development, and organisational behaviour.
Project Reference No.: UGC/FDS16/H08/23
Project Title: From Transformation to Empowerment: A Study of Heritage Tourism in Four Asian Tigers and Its Implications
Principal Investigator: Dr LAW Lok-yin (HKMU)
Abstract
Tourism in the Four Asian Tigers (Hong Kong, Singapore, South Korea, and Taiwan) has been of substantial social and economic significance. Other than urban tourism attractions, the authorities have heavily promoted both built and intangible cultural heritage to domestic and international tourists to display the traditions of their culture. Nonetheless, shaped by an array of social, political, and economic factors, the official narratives did not necessarily align with the local perception of heritage, resulting in dissonance of identities and memories. The proposed research aims to address the dynamic of heritage narratives in post-World War 2 (WW2) Four Asian Tigers, by applying theories and methods from critical heritage studies, tourism history, and digital humanities. The project is divided into two parts: 1) Investigation of heritage discourses in tourism history in the Four Asia Tigers, and 2) development of a Four Asian Tiger Heritage Map System using Geographical Information System (GIS) for teaching and further research purposes.
As for the first part, this project will explore how the Four Asian Tigers societies selected, interpreted, and demonstrated their ‘tradition’ and the ‘past’ in different layers through heritage tourism since the second half of the 20th century, under the context of Cold war tourism and western heritage conservation. The similarities and differences between the Four Asian Tigers’ view on the colonial legacy or cold war politics will be highlighted in this project to search for a new paradigm of heritage discourse from the Asian perspective. While heritage narratives and tourism in Asia are commonly examined through ethnographic studies, this study will delve into numerous unexplored travel-related archives and materials that promote local cultural heritage (e.g., brochures, posters, postcards, advertorials, and other non-printed media by the authority) to look into how tourism depicted or reinvented the narrative of heritage from a novel perspective.
It will also examine unofficial sources recorded by local communities (e.g. heritage tour maps or guidebooks) to compare the heritage narratives expressed by the authorities and the communities in the Four Asian Tigers. The latter usually concealed by the former. It will shed light on the consonance and dissonance between different heritage discourses to demonstrate the intentions and impacts of the authorised version of heritage stories.
On the other hand, this project will develop a Four Asian Tiger Heritage Tourism Map platform to demonstrate the cultural mapping or heritage tour maps in Four Asian Tigers from a spatial perspective. The platform will comprise substantial multi-media information about the transformation of the heritage tour maps in different periods. The team will also invite the public to contribute their private collections about different heritage tours maps or guidebooks. By connecting heritage tourism activities with the sense of space, the geolocation of heritage in the system will visualise heritage data from a new perspective. The platform will not only facilitate heritage research among scholars and students about the evolution of heritage discourse in Four Asian Tigers but also enrich public understanding of the ignored heritage sites and the stories behind the sites to promote in-depth cultural tourism.
Together, the two parts of the proposed project will highlight the hidden voices of heritage practitioners and local communities, which are mostly considered insignificant from the authorities’ point of view. Our endeavors to empower the minorities will contribute to the diversity of interpretations of ‘traditions’ and the ‘past’, particularly in the Asian context. Besides, the project aims to publish peer-reviewed papers in international journals and a monograph to foster academic discussions of heritage discourse in Asia, as most studies addressing heritage discourse are western-oriented.
Project Reference No.: UGC/FDS16/E05/23
Project Title: Development of using retired batteries from electrical vehicles to build smart emergency power systems for buildings in Hong Kong
Principal Investigator: Dr LEE Chi-chung (HKMU)
Abstract
Power outage is not common in Hong Kong but severe incidences can be caused once it happens. For example, a large number of people being trapped in elevators is the norm because of the high density of high-rise buildings. Surely, it will significantly reduce the incidences if most buildings have installed proper emergency power systems with automatic rescue devices (ARDs). With ARDs, it will bring the lift car to the nearest floor and open the lift door during power outage. At the moment, however, few buildings in Hong Kong have such emergency power systems. ARDs are powered by batteries but traditional large capacity emergency power systems rely on diesel generators. Apart from being expensive, emergency power systems with mixed energy sources in general are more difficult for daily operation and maintenance. Recently, batteries have been proposed to energy related applications such as large capacity emergency / stand-by power systems. It has been demonstrated that batteries can replace diesel generators in many occasions to provide the required services. Moreover, batteries are environment-friendly and can much simplify the operation and maintenance of emergency power systems, i.e., professional is not needed. In spite of having such advantages, only using batteries to emergency power systems will be rather expensive if large power capacity is required. Nevertheless, a possible solution has recently emerged from the popularisation of electrical vehicles (EVs) in Hong Kong. For safety and performance considerations, the batteries of EVs are advised to retire within 10 years or their capacities having reduced to 80% of their initial values. Apart from recycling the materials of retired EV batteries, second-life applications of the batteries are highly encouraged since it is more environment-friendly and has potential economic benefit. As supply of retired EV batteries continually increases, cost should no longer be the major concern for using batteries in large capacity emergency power systems. We will have the viable battery based emergency power system after solving the related technical problems. According to this observation, this project proposes a retired EV batteries based emergency power system for Hong Kong buildings.
For the proposed emergency power system, a possible scenario is considered: owing to the shortage of common spaces in Hong Kong buildings, the batteries may be installed at multiple locations instead of a single room in a building. Moreover, batteries in different locations may be retired batteries from different brands / kinds of EVs. They may have different parameters such as capacities, power rates, and ageing levels. Nevertheless, all batteries must be synchronized or properly scheduled to provide sufficient power for the emergency services such as completing the safe landing of elevators at power outage. Without a doubt, the emergency power system must be robust and better have self-healing.
According to the possible scenario, this project will first determine the proper loading model to include all possible emergency services such as lighting, ventilation, and ARD. Based on the loading model, suitable power buses for connecting batteries in different locations of building will be investigated. A distributed battery management system (BMS) will then be developed using Internet of Things (IoT) technology. The status of charge (SOC) and status of health (SOH) of batteries at different locations are monitored to ensure that proper amount of energy has been distributed to / retrieved from each battery according to its status. For BMS operating correctly, however, proper methods for checking / estimating SOC and SOH must be further determined since multiple kinds of retired batteries may be used in the emergency power system. With accurate estimation of SOH, a proper redundancy of batteries will be used to improve the reliability of the system.
Project Reference No.: UGC/FDS13/H01/23
Project Title: Idol Culture, Masculinity, and Hong Kong Identity: A Reception Study of the Mirror Phenomenon
Principal Investigator: Dr LEE Kwok-fong (Chu Hai)
Abstract
The project will be the first scholarly inquiry into the Canto-pop sensation Mirror and the aesthetic of male beauty embodied by the group. The 12-member boy band has taken Hong Kong by storm since its premiere in 2018 amid political upheavals and the Covid-19 shutdown. They have been described as “a bright light amid a gloomy world” and are particularly popular among youths and female fans. The popularity of Mirror both reflects and impacts people’s perception of masculinity, especially among the youth. A context-sensitive reception study of the phenomenon will thus unravel important features of youth culture, subjectivity, and gender dynamics in the rapidly shifting Hong Kong society.
The interdisciplinary project will situate the consumption and reception of Mirror and their connection to the wider processes of identity, subject, and subjectivity formation in the context of contemporary Hong Kong society. It will explore the various factors contributing to the success of Mirror and the gendered impact of this idol culture on Hong Kong youths. The research design consists of three interlocking components: 1) textual analysis of the lyrics of Mirror’s musical productions, as well as media representations and comments about Mirror on social media, 2) focus group discussions with Mirror fans, non-fans, and anti-fans, aged between 15 and 30, and 3) in-depth interviews with practitioners in the idol culture industry. The multifaceted synthesis will make possible a comparison between media representations and real-life enactments of gender among youths. Through empirical study, this project will clarify how we should understand, evaluate, and theorize the phenomenal popularity of Mirror. Through the lens of masculinity, the study will address several important cultural issues in contemporary Hong Kong and will further our understanding of how popular culture constructs Hong Kong identity and gives expression to the associated sentiments.
Project Reference No.: UGC/FDS16/E13/23
Project Title: Fusing Intra-modality and Inter-modality Interactions with Self-attention for Multimodal Sentiment Analysis
Principal Investigator: Dr LEE Lap-kei (HKMU)
Abstract
With the popularity and advance of mobile devices and social platforms (e.g., YouTube and Facebook), the Internet has become a major channel for people to access information and express their views and comments on their daily life, societal incidents, services, and products. There is a huge amount of user-generated content on the Internet every day, usually in the form of diverse modalities, including textual data (e.g., articles, transcripts from audio/video), acoustic data (e.g., speech recordings, voices over video), and visual data (e.g., images, photos, videos). Since the multimodal data is heterogeneous from one modality to another, it is challenging to make full use of the complementary information in different modalities effectively under the presence of contradictory information between modalities.
Sentiment analysis aims to analyze people’s opinions, attitudes, and emotions toward entities such as services, products, stocks, or topics. The sentiment can be broadly classified as positive, negative, or neutral. The ability to extract users’ sentiments can help decision-makers to understand the past, predict the future, and make the right decisions, which is useful for many applications like politics, stock market prediction, movie box office revenue prediction, and customer feedback. Sentiment analysis on texts has been a well-studied problem, as texts are usually easier for extracting semantic information. Yet as shown in recent works, machine learning models based on multimodal information may perform much better than models based on a single modality. For instance, the negative sentiment of sarcastic sentences with positive words like “amazing” and “great” may be impossible to recognize, but can be detected with visual and acoustic modalities, e.g., unpleasant facial expressions, gestures, and tones of speakers. This motivates multimodal sentiment analysis, which aims to analyze sentiments from visual, acoustic, and textual information collectively.
In terms of analyzing multimodal data, early multimodal works are mainly based on manually-extracted features, which are limited in terms of their expressive abilities, designed under limited human knowledge, and also require a fair amount of human power. With the development of deep learning, finding a proper way to fuse features from different modalities is one of the primary tasks in multimodal data fusion. There are two critical challenges. The first is to preserve the information on every single modality and obtain the potential correlation between modalities. Some deep learning models simply concatenate feature representations of each modality directly, preserving unimodality information to the great extent but fails to explore the more hidden interactions between modalities. On the other hand, quite a few works were proposed to model both the intra-modality and inter-modality dynamics but resulted in the compromise of both kinds of information. The second challenge is to pay more attention to important characteristics and less attention to insignificant ones. Most of the existing works employed the attention mechanism in a coarse-grained manner, where more important channels or frames were selected for analysis. It is unclear how to perform fine-grained attention without the need for a high computational cost.
This project aims to design fusion mechanisms to capture and process intra-modality and inter-modality interactions in parallel. To this end, we will fuse multiple classifier tokens received from the feature extraction module, based on a new variant of the self-attention model, Transformer, for multiple modalities. In each of the three modalities, we will learn three classifier tokens instead of one classifier token as in Transformer: one classifier token preserves unimodal information and the two other classifier tokens collect the inter-modality interaction with the other modalities. In particular, to focus on more important characteristics in a fine-grained manner, we introduce a hierarchical element-wise self-attention mechanism to merge the classifier tokens such that greater attention is paid to significant token elements within a channel or frame, i.e., meaningful intra-modality interactions, while suppressing noises and unimportant elements, benefiting the analysis of multimodal sentiments.
Project Reference No.: UGC/FDS16/H17/23
Project Title: A study of modals and modal negatives in the Shantou Southern Min dialect
Principal Investigator: Prof LEE Peppina Po-lun (HKMU)
Abstract
Southern Min is spoken by more than 45 million people, with some 40 million in China, and this project focuses on the Shantou Southern Min (the Shantou dialect). The negation system of Southern Min dialects is a complex one, reflected in at least two ways. Firstly, it is their large array of negatives, consisting of around ten. Secondly, when Mandarin negators are negating modals or aspect markers, relevant negation is done by scope interpretation, with the negator negating element within its c-commanding domain. Unlike Mandarin, Southern Min has the negator and the modal/aspect markers fused together. To account for the large array of negatives in Southern Min, this project focuses on four research issues (i) morpho-syntactic properties of negatives in the Shantou dialect; (ii) modal negatives and the relation of negation with modality in the Shantou dialect; (iii) morpho-syntax and semantics of modal negatives and perfective negatives, and (iv) their implications on the modal and aspect systems in the Shantou dialect. The data used in this project will be collected from targeted investigation via questionnaires and natural data via interviews. Data collected would be good not only for properly documenting our understudied dialects, but for conserving them through compiling preliminary data archives as well.
Project Reference No.: UGC/FDS16/M09/23
Project Title: Ageing without children: the lived experience of childless Chinese older couples
Principal Investigator: Prof LEE Yin-king (HKMU)
Abstract
Background: The number of childless older couples is foreseen to increase with the increase in infertility, voluntary childlessness and delayed marriage. Childlessness is frequently associated with negative connotations, such as guilt, isolation and the discontinuation of legacy. In Western culture, the literature has identified some benefits of childlessness; however, in Chinese culture wherein the responsibility for care weighs heavily on children, older couples without children have no one to take care of near the ends of their lives and arrange their funeral matters after they have passed away. This situation is described as no one pays the last respects in the Chinese culture. The childless older couples are being marginalised as an unfortunate group because they can be left without care and support at a time when they need it most. The World Health Organisation considers ageing well as a global priority. However, whether ageing is perceived positively (ageing well) or negatively (not ageing well) by these childless Chinese older couples is still unknown.
Aim: The aim of this study is to examine the lived experience of ageing in childless Chinese older couples. Caring patterns, expected support network and provisions and strategies for ageing well can be identified from their experiences.
Design: This study will adopt the interpretive phenomenological approach with photovoice as a method.
Participants: Approximately 20 childless older couples (40 participants) will be recruited from one of the co-investigators’ network of childless Chinese couples. The participants will participate in two face-to-face interview sessions with one photo-taking training session in between. The maximum variation sampling method will be adopted to ensure the heterogeneity of the participants in terms of years of marriage, educational background and financial status.
Methods: Interviews will be conducted to obtain an in-depth understanding of the phenomenon of ageing in childless Chinese older couples. In addition to in-depth interviews, this study will employ the photovoice method to enhance its veracity. Photovoice is a qualitative visual research method that requires the participants to take photographs on the research topic by themselves, providing data for analysis. Researcher analyzes the data from the photographs, thereby considerably enriching the data from the interviews. Data analysis will commence at the beginning of data collection and continue for the duration of the research. The themes of the study will be generated by performing van Manen’s thematic analysis together with Oliffe’s photographic analysis.
Results and Conclusions: The result of the analysis will add to the understanding of the meaning of ageing without children from the perspectives of childless Chinese older couples. The various factors contributing to or hindering ageing well, caring patterns and support provision will be revealed. These findings will generate new insights for policy making, health education and the silver market.
Project Reference No.: UGC/FDS11/E01/23
Project Title: Providing Clean Air to Breathe by Wet Electrostatic Precipitators
Principal Investigator: Prof LEUNG Andrew Yee-tak (Caritas)
Abstract
Hong Kong has approximately 600 temples, shrines, and monasteries. Elimination of smoke pollution from burning incense and diesel engines is a challenge. A self-cleaning wet electrostatic precipitator (WESP) is an important pollutant removal equipment and has great potential for greenhouse gas emission reduction. We investigate and enhance its performance by techniques of Computation Fluid Dynamics, charged water films, turbulence creation and hydrophilic coating. The watering system of the WESP is heavy and bulky, we plan to improve it by using vacuum air instead of water for cleaning purposes. Carbon will be captured and sequestrated in the collected fly ashes for carbon neutrality. The collected fly ashes will further be mixed with concrete to enhance its flow property and make it easier to pump in construction. However, the wastewater generated from the WESP process is considered hazardous due to high pollutant loads and cannot be treated efficiently using conventional treatment methods. We use supercritical water oxidation and bag filters to clean up the polluted water for recycling. The WESP will be used with air-handling units to combat airborne diseases on a large scale. Finally, the WESP will be redesigned for use in (1) ships' and power generators' diesel engines and (2) localized solid waste incinerators. In summary, the project aims to balance the freedom of worship and environmental protection, use the heat generated by the furnace to partially power the WESP, utilize fly ash for carbon capture and sequestration, clean and recycle the wastewater, and disinfect the air in a large scale.
Project Reference No.: UGC/FDS15/H02/23
Project Title: Court news in the post-national security law Hong Kong: Examining the significance and journalistic roles, production, and contents of protests-/ politics-related court news
Principal Investigator: Dr LEUNG Ka-kuen (Shue Yan)
Abstract
Throughout the world, the mainstream news media have tended to attach more importance to political news than court news. However, in recent years, Hong Kong (“HK”) has witnessed a very different scenario, where court news has increasingly become an important news genre. In 2019, Hong Kong experienced arguably the most tumultuous political crisis in the contemporary era. The anti-extradition bill amendment movement (“Anti-ELAB Movement”) has resulted in a large amount of court trials, with numerous political figures and protesters being sent to the courts to defend their alleged illegal activities during the period of social unrest. According to the Hong Kong Police’s figures, up until May 2022, more than 10,200 people were arrested among which 2,850 people were charged (Mok, 2022). After the street protests and the political conflicts in political institutions (such as the legislature) had subsided following the end of the movement and the promulgation of the national security law (“NSL”) in 2020, legal conflicts have persisted in the courtrooms between prosecutors, defendants, and judges. Over the past three years, HK media have closely and relentlessly monitored all these court trials, elevating the salience of court news in both media and public agendas (Tang, 2022). It is perhaps for the first time in recent decades that HK citizens are being exposed to legal information on such a large scale.
The growing prominence of court news related to protests and politics deserves more scholarly attention. This study aims to examine the conceived significance and journalistic roles, the production practices, and the content of the associated news reports. For the conceived significance and journalistic roles, it is imperative to find out why HK journalists have paid so much attention to court news related to protests and politics against the city’s democratic backsliding in the post-NSL era. As HK journalists have long held a liberal conception of the press, believing in the monitorial and adversarial roles of the media in providing checks and balances on the authorities, including the judiciary, it is also important to examine whether they would still uphold these types of professional roles when covering court news.
For production practices, this study will identify the political and other non-political factors that may affect the production of court news related to protests and politics. Politically, in face of growing concerns over the decline of press freedom, it is important to examine whether the growing political and legal pressures in the post-NSL era have affected the production of the concerned court news. Organizationally, we will pay attention to the relationship between the media and the court and examine whether and how it may also affect the room for newsgathering practices for court news. Technologically, with the digital transformation of the news landscape, there has been a rise in online alternative media and social media influencers specializing in court news in recent years. We will examine how they make use of digital platforms to produce alternative court news. At the individual level, the mainstream and alternative court news journalists may also differ in their conceived significance and journalistic roles in covering court news. In order to map out the new ecology of court news production in the post-NSL era, the differences in the production practices as well as the contents between the mainstream and alternative media necessitate investigation. Finally, with regard to contents, this study will examine what sorts of court information have been communicated to the public. We will analyze their content features, including their news values (e.g., deviance, conflicts, and importance), framing, and enacted professional roles.
To address the above conceptual issues, we will mainly conduct qualitative in-depth interviews with the court news journalists and a quantitative content analysis of the court news.
As HK has witnessed paradigmatic changes in the legal system, the political field, and the news landscape in the post-NSL era, a study of court news would provide a unique theoretical window to re-examine and reflect upon their changing relationships in the new political and legal context of HK. To the best of our knowledge, this study will be the first of its kind in HK, which will fill the long-standing theoretical void of court news research locally.
Project Reference No.: UGC/FDS11/E02/23
Project Title: Developing Efficient Cartoon Animation Editing Pipeline via Deep Entity Recognition and Motion Analysis
Principal Investigator: Dr LI Chengze (Caritas)
Abstract
The cartoon and animation industry has a long history of using hand-drawn techniques to create visually appealing and engaging content. However, the production of traditional cartoons and animations can be labor-intensive and time-consuming, requiring significant human resources to draw and color each frame (key animation, or genga, in Japanese). This can limit the overall production capacity of the industry and make it challenging to meet the demand for high-efficiency cartoon editing to deliver new content from existing works. The new contents could be the HD remastering of old classic animations, 3D stereoscopic remakes, editing of the characters and the backgrounds, or a retargeting to mobile-friendly vertical short-video format (e.g., for TikTok.) This project intends to propose a deep learning based system to implement an automatic analysis and editing pipeline for cartoon animations. The proposed system shall be able to learn from large amounts of data from cartoons and animations to analyze and recognize the characters, subjects, and objects depicted in the content as well as their movements and interactions. With the learned knowledge, the proposed system is further designed to empower automation in multiple anime editing tasks that are currently done manually, such as re-shading, character and background modification, depth analysis, and scene compositing. More importantly, the system does not require any key animations, which will allow direct processing and editing for any arbitrary animations with minimum human effort.
In implementing such a system, several challenges will need to be addressed. First, obtaining sufficient high-quality data from cartoons and animations with fine-grain annotation of the objects and their motions shall be necessary. Second, we shall propose using a cross-modal transformer model with strong generalization and expression abilities and few-shot learning capabilities to understand and interpret the content of the cartoons and animations in conjunction with their motion information. To the best of our knowledge, there are no existing solutions to this task, due to the stylized and exaggerated nature of cartoons and the lack of relevant data. We believe this approach can outperform traditional key animations in terms of accuracy and speed. More importantly, the few-shot learning ability of transformers should make it effective to ground any arbitrary objects even without prior knowledge. Finally, effectively using the information obtained from the model to achieve downstream tasks is a challenge that has not been researched before. Our goal is to identify the critical, labor-intensive steps in the downstream editing applications such as retargeting, stereoscopic remastering, and text-guided object editing. Then, we will incorporate object and motion information from the previously learned model into deep learning solutions to automate these critical steps to improve overall efficiency and precision in cartoon editing tasks.
In the final stage of this project, we plan to integrate the proposed system into the existing workflow of the cartoon and animation industry. To do this, we shall develop a user interface that allows users to input the desired processing tasks and parameters, and the system will automatically apply the appropriate algorithms and techniques to achieve the desired results. We will also set performance goals for the system in terms of accuracy, speed, and efficiency, and will optimize the system architecture and algorithms to meet these goals. This proposed system is expected to significantly streamline the process of updating and repurposing existing cartoon and anime works, making it easier and more cost-effective. In addition, the pipeline could help to democratize the creation of high-quality cartoons and animations by enabling individuals and small teams to produce professional-grade content, fostering a more vibrant and diverse anime community. Besides the impact on the cartoon industry and the community, this project should also benefit the research community by advancing state of the art in computer vision, computer graphics, and natural language processing. It shall also provide valuable learning opportunities for institute students studying digital media and AI technologies.
Project Reference No.: UGC/FDS11/H06/23
Project Title: Rearticulating Modernity in China: Trains and Railways in Wartime Literature and Film, 1937-1958
Principal Investigator: Dr LI Siyi (Caritas)
Abstract
The study of trains and railways as well as their cultural representations have received increasing attention from scholars across the disciplines, as they played an important role in the shaping of modernity through mechanical acceleration and capitalized globalization. There has been a great deal of research on trains and railways as related to literature, cinema, nation-state, urban space, technology, women and gender in Western scholarship. Thinking comparatively, what is the role of trains and railways in shaping the perception and experience of modern time and space in China? What is the relation between railways and the formative process of China from an empire to a modern nation-state in comparison with the situation in the West and the other Non-western states? What is the influence of trains and railways on modern Chinese culture and literature? Did the representations of trains and railways provide a new sense of speed and an alternative expression of modern experience for China different from the West? To answer these questions and fill the lacuna of research on railway modernity in China, the PI would like to undertake a comprehensive cultural and literary study on trains and railways in modern China.
The proposed research is built on the PI’s previous project on railway modernity in China from the late Qing to the Republican period (1840-1937). The PI attempts to extend the research to 1937-1958, which in the meantime China was undergoing three wars (the Second Sino-Japanese War, the Civil War, and the Korean War). It will investigate the literary works and cultural representations of trains and railways within the framework of modern Chinese writers and intellectuals’ wartime experience of speed and mobility. Drawing extensively on historical materials (newspaper reports, private records, databases, oral histories) and cultural works (literary texts, paintings, photographs, and films), the proposed project will be divided into three parts to investigate railway modernity in China in the state of exception, especially how the experience of mobility, the new sense of speed, the bodily perception of violence affected on the modern Chinese literature and culture.
The first part of the project will analyze the related works created by the Northeastern diasporic writers, the literatus in the isolated-island Shanghai, the intellectuals and artists based in National Southwestern Associated University (Lianda 西南聯大) as well as in the occupied Hong Kong and Taiwan, exploring the dynamic correlation between the writers’ experience of mobility and warfare and their literary writings on trains and railways. The second part will look into the representations of trains and railways in various visual media, especially the films concerning the three wars, aiming to identify the different metaphoric meanings and narrative functions of trains and railways in Chinese visual culture from the republican period to the socialist era. The third part will examine the intellectual discourses and political trends on modernity and warfare from 1937 to 1958, highlighting how nationalism, modernism, the cosmopolitan left-wing and anti-imperialist thoughts interplay with the social reality and cultural imagination produced by trains and railways.
This project will fill in the gaps from existing research on twentieth-century China, by adopting a new research paradigm as railway modernity to investigate the two understudied areas in the experience of mobility and speed and the cultural representations of warfare. It will enrich the study of modern Chinese literature and film and advance our understanding of literature and war, the critique of modernity, and the Sinophone studies in today’s world.
Project Reference No.: UGC/FDS16/E08/23
Project Title: Long-term adaptation strategy for climate-resilient coastal bridges subjected to extreme risk under typhoons
Principal Investigator: Dr LI Yaohan (HKMU)
Abstract
Catastrophic typhoon events have resulted in significant economic and social consequences in recent decades. Due to climate change and population concentration, the vulnerability of coastal communities and civil infrastructure systems (e.g., coastal bridges) subjected to typhoons keeps increasing. It is paramount to ensure the climate resilience of coastal bridges to minimize direct losses and reduce losses due to disruption. However, the extreme typhoon risk and the changing climate have not been adequately considered in most existing bridges designed based on past design specifications, thus these bridges can be susceptible to future climate scenarios. Therefore, a long-term adaptation strategy is needed to achieve climate-resilient coastal bridges against the extreme risk of typhoons under climate change uncertainties.
Appropriate adaptations for climate-resilient coastal bridges rely on a proper evaluation of several performance indicators: structural vulnerability, resilience, and risk. However, in terms of vulnerability, most of the previous studies investigated the bridge performance against typhoon-induced waves and storm surges based on empirical formulations, while such a method may result in an improper estimation of bridge vulnerability and the following risk assessment. This project aims to address this issue by establishing a three-dimensional computational fluid dynamics model with the validation of experimental results. For resilience and risk assessment, past studies have widely assessed resilience and risk under a single hazard event but ignored the long-term impact. Under a changing climate, the nonstationary climate drivers may result in an increase in typhoon frequency and intensity. Thus, a nonstationary model is needed to quantify the associated uncertainties in typhoon frequency and intensity. In addition, the extreme typhoon risk under the deep climate change uncertainties has been ignored in previous research. Nevertheless, the extreme risk is associated with the tail risk of economic loss. Neglecting the associated tail risk can lead to low probability but tremendously high economic consequences. This project will solve these problems by developing a long-term assessment approach to quantify long-term resilience, risk, and extreme typhoon risk considering climate change impact. Instead of the conventional method purely based on the expected economic loss, this project measures the typhoon extreme risk by proposing higher-order analysis and tail risk measures.
Given the assessment of structural vulnerability, long-term resilience and risk, and extreme risk, a robust multi-criteria optimization approach can be ultimately developed to determine the optimal long-term adaptation strategy for coastal bridges. Then the performance of adaptation strategies can be effectively evaluated under different uncertain climate change scenarios. The proposed long-term adaptation strategy will be robust and cost-effective to meet the various preferences of decision-makers against extreme risk. In general, the proposed approach will address the limitations of existing methodologies in terms of long-term impact, nonstationarity, deep uncertainty, and extreme risk associated with climate change. The research outputs of this project can be applied to coastal bridges in the Greater Bay Area and help achieve climate-resilient civil infrastructure systems and networks.
Project Reference No.: UGC/FDS16/E10/23
Project Title: A Generic Label Extension and Enhancement Framework: from Emotion Domain to General Domain
Principal Investigator: Dr LI Zongxi (HKMU)
Abstract
Multiclass text classification tasks, such as emotion classification, sentiment analysis, and spam detection, have broad applications in daily life. Conventional paradigms for such tasks in natural language processing (NLP) use a one-hot label to calculate the cross-entropy loss as the optimization objective. Although this approach is efficient, research has found that the one-hot label can lead to overfitting and cannot provide sufficient guidance in training, especially in sentiment- and emotion-related tasks, where the labels are subjective and ambiguous.
Researchers have recently attempted to use an enhanced label in classification tasks. The enhanced label is a distribution describing the relationship between the text and all the labels. Studies have shown that the enhanced label helps train a more robust classifier than the one-hot label. However, this approach is problematic. Training with an enhanced label potentially harms the classifier’s prediction ability because the distribution learning loss contributes to all directions in gradient descent. Furthermore, as the manually annotated distribution label is rare, the label enhancement process must be conducted unsupervised, thus yielding a less reliable distribution. Therefore, the power of the distribution label has to be suppressed.
We have proposed a label extension scheme to address the abovementioned issue for emotion-related tasks by extending the labels with fine-grained subcategories. In the framework, distribution learning is applied in the extended label space. The learned distribution is mapped back to the original space for refined label prediction. As distribution and classification learnings are configured in separate label spaces, they do not interfere with each other, thus avoiding the dilemma. Preliminary results indicate that the label extension scheme can substantially improve the baseline models without introducing additional parameters. Relevant studies have been published in international conferences and academic journals with high impact-factors, and we have received very positive comments from reviewers and editors.
The framework’s essential components are the construction of extended label space and the generation of sentence distribution. Focusing on these two components in the proposed project, we will aim to explore possible methodological improvements from two perspectives based on our prior studies: (1) discovering the optimal strategies for extension and enhancement in emotion-domain tasks; and (2) devising a generic framework for tasks in different domains.
As the intermediate steps, both extension and enhancement can be carried out through various methods but have trade-offs. In the extension stage, having more subcategories for each label can incorporate more dependency information of labels, but it can cause difficulties in distribution learning and may affect prediction accuracy. In the enhancement stage, the sophisticated learning-based method can produce a descent sentence distribution with computational overhead as the cost; a naïve rule-based method can be efficient in use but requires human effort in handcrafting rules, and the resultant distribution may be less reliable due to the bag-of-words nature. In this project, we will investigate different configurations of both stages to determine the optimal strategies for achieving the best model performance.
The major challenge of applications in general domain tasks is the inability to determine definite subcategories and calculate the distribution given the lack of domain knowledge. To address this challenge, we will propose a semi-automated method where the subcategories are defined manually using common sense. A semantic tree will be used to find seed words for constructing a concurrence graph. The distributions can be learned via graph embedding. Furthermore, we will reconsider the necessity of having definite subcategories for the extension. Like latent topics in topic modeling, subcategories can be denoted as latent semantic clusters that are discovered via a data-driven approach. Therefore, an automated extension and enhancement framework that can be applied to different domain tasks would be practical.
The proposed label extension and enhancement framework aims to improve the model performance by augmenting label information. This project will generalize the framework from emotion-domain to general-domain applications. Our research will benefit the NLP community by shedding light on how to effectively utilize the labels.
Project Reference No.: UGC/FDS16/H27/23
Project Title: “Shanghai fever”: Japanese travel writers and modern Chinese writers in the 1920s
Principal Investigator: Dr LO Man-chi (HKMU)
Abstract
During the 1920s, many Japanese writers travelled to Shanghai and interacted closely with the Chinese writers in the city, constituting the phenomenon of “Shanghai fever”. However, this intimate cultural connection between Japanese modernity and cosmopolitanism in modern China has long been relegated to historical obscurity. The Shanghai fever among Japanese writers in the 1920s constituted a short but significant episode in the history of Sino-Japanese exchanges during the interwar period, writers from both countries shared common literary values that transcended national borders. By illustrating a fuller picture of the writer-tourist scene at that time based on hitherto-neglected source materials, the project will explore the large-scale Shanghai fever present among Japanese writers from different literary camps in the 1920s, the friendships and conflicts between modern Chinese and Japanese writers, as well as modern Japanese writers’ impressions of Shanghai and its literary world.
Project Reference No.: UGC/FDS13/E01/23
Project Title: Automatic Landmark Object Extraction and Image-based Visibility Estimation System
Principal Investigator: Prof LO Wai-lun (Chu Hai)
Abstract
Visibility is an important safety indicator for road traffic, as under bright background conditions, an object on the ground can only be detected or recognised if it is within the visibility distance threshold. Visibility is also used as an environmental parameter for pollution and weather conditions’ monitoring.
In past research, we extracted image features and used meteorological laws for visibility estimation calculations. However, the accuracy of these approaches depends on image quality and is affected by noise. It is also difficult to extract all of the environmental factors via such approaches and formulate an equation relating these factors to visibility. Thus, artificial intelligence (AI) approaches are currently used for visibility estimation.
In preliminary research on AI-based visibility estimation, we used a pre-trained convolutional neural network to extract image features from webcam images. Instead of using whole digital images, effective areas were identified, and a single correlation variable was used to select suitable image features. A generalised regression neural network (GRNN) and a deep learning algorithm were designed and used to estimate mapping between visibility and selected images, with use of a visibility database provided by the Hong Kong Observatory (HKO). The results showed that this method is more accurate than the classical method based on handcrafted features. However, this method is not fully automated, as effective regions of image data are selected manually using expert judgement. Furthermore, the accuracy of this method is limited, as image features are selected using only a single correlation coefficient and low-resolution webcam images. Finally, the data training and application scope of this method are limited by the HKO-provided database. Thus, the developed algorithm needs to be tested with different datasets.
In this proposed project, we will investigate and develop an automatic landmark object (LMO) extraction system that can extract LMO regions from weather photos and identify the LMOs’ effective visibility estimation ranges. The mapping between LMO’s regions’ features values and visibility ranges for different ranges are approximated by an artificial neural network (ANN). A multi-class ANN visibility estimator will be developed in this project. We will also build a hardware prototype of the multi-class ANN Visibility Estimation System have the following functions: (i) automatic visibility meter data collection; (ii) automatic weather photos, image data collection; and (iii) wireless communication with a host computer (e.g. via Wi-Fi). Finally, the multi-class estimation algorithms will be implemented in a host computer. The team of this proposed project will evaluate the LMO extraction system and the Visibility Estimation and Training System by conducting experimental studies. The major outcome of the project is a low-cost and automatic visibility estimation system which contribute to environmental monitoring technology.
Project Reference No.: UGC/FDS24/E12/23
Project Title: Exploring the Use of Sonic Crystal in Noise Barriers
Principal Investigator: Dr LOH Anthony Wai-keung (PolyU SPEED)
Abstract
Road traffic noise is a major concern in Hong Kong since a high proportion of our population dwells in close proximity to heavily-trafficked highways. Besides a deterioration in the quality of life, it has serious health impacts, particularly on the cardiovascular and immune systems. Road traffic noise levels depend on a large extent on the volume, speed and mix of vehicle types. To protect the environment, the Government has implemented policies to encourage the use of electric vehicles. However, they produce noise from road/tire friction at comparative levels with vehicles with internal combustion engines (ICE). The noise barrier is the most commonly used technical sound abatement solution in Hong Kong. The Government has issued strict guidelines on their design and construction, and there are constraints, including their wind load, availability of pavement space for their structural base, aesthetic appearance, costs, etc. Over time, much progress has been achieved on designs and materials used in noise barriers.
Research work in prior years focused on top-edge designs for a reduction in sound diffraction behind the barrier. In recent years, however, there has been a proliferation of research work on Sonic Crystal (SC) acoustic screens, consisting of cylinders arranged in a predetermined lattice, to attenuate the sound of certain frequency ranges (called Band Gaps). This is possible due to destructive Bragg interference caused by the different periodicity of the heterogeneous material (rigid cylinders and air). Cavities built into the cylinders and absorbent material covering the cylinders or their cavities can further tune and widen the Band Gaps and enhance their sound attenuation performance. This can be achieved through a process of Band Gap engineering and optimisation. Our proposed study focuses on finding practical solutions. Via numerical simulation and prototype testing for proof of concept, we search for an optimal SC noise barrier design with improved acoustic performance. We will compare the SC noise barrier performance against a conventional barrier by building the SC noise barrier prototypes from the same material as the conventional noise barrier. Tests will be performed in an anechoic chamber, and absorbent material covering as well as cavities on the cylinder walls will be introduced as long as the budget is available. Experiment results on overall frequency attenuation, Insertion Loss (IL) and Transmission Loss (TL) will be analysed for the effectiveness of our proposed design and its potential practical application. Hopefully, the results of our study will be useful for motivating academics and local industries to research development in this area.
Project Reference No.: UGC/FDS14/E06/23
Project Title: Impacts of Fairness on Cabin Crew Scheduling: Crew Preferences and Allowances
Principal Investigator: Dr MA Helen Hoi-lam (HSUHK)
Abstract
Fairness in workplace is crucial. Many companies strive their best to create a fair employee experience to maintain good spirits and thus competitiveness. To maintain fairness, it requires transparent systems and implementation mechanisms. In airline industries, transparent systems in regulating salary, promotion, holiday, etc. are usually present to facilitate the culture of workplace fairness. To treat cabin crew well, airlines have implemented a crew preference scheme, which allows cabin crew to express their preferences on various items, e.g., holiday, destination city, teammates, etc. Then, airlines will try to satisfy these items to the greatest extent during the crew scheduling process, i.e., during the generation of their rosters (itineraries). Crew scheduling problems aim to determine the proper assignment of crew to provide services on flights in the flight networks. Thus, the quality of the crew schedules generated significantly affects not only the airline operating costs but also the quality of life of the crew members. This is because, to airlines, a high-quality schedule can minimize the total idling or total transit time between flights. As a result, the total number of crew members required to support flight operations can be reduced. To crew members, the quality of crew schedules influences their work patterns, holidays, teammates, etc., and hence implies the number of desired preferences being fulfilled.
In the economic point of view, airlines can generate rosters entirely for their own benefit as long as the rosters satisfy the regulations governed by the civil aviation department. However, many renowned airlines care for their crew members and significantly consider the fulfilment of crew preferences. However, in the existing implementation mechanisms, crew preferences are only considered during the last phase of crew scheduling, which is called the crew assignment phase. Therefore, the fulfilment of crew preferences is limited. More importantly, fairness in the number of preferences being fulfilled between different rosters is not being considered. As a result, certain crew members may be assigned schedules that are more favorable to them than to others, thereby causing unfairness. Furthermore, in the existing airline practice, (in particular legacy airlines), the salary of a cabin crew consists of the basic salary plus allowances. If crew members are being scheduled with only turnaround flights, the crew members will not earn any allowances due to the lack of overseas staying. As a result, their salaries will be lower and again it causes unfairness. This kind of unfairness recently has been seriously raised and complained by cabin crew as seen on headline news (e.g., Cathay Pacific Airways), “cabin crew member rousing for a fair crew scheduling mechanism”. Motivated by the abovementioned issues, we propose to develop a novel Crew Scheduling with Fairness in Crew Preferences and Allowances (CS-FCPA) model. Our objective is to study the impacts of considering fairness in crew preferences and allowances on airline operating costs and crew members. To achieve so, long-haul and short-haul flights should be scheduled together during crew scheduling because of the assignment of overseas staying. However, as this will exponentially increase the problem complexity, our proposed novel CS-FCPA model combines the Benders decomposition method with the column generation algorithm to decompose the problem and reduce the computational burden. This model is novel in the field of study. Moreover, we will propose novel acceleration strategies based on the problem characteristics. This project provides not only theoretical breakthroughs in crew scheduling studies but also practical contributions to airline industries and cabin crew. Our findings can provide novel insights into airline cabin crew management which consequently are essential to the development of the airline industry. The insights also provide scientific guidance on the implementation of fairness in workplace culture in airline industries.
Project Reference No.: UGC/FDS14/B06/23
Project Title: A multilevel model of workplace happiness: Examining the effects of internal market orientation, supervisory trust on employee work outcomes and organisational performance
Principal Investigator: Dr MAN Thomas Wing-yan (HSUHK)
Abstract
This research project aims to examine the relationship between leadership internal orientation and employee customer orientation, mediated by employee workplace happiness. In today's rapidly changing and uncertain world, global leaders face not only external challenges like pandemics and climate change while also grappling with unethical leadership practices. At the same time, Various industries are witnessing a significant workforce attrition known as the "great resignation." To address this issue, organisational leaders must prioritise employee retention and well-being. By adopting an internal market orientation characterised by connection and care for employees, leaders can contribute to employee workplace happiness and engagement. This, in turn, leads to a happier and more satisfied workforce, resulting in improved business performance and customer satisfaction. The project will also explore the moderating factors that influence this relationship, such as supervisory trust, organisational identification, and psychological ownership. Data will be collected from 150-200 organisations to investigate these relationships, and a workplace happiness index will also be developed. The research findings will have both theoretical and practical implications, advancing the understanding of workplace happiness and its integration into management and organisational behaviour theories. Furthermore, the project will inform business practices by emphasising the importance of a supportive organisational orientation in talent management. By addressing ethical leadership and employee well-being, this research project provides valuable insights for organisational leaders and managers striving to create a positive work environment and enhance business performance in the face of contemporary challenges.
Project Reference No.: UGC/FDS24/E15/23
Project Title: Development of Mycelium-Based Composites with Enhanced Mechanical and Thermal Properties for Diabetic Foot Insoles
Principal Investigator: Dr NG Sun-pui (PolyU SPEED)
Abstract
Diabetes mellitus (DM) is one of the most concerning diseases in the world. According to the Diabetes Atlas 2022 Report by the International Diabetes Federation, around 537 million adults aged from 20 to 79 were suffered from DM in 2021 and it is predicted to have an increase to 643 million by 2030. Diabetic foot is one of the common complications resultant of DM and the estimated proportion was 72%. Due to the changes in the mechanical properties of the plantar soft tissues and sensory neuropathy, foot ulceration and infection are frequently found which deteriorate the quality of life of diabetic patients significantly. The function of diabetic insole is to reduce the magnitude of the pressure on the plantar side of the foot so as to minimize the risk of foot ulcers, calluses and foot pain. Diabetic insoles are normally made of leathers, plastics or other moldable materials to fit the feet shapes for individuals. These traditional materials having suitable properties for pressure off-loading are primarily used in current practices. However, those synthetic materials have poor breathability and heat retaining properties resulting in a high level of foot discomfort from heat and perspiration. Sustainability issue is another critical concern since they are non-renewable and non-biodegradable. In particular for fibre-reinforced polymers (or plastics), this kind of composite material is made of a polymer matrix reinforced with a fibre material. With regard to the prevailing composite materials in the market, carbon, glass and aramid fibres are commonly used but the raw materials for making these fibres are limited, and the recyclability processes for these fibres are costly and inefficient. On the other hand, the manufacturing processes of the resin materials lead to high volatile organic compound, pollutant emissions and energy consumption. Hence, the trend towards the development of “green” fibres as the reinforcement of polymer matrix materials in composites is growing rapidly nowadays.
Mycelium is the most preferable replacement of inorganic fibres and plastics because of their lower density, low cost, good readiness for the environment and with lesser associated health and safety problems. Mycelium is the root fibres of mushrooms composed of multi-layers of natural polymers forming a network structure. It acts as a self-assembling binder to hold organic materials together which functions as the reinforcing fibres and thus, mycelium-based (MB) composite is fully biodegradable. By changing the fungi species, substrate mixtures and processing techniques, mycelium materials have a wide variety of properties and potential applications. Mycelium has been used tremendously in various applications including both low-value and high-value products like packaging, thermal insulator, acoustic absorption panels, flame-retardant materials, footwear products, etc. However, weak mechanical properties and high thermal insulation of mycelium are the main problems that hinder the application in orthotic insoles. To overcome these problems, this proposed project aims to develop new MB composites with enhanced mechanical properties using “eco-friendly” fabrication methods. To provide further resilience and pressure-relief property to the MB composite insole, bio-resin will be infused into the mycelium fibre by resin transfer molding method. The resin material will be mixed with the micro-size graphene platelets. This kind of high strength reinforcing filler in polymer resin not only enhances the strength of the MB composite, but also improves the thermal conductivity of the material. Since the distribution of plantar pressure over the entire insole surface by the new MB composite material cannot be accurately measured and analyzed by experimental means, a finite element model for various insole designs with different material properties and geometries will be developed and hence the mechanisms of peak plantar pressure reduction can be investigated. Upon project completion, a MB composite insole with optimum material property and thickness together with a finite element modelling method for designing MB composite insole will be obtained.
Project Reference No.: UGC/FDS14/P02/23
Project Title: Principal contrast analysis of high-dimensional time series
Principal Investigator: Dr NG Timothy Chi-tim (HSUHK)
Abstract
In the Big Data Era, econometric data, and air pollution data etc. contain thousands of time series collected over a long period of time. For example, a dataset can cover N=1017 stocks trading in Hong Kong Stock exchange during the period of T=1315 trading days from January 1, 2008, to January 30, 2013.
The goal of this research is to group the time series with similar statistical properties into groups and/or split the period into regimes. Time series belonging to the same group are considered more closely related. In addition, the statistical property of the data changes from time to time. It is important to obtain information from the most recent regime for the prediction and identify the obsolete data. For such purposes, novel concepts of principal contrast are introduced to capture the information about the group membership and regimes in the high-dimensional settings with thousands of series. The proposed method can identify the effects of the regime switching on different time series.
Computation method and software are developed to find the contrasts and therefore identify the groups and regimes for applications in finance, business, and geography. Statistical properties of the proposed method are established to support its correctness.
The results can be applied to for example, the stock return data of huge number of stocks, econometric data of huge number of cities within a country, sales data of huge number of chain stores, paleoclimatic data collected from different sites all over the Earth.
Project Reference No.: UGC/FDS14/H13/23
Project Title: Poetic Justice and the Anglophone Novel, 1678-1928
Principal Investigator: Dr PARKER Jay Thomas (HSUHK)
Abstract
Literature has long been valued for its ability to support ethical and political thought. This project examines poetic justice and the novel in English. It will use digital and traditional approaches to understand the resources that the Anglophone novel has to offer in debates surrounding justice. It focuses on the under-theorised notion of poetic justice, proposing first, to study how this concept has changed over time, and second to use the emergent theoretical to examine how forms of poetic justice in the Anglophone novel can help us to think about justice in general. Poetic justice is considered an unrealistic trope. At its inception, this reflected an aesthetic demand that literature should be morally better than the world, representing just outcomes in fiction that were implausible or unlikely in reality. Today, the moral duty of serious literature is more likely to be the opposite, namely to embody a moral realism that reflects and highlights the world’s injustices. Yet there is also something utopian about Poetic justice: both utopia and poetic justice seem to postulate an impossibly ideal society or world. Fredric Jameson has argued that conceptually “reviving utopia” is an essential step in defining “the outlines of a new and effective practical politics” (2004), because it is only this concept that can fuel our ability to imagine political transformation for the real world. This project proposes that examining poetic justice in the novel might provide similar resources for thinking about or theorising real world justice, and seeks to understand what those resources might be by redescribing poetic justice and exploring how that redescription sheds new light on representations of justice and injustice in the Anglophone novel.
Project Reference No.: UGC/FDS15/B09/23
Project Title: Mastery and helpless responses to proactivity setbacks: The role of implicit person theory
Principal Investigator: Dr PENG Zhengmin (Shue Yan)
Abstract
Employees are expected to proactively improve deficient processes and address issues caused by the increasing unpredictability and uncertainty of workplaces. It is also found that such proactive work behaviors are important to ensure workplace effectiveness (Campbell, 2000; Carpini, Parker & Griffin, 2017; Frese & Fay, 2001; Parker, 2000). However, employee proactivity research focuses on the motivational states that drive employees to proactively initiate change but pays little attention to the setbacks that they may encounter in doing so. We know little about why some employees who encounter proactivity setbacks attempt to overcome them, whereas others do not. That is, why do some employees regard proactivity setbacks as energizing their goal-regulation processes, whereas others regard them as inhibiting their goal regulation processes? The proposed project will answer this important question because this difference determines the ultimate success or failure of the proactive endeavor.
Based on an integration of the proactive goal-regulation model with implicit person theory, we argue that employees may exhibit either mastery or helpless responses to proactivity setbacks, depending on what kind of implicit person theory they hold. Specifically, employees who hold an incremental implicit person theory are likely to engage in mastery reflection on proactivity setbacks, which sustains their proactive efforts to envision, plan, and enact their goals. By contrast, employees who hold an entity implicit person theory are likely to engage in helpless reflection on setbacks, which leads them to decrease their proactive efforts to envision, plan, and enact their goals. Therefore, the type of reflection engaged in by employees dictates whether they continue changing or taking on challenges. We will examine these arguments in two experimental studies and a field weekly diary study. We will also conduct studies to develop a measurement scale for proactivity setback, which is newly developed in this research.
Overall, we have identified a major gap in current proactivity research: a lack of understanding on proactivity setbacks and how employees’ responses to them affect employees’ proactive goal-regulation process. In this proposed project, we extend the employee proactivity literature by conceptualizing proactivity setback based on the proactive goal-regulation model and two forms of reflection based on the implicit person theories to examine the dynamic relationships between the proactive goal regulatory activities. Theoretically, this work delineates employees’ differential responses to proactivity setback, and shed light on proactivity literature concerning how to make effective proactivity resiliently. Practically, this work will provide an evidence-based guidance on facilitating and nurturing employees’ proactive goal-regulation through cultivating an incremental implicit person theory and facilitating mastery reflections in employees. These insights can be integrated in employee/manager training and organizational culture development programs to benefit employees, managers, organizations, and communities.
Project Reference No.: UGC/FDS16/H16/23
Project Title: Stage for Positive Ageing: Oral History Theatre by Older People
Principal Investigator: Dr SHU Chi-yee (HKMU)
Abstract
This project aims at exploring the possibility of oral history theatre as a healthy form of elderly theatre, i.e. theatre of, for and by older people. In an era of the ageing population, this research on the theatre form, together with its impact and methods, will be an important reference for the fields of applied theatre as well as positive gerontology, with respect to the concern about successful and productive ageing.
The PI will observe the practice of a Hong Kong professional theatre company, namely Chung Ying Theatre Company (中英劇團), and follow their creation of oral history theatre from the play-writing process to the performance. Since 2022, around 300 retired older people and 30 young people have been recruited by Chung Ying for devising the plays, who have been performers themselves in one to two performances. It is foreseen that after 2024 Chung Ying will continue to stage oral history theatres and elderly theatres because such theatre forms have been their signature art forms since 2008.
In this performance ethnography research, there will be mixed methods of questionnaire and case studies. First, there are case studies on the oral history theatre’s impact on both the older participants and young participants, to find out the older participants’ theatre experience in terms of positive ageing and productive ageing, and also their views on any improved intergenerational relationship. The experience and views of the young participants will be studied too. Questionnaires will be distributed to these older and young participants, and also to the audience members of selected shows. Quantitative and qualitative data will be triangulated and analysed, with coded analysis of interview data and performed scripts and images.
Second, the directors’ devising methods involved in the creative processes will be observed through their conducting of workshops and rehearsals, as case studies. This aims to study how the devising directors maintain the artistic quality of the theatrical works, including content selection and fictionalization (i.e. play-writing) and mise-en-scène (i.e. directing) while considering the empowerment of the older and young participants. In-depth interviews of the directors will be conducted and triangulated analysis will be done.
The project will last for 30 months from beginning to end. The deliverables of the project will include an online and printed practical book including a video-and-script package for dissemination to practitioners and researchers in related fields. Also, at least two journal articles will be produced from the impact study and the case studies mentioned above.
Project Reference No.: UGC/FDS16/M16/23
Project Title: Storage conditions and brewing methods for preserving catechin contents, reducing fungal contaminants, and minimizing mycotoxin residues in four types of tea in different packaging
Principal Investigator: Dr TAM Emily Wan-ting (HKMU)
Abstract
After water, tea is the most consumed beverage worldwide and is regarded as a safe drink. As many studies have claimed that tea has a wide variety of health benefits, there is a growing trend in tea consumption. For instance, tea contains catechins, which have proven beneficial effects. However, tea can become easily contaminated by fungi during processing and because of improper storage conditions, and the mycotoxins produced by the fungi can pose a health hazard for tea consumers. Previous studies have reported that the levels of the mycotoxins aflatoxin and ochratoxin A in tea leaves highly exceed the maximum permitted limits. If tea contains mycotoxins, the long-term and high consumption of tea may lead to devastating health effects that would outweigh any benefits.
Although the beneficial health effects of tea consumption have been widely investigated in the past decades, surprisingly, its safety has not attracted much attention. To the best of our knowledge, no study has reported how storage conditions, such as time, temperature, and relative humidity, affect catechin levels, fungal contamination, and mycotoxin abundance in different tea leaves. Since high contamination levels are usually found in the pre-purchase stages, we can assume that the contamination worsens in the post-purchase stages due to the opening of the packages, prolonged storage time, and the effects of the external environment. However, no one has ever recommended tea manufacturers and consumers optimal storage conditions and brewing methods to reduce fungal and mycotoxin contamination while maximizing catechin levels. If storage conditions and packaging are defective, fungi can rapidly grow and produce mycotoxins, leading to health risks.
To address these issues, we propose to investigate how temperature, time, and relative humidity of storage conditions affect the levels of catechins, fungal contamination, and mycotoxin residues in tea leaves in the post-purchase stage. Response surface methodology will be used to identify the high-risk factors for tea contamination and find the best storage conditions to reduce fungal and mycotoxin levels while maintaining catechin contents. We also aim to develop a brewing method that preserves catechin levels and reduces the health hazards posed by mycotoxin residues. The results of this study will provide tangible benefits not only to tea lovers and the general public but also to tea producers, restaurants, and the whole tea industry chain.
Project Reference No.: UGC/FDS16/M14/23
Project Title: Effects of microplastic ingestion on the growth of marine microplankton grazers and their grazing control on algal blooms
Principal Investigator: Dr TANG Chi-hung (HKMU)
Abstract
Plastic pollution has been an emerging issue worldwide. Microplastic particles, with the size < 5000 micrometers (μm), are found to be ubiquitous in both the terrestrial and aquatic environments. While ingestion of microplastics has been documented in large animals (e.g., mammals, fish, turtles, and seabirds), ingestion of plastic particles by a group of more important organisms in the aquatic ecosystems has been rarely studied. Microzooplankton, in the size range of 20-200 μm, consist of mainly ciliates and heterotrophic dinoflagellates. These microscopic organisms are the major consumers of primary production in the oceans and thus are believed to be able to affect algal blooms through a top-down control mechanism. However, the actual effects of ingestion of small microplastic particles (< 200 μm) on microzooplankton’s biology and their ecological roles in removing algal prey from the waters have been largely unknown. This project aims to test the hypothesis that ingestion of microplastics will cause adverse effects on the growth and grazing of microplankton grazers and will impair their role as a biological control of algal blooms (red tides). It is one of the few that investigates the effects of microplastic pollution on protozoan grazers since most previous studies have ignored this group of organisms. It is also one of the few studies that examines the ingestion of weathered microplastic beads by the grazers to reflect the realistic conditions of microplastics in the natural environments. To the best knowledge, this research is the first study that explores the potential linkage between microplastic pollution and algal blooms due to reduced microzooplankton grazing on algal prey.
This research will bring positive impacts to the environments by making the new knowledge available to the government, academia, and general public through means of scientific publications, media exposure, workshops, and face-to-face presentations. It will potentially strengthen the public’ initiative in reducing microplastic pollution and participate in activities to tackle the pollution problems. In the long run, better management plans and administrative policies could be established by the authorities to effectively reduce the emission of plastic waste to the environments and to improve the conservation and protection strategies of the environments.
Project Reference No.: UGC/FDS14/H15/23
Project Title: An exploration of civic resilience of Hong Kong’s citizens and society in times of political change
Principal Investigator: Dr TANG Gary Kin-yat (HSUHK)
Abstract
This project aims to explore the changes in social values and civic norms following the significant shift in the political institution in 2020. It seeks to understand how Hong Kong citizens and society have worked to maintain and build civic resilience in this new environment. The data collection methods for this project include population surveys and in-depth interviews. The results from the population surveys will be compared with those conducted before 2020 to examine the persistence of civic norms. In-depth interviews will be conducted with actors in civil society to observe how they perceive social and political changes. The project has potential for both immediate and long-term impact. The findings can inform policy-making and reconciliation efforts in Hong Kong and contribute to youth studies and youth organizations.
Project Reference No.: UGC/FDS24/H15/23
Project Title: Can a Transgender Storytelling (TGST) Workshop Change Students’ Empathetical Feelings and Attitudes about Transgender Individuals?: A Triangulation Approach
Principal Investigator: Dr TAO Kimberly Wei-yi (PolyU SPEED)
Abstract
This project explores the role of storytelling in sexuality education in changing students’ empathetical feelings and attitudes about transgender individuals and gender diversities topics. Human stories play a significant role in aiding individuals in better comprehending different individuals’ lives and establishing a broader understanding of individuals whose experiences and mindsets are different from them. Through analyzing how a transgender storytelling (TGST) workshop can change students’ empathetical feelings and attitudes about transgender individuals, this project aims to demonstrate the role of storytelling in promoting sexuality education and fostering students’ capacity to comprehend individuals who have different gender expressions, identities, and experiences. This can cultivate students’ ability to embrace gender diversities and face gender possibilities, contributing to the building of “civic imaginations” among students, a quality of understanding and building empathy for sufferings of distant others placed in different social contexts (Nussbaum, 1997). “Empathy” is the term for feeling feelings that are thought to be shared by the character(s) (“feeling with”) (Busselle & Bilandzic, 2009; Mar et al., 2011). The building of it can help students better understand and respect experiences and thoughts that are different from them and promote equality in the society. In Hong Kong, transgender individuals are still facing different sorts of discriminations and rejections in the society. The reasons behind this are mainly due to the limited understanding and misconceptions about transgender individuals in society. In Hong Kong, schools normally adopt a gender non-inclusive approach in sexuality education for discussion of topics related to non-cisnormativity or non-heteronormativity is normally deemed as controversial and challenging.
To fill in this gap, this project proposes an innovative pedagogy that involves human storytelling in Hong Kong sexuality education. By including gender and sexual minorities to tell their own stories to our students in TGST workshop, this human storytelling approach is believed to be able to transgress the conventional gender and categorical boundaries and to strengthen students’ understanding of gender diversities and possibilities. A 5-hour TGST (including a 2-hour transgender talk and a 3-hour Human Library that allows other students to interact with transgender individuals) will be held during the project period. To demonstrate the importance of storytelling in this TGST workshop, this project adopts a triangulation approach that involves quantitative study and qualitative studies (analyzing around 600 sets of questionnaires on students’ changes of empathy level before and after attending a transgender workshop, conducting discourse analysis of around 600 pieces of students’ pre- and post-workshop written reflections on (trans)gender-related topics and organizing focus groups (around 120 students) to find out students’ reasoning behind the word choices in their pre – and post- workshop written reflections and the influence of the workshop on helping them better understand gender diversities topics). The findings contribute to the design of sexuality education pedagogy and develop an online corpus to systematically document the changing language uses when discussing (trans)gender topics over time.
Project Reference No.: UGC/FDS16/E07/23
Project Title: Developing an automatic shading control approach for improving visual comfort in urban areas
Principal Investigator: Dr TSANG Kin-wai (HKMU)
Abstract
Automatic shading control system is one of the most important daylighting devices to prevent occupants from experiencing excessive glare and solar heat gain. However, this system is usually overridden by manual operation due to unsatisfactory performances. Glare is one of the major factors affecting the occupants switching on and off the shading device. Even so, glare is rarely measured or as an input parameter for shading control system. Hence, the system performance cannot satisfy the visual comfort of indoor occupants. Other problems like frequently switching, excessive glare entering the indoor space and energy wastage are recurrently reported. An ineffective operation of shading devices does not only expose indoor occupants to excessive glare and damaging their visual systems but also causing the failure of daylight-linked lighting control system.
Hong Kong faces more challenges in designing an efficacious shading system to prevent glares. Hong Kong is one of the densest urban cities over the world. Vast of high-rise buildings are constructed closely together. New constructions are usually equipped with curtain walls, which induces unexpected glare even under sunshading conditions. Consequently, the opportunities of automatic shading device malfunctions increased. A poorly designed and commissioned automatic shading control cannot prevent excessive glare entering the occupied zones especially in the presence of reflected sunlight via the external obstructions. It increases the false operations of daylight-link lighting control system in two aspects. First, if the shading device switched on when there is no excessive glare, it reduces the amount of natural light received by the buildings and hence, increases the lighting energy usage. Second, the shading device modifies the direction of incident sunlight which may create another source of glare and changes the illuminance distribution. If the redirected light hits the photosensor, overdimming occurs and the illuminance level over the working plan drops below the designed lighting level. In urban cities, automatic shading control should provide accurate control action under both direct and reflected sunlights.
In this study, design guidelines covering sensor location, training, testing and commissioning procedure will be developed for shading control system in urban areas. Scale-model measurement and computer simulation will be employed in this study. The control signals of shading devices from different control algorithms will be measured, simulated and analyzed. The performances of traditional, simulation-assisted control and deep-learning approaches will be evaluated. Artificial neural network shading control strategies will be developed based on high dynamic range image of sky and other climatic parameters. The performances of existing open-loop and new deep-learning shading control systems and their implications to the daylight-linked control lighting system & building energy performance will be determined via simulation studies of a generic buildings. The findings will be generalized to simple rules for developing a design guidelines. The proposed design guidelines will be focused on selection and location of sensor, choosing of climatic parameters for monitoring, calibration conditions, testing and commissioning procedures. This study will specifically address the shading control design issues related to heavily obstructed urban areas.
The findings would be useful to architects, building engineers in building layout, shading and lighting system designs, energy prediction and daylighting designs. It can also enable facility management professionals to effectively operate their properties and reduce operational costs.
Project Reference No.: UGC/FDS14/B08/23
Project Title: Leveraging customer co-creation platform for firm innovative performance
Principal Investigator: Dr TSE Fiona Sin-yan (HSUHK)
Abstract
In this digital era, companies are increasingly adopting digital platforms as a strategic tool for stimulating product innovations and managing customer relationships. Notably, businesses are eager to build digital customer platforms that leverage the power of collective wisdom (through customer co-creation) to develop and market new products. However, the exact link of customer co-creation to innovation performance has remained unclear (Chang & Taylor, 2016; Fang, 2008). And recent research has reported intriguing—and contrasting—findings on the business implications of digital platforms (Cenamor, Parida, & Wincent, 2019; Tse et al., 2023).
This proposal was partially inspired by the PI’s forthcoming article in European Journal of Marketing (Tse et al., 2023), which documents the divergent effects of digital customer platforms on customer perceptions of corporate brand image. In the proposed research, we highlight the conditions under which digital customer platforms would increase customer contributions to corporate innovation performance. We aim to add to the emerging literature on digital platforms and customer co-creation by delineating the mechanism through which customer platforms bolster innovation performance: namely, by facilitating competence ambidexterity (i.e., the capability to achieve both competence exploration and competence exploitation simultaneously).
Drawing on the resource-based view and contingency theory, we propose that customer platforms are conducive to competence ambidexterity, which is a key driver of innovation performance, as revealed in previous literature (Rothaermel & Deeds, 2004). To illuminate the link between customer platforms and competence ambidexterity, we further identify the moderating influences of two types of contingencies: (1) internal contingencies (customer assets breadth and interfuctional coordination) that concern knowledge acquisition and integration and (2) external contingencies (market uncertainty and competitor imitation) that reflect market dynamism.
The research context will be a group of Mainland Chinese manufacturers which face intense competition and are trying to explore new competencies as well as exploiting existing ones. Data collection will involve a large-scale survey of Mainland Chinese manufacturers from different industries with a multi-informant design. Specifically, the survey instrument will be completed by two informants (a director and a manager) from the same firm. Data analysis will involve pertinent multivariate techniques. Specifically, the proposed model (and hypotheses) will be examined with structural equation modeling and moderated regression analyses.
In summary, the prevailing evidence suggests that not all companies can benefit from vaunted customer co-creation activities, which require substantial investments in building customer platforms. How to make better use of digital customer platforms to improve innovative outcomes and business performance is therefore a topic of interest to both researchers and practitioners. Integrating insights from the resource-based view and contingency theory, we argue that companies must balance competence exploitation and exploration to achieve ambidexterity in order to translate the benefits of customer platforms into innovative outcomes. Moreover, we predict that both internal (customer assets breadth, interfunctional coordination) and external (market uncertainty, competitor imitation) contingencies will function as boundary conditions to moderate the influence of customer platforms. As we envision, this research has the potential to advance academic inquiry and improve business practices.
Project Reference No.: UGC/FDS16/E01/23
Project Title: Gunn diode-based Terahertz emitter using Topological Insulators
Principal Investigator: Prof VELLAISAMY Arul Lenus Roy (HKMU)
Abstract
Terahertz (THz) technology is proving its importance in medicine, materials, communication and security applications. Non-ionising characteristics and see-through property of the terahertz radiation gains attraction for its use in daily life. Terahertz frequency region lies between 0.1-10 THz, and operating an electronic device in this range is one of the most complex tasks in microelectronic industry. Conventionally, techniques like optical rectification and photoconduction are widely used for the generation of broadband terahertz radiations (0.1-10 THz), however, the use of femtosecond laser in these systems limits their use for commercial applications. On the contrary, though solid state electronic devices such as resonant tunnelling diode, Gunn diode and quantum cascade lasers exist, their operations under terahertz frequencies are limited due to the fact that these devices are mostly based on Si, GaAs and InP, which requires complex device fabrication steps and encounter operational instabilities or breakdowns. Hence, developing an efficient design of semiconductor materials for realising a new class of terahertz emitters is an important goal among materials researchers. In this proposal, we aim at developing a low-cost terahertz emitter source using topological insulator based materials.
With this achievement in developing efficient terahertz emitter diode device, the field of terahertz technology for medical diagnosis, security surveillance and high data rate communication can be further accelerated at low cost with high efficiency due to the fact that diodes can be easily embedded in any electronic circuits. Moreover, since Hong Kong is a major transport hub with high population density, a compact diode-based terahertz technology will benefit customs and other security organisations in security screening. Our diode-based terahertz technology will put Hong Kong as a pioneer in 6G communication equipment and hub for non-destructive testing technology.
Project Reference No.: UGC/FDS16/E16/23
Project Title: Protein-ligand Binding Affinity Prediction Based on Grid Representation and Deep Learning: Paving the Way to Efficient Structure-based Drug Design
Principal Investigator: Dr WANG Dan (HKMU)
Abstract
Background: From a molecular perspective, the formation and progression of many severe diseases, such as cancers, are closely related to specific proteins in our body. Using small-molecule drugs (ligands) to target these proteins and modulate their functions is a broadly-applied therapeutic strategy. A ligand often binds to its target protein, and the strength of such binding, termed ‘binding affinity’, crucially determines the therapeutic effects. Efficiently predicting the binding affinities of protein-ligand pairs has therefore become a key challenge in computational drug design and its branches.
Project objectives: With the rapid growth of structure-determination techniques and the recent advances in deep learning, current binding-affinity-prediction works are seeking to decode the binding affinities from given protein-ligand complex structures in a data-driven manner. However, problems like inefficient feature representations, lacking variety of learning architectures and insufficient practical applications have hindered their further development. To rise to these challenges and work out adequate strategies for handling them, we initiate this project.
Main strategies: (1) Molecular grids are the de facto representations of protein-ligand complex structures in deep-learning studies. But conventional grid representations are commonly biased and inefficient because of involving many geometric rules in the generation process. A straightforward way to improvement is defining key interactions by critical atomic information and filling them in the molecular grids. In this project, we will develop less rule-based but more efficient feature representation models, with key molecular interactions and atomic conflicts taken into account. (2) Convolutional neural networks (CNNs) have started to play a role in binding-affinity predictions. However, mostly the classic architectures were used, producing insufficient prediction accuracy and low interpretability. To mitigate this problem, we will orient the learning architectures to affinity-prediction tasks by incorporating appropriate physicochemical concepts and processes, which will prospectively lead to improved predictions. (3) Deep-learning models often induce large time consumption. In this regard, we will employ parallel and GPU computations to accelerate the predictions. To further enhance the practical applications of such deep-learning models, we will combine them with existing docking tools to perform efficient rescoring tasks. Additionally, this project will produce a web-application that can provide binding-affinity-prediction services in multiple levels, thereby maximizing the value of proposed models.
Impact of research: By successfully tackling aforementioned challenges, the proposed project will benefit tasks like virtual screening, lead optimization and reuse of drugs, thereby having a positive impact on structure-based drug design. Meanwhile, it will output useful publications and applications that will facilitate new affinity-prediction works. Last but not least, it will provide teachers and students with sufficient experience in programing, data manipulation and interpersonal communications.
Project Reference No.: UGC/FDS16/E17/23
Project Title: Knowledge Alignment Framework Based on Contrastive Learning in Stock Prediction and other Emotion-Related Natural Language Processing Tasks
Principal Investigator: Prof WANG Fu-lee (HKMU)
Abstract
We can extract useful information from text generated in various ways. Timely and accurate detection of the sentiments and opinions embedded in user-generated text can greatly benefit stakeholders, such as government agencies and service providers. Unsupervised sentence representation learning, which is an important information retrieval technique, aims to derive meaningful sentence embeddings from textual input that has not been tagged with labels. The learned sentence representations have a profound impact on downstream tasks, such as sentence classification, inference, and matching, and have been extensively investigated in the literature.
In various NLP tasks, researchers exploit sentence representations from bidirectional encoder representations from transformers (BERT)-like pretrained language models such as BERT and LBERT. These models are trained over a large corpus by encoding semantic continuity in the general domain. They follow the distribution hypothesis that contextual sentences have similar representations and that words occurring in the same context have similar meanings even if they have different semantics. This property may limit the improvement achievable by BERT models in domain-specific tasks in which domain knowledge is prominent in the decision-making process. For example, ‘happy’ and ‘sad’ are considered similar in BERT-like representations, as these terms tend to have similar distributions in a large corpus. A classifier cannot easily identify the emotion of a sentence with such representations in tasks in the emotion domain (i.e., emotion-related tasks such as sentence emotion classification). Therefore, BERT-like representations are not naturally suitable for domain-specific tasks, such as spam and hate speech detection, financial sentiment analysis, and opinion mining. The state-of-the-art methods attempt to overcome this limitation by incorporating domain knowledge into domain-specific tasks via supervised learning, for example, by integrating emotional lexicons (which are lists of words and the emotions associated with them) that distinguish the semantics of words such as ‘happy’ and ‘sad’ with pretrained BERT-like representations for emotion-related tasks. However, the representations obtained by supervised learning are not transferable across tasks within the same domain. Moreover, such supervised methods require numerous labelled samples, the acquisition of which is expensive and involves considerable manual effort by domain experts.
To address the abovementioned problems, we will establish an unsupervised method based on contrastive learning to align the general semantic representation with domain-specific knowledge without using labelled samples. In the proposed project, we will mainly select tasks in the emotion domain (e.g., emotion-related tasks) because research communities have acknowledged emotions to be as crucial as semantics for language representation. Contrastive learning aims to decrease and increase the distance between similar and dissimilar instances, respectively. We will redefine similar instances and design a framework for knowledge alignment to introduce various knowledge sources by substituting embedding models. By leveraging domain knowledge, this method will have the potential to address the problem that the refinement of BERT representations in conventional contrastive learning is more semantic-oriented than domain-dependent. Further, we will incorporate a momentum update strategy that can control the proportion of general-semantic and domain-specific features in the learned representations. In addition, the optimal augmentation strategy and proportions of general-semantic and domain-specific features will be identified for each domain. Subsequently, the representations encoding diverse knowledge will be evaluated in downstream text mining tasks. In the proposed project, we will use financial market prediction to showcase and test the learned representations in stock prediction incorporating sentiment analysis. We will aim to investigate whether the semantic meaning or sentiment encoded in financial articles and news influences stock movement.
Project Reference No.: UGC/FDS16/E02/23
Project Title: Automatic and Accurate Evaluation of Placental Abnormalities via Multi-modal, Multi-task and Federated Semi-supervised Learning
Principal Investigator: Dr WANG Weiming (HKMU)
Abstract
The placenta is an essential organ that directly affects fetal growth and mother health. It plays a significant role throughout pregnancy because it allows the exchange of nutrients, gases and waste between fetus and mother. Abnormal placentas are the most culpable causes of many pregnancy complications, such as fetal growth restriction, pre-eclampsia and fetal death. Thus, accurate evaluation of placental abnormalities is very important in prenatal diagnosis so that appropriate treatments can be planned to reduce fetal and maternal morbidity and mortality.
In clinical practice, placental maturity grading and detection of suspected placenta accreta spectrum (PAS) are typically performed for pregnant women to evaluate placental abnormalities with ultrasonography (US) and Magnetic resonance imaging (MRI). However, due to the complexities of placental morphologies and various imaging artefacts, placental abnormality evaluation is mainly based on visual observation of US and MRI images by professional radiologists, which is tedious and time-consuming, as well as subject to large inter- and intra-observer variations. As a result, a lot of automatic algorithms have been developed to facilitate prenatal examination based on machine learning and deep learning. Unfortunately, there are still many challenging issues that hinder clinical applications of these algorithms. First, existing methods for placental maturity grading and PAS detection mainly utilize the information from single modality without considering complementary information from multiple modalities, and thus suffer from unsatisfactory accuracy. Second, there is limited annotated data to train the network for PAS classification because it is very difficult to identify PAS from US and MRI images even by experienced radiologists due to unobvious clinical symptoms. Third, domain shift is a common issue in medical images obtained from different data sources, and CNN models that are well trained on one domain may perform poorly on new domains due to the interference of domain shift. Fourth, the performance of neural networks is usually restricted by limited number of training images, while a large number of datasets from distributed hospitals cannot be shared to boost the model performance due to data privacy.
In this project, we shall conduct comprehensive research work to tackle the aforementioned challenges. First, we shall propose a multi-modal and multi-task learning network for placental maturity grading. Multi-modal information from both image and non-image data will be fully exploited to increase the grading accuracy of placental maturity. In addition, placental segmentation is utilized as an auxiliary task to enhance placental grading. Second, we shall propose a multi-modal semi-supervised learning network for PAS classification by combining complementary information from US and MRI data. Numerous unlabeled images will be incorporated into the training process to compensate limited labeled data in order to boost the model performance. Third, we shall propose an uncertainty-aware adversarial learning network to alleviate the interference of domain shift between different data sources. The generalization capability of CNN models is improved by suppressing unreliable information (high uncertainty) during domain adaption. Fourth, we shall propose a Fourier-based federated semi-supervised learning (FSSL) network to deal with data privacy among distributed hospitals for collaborative training. The relationship between different unlabeled data is utilized to enhance multi-hospital collaborative training. Moreover, Fourier-based synthesis is adopted to explore the information among unlabeled data to improve unsupervised learning.
The deliverables of this project are a series of advanced deep learning algorithms to facilitate placental maturity grading and PAS classification, as well as novel networks for multi-source domain adaption and multi-hospital collaborative training. Furthermore, we shall integrate the proposed algorithms into a user-friendly computer-aided system, which could potentially serve as an automatic, accurate and objective tool to assist prenatal diagnosis.
Project Reference No.: UGC/FDS14/E08/23
Project Title: Generating novel customer needs for new product development
Principal Investigator: Dr WANG Yue (HSUHK)
Abstract
Identifying new customer needs is critical for developing innovative products and starting new businesses. For instance, when Ford recognized baby boomers' desire for inexpensive sporty cars in the 1960s, they developed the Mustang, selling over 400,000 units in the first year. Companies want to identify emerging needs early for profit opportunities, but traditional market research is slow and costly. Recent advances in AI like ChatGPT present new chances to efficiently generate novel customer needs for product development. We propose incorporating domain knowledge into large language models, then using the enhanced models and deep learning to generate fresh customer need texts. This methodology can: 1) understand relevant domain knowledge; 2) rapidly create novel needs beyond current ones; and 3) require fewer company resources than traditional techniques. Our approach could transform product ideation and development by enabling designers to derive more creative ideas from AI-generated insights. The results will advance product design research by demonstrating the power of generative AI for innovation. This project has practical implications for mining insights for new products and businesses.
Project Reference No.: UGC/FDS16/H09/23
Project Title: The Casual Official and the Tang Minister: A Study of the Reception of Yuan Jie and His Literature
Principal Investigator: Dr WONG Chi-hung (HKMU)
Abstract
Yuan Jie 元結 (719–772) was both an important writer and a capable local official during and after the An Lushan Rebellion (755–763). In the history of Chinese literature written since the early years of Republican China, Yuan Jie has gained the attention of literary historians who promote realism, and he is generally regarded an advocate of xin yuefu 新樂府 (new music bureau poetry) in the Tang (618–907) dynasty. Literary historians believe that mid-Tang poetry belongs to the style and theme of literary function, which is closely related to the reflection of common people’s tragic life and the expression of deep sympathy for people during times of chaos. Through the efforts of Du Fu 杜甫 (712–770), Yuan Jie and other xin yuefu poets, literary historians believe that the “xin yuefu movement” and its associated Chinese realism was an influential trend in Tang poetry and made a major contribution of Chinese literature.
Scholars in recent years have conducted research on Yuan Jie’s xin yuefu poetry, his description of unique landscapes, and his contribution to guwen 古文 (ancient prose) and archaist poetry in the Tang dynasty to delve into the above views. To extend the present research, this project intends to study the comments on, impressions of, and reception to, Yuan Jie from literati and scholars in the Imperial period. In their eyes Yuan Jie was not merely a “realist” poet and a good official for the emperor and the people, but also a recluse, who was not inclined to power and was fond of secluded and extraordinary scenery, as well as a herald of the landscape prose of Liu Zongyuan 柳宗元 (773–819) and a representative figure of Tang landscape literature. By exploring the literary status, influence and understanding of Yuan Jie and his writings in past dynasties, it would be beneficial to reposition the writer in the long history of Chinese literature, and reveal his literary contribution and legacy.
The aim of this project is to study Yuan Jie from two perspectives: (1) investigate how the writer was discussed and how he was categorised in literary analyses to reveal the author’s image and influence in the past dynasties, and (2) study how Yuan Jie’s landscape writing has been discussed and how he was rediscovered as a pioneer of landscape literature. These perspectives will allow us to discuss his literary contribution more fully, including the views of xin yuefu and realism, and to further understand his achievements in landscape literature and his importance to later generations of literati.
Project Reference No.: UGC/FDS16/M17/23
Project Title: Growth stimulating effect and related mechanisms of concentrated Chinese medicine granules of Astragalus membranaceus and Glycyrrhiza uralensis in Nile tilapia and mud carp under overcrowding stress
Principal Investigator: Dr WONG Emily Sze-wan (HKMU)
Abstract
Freshwater aquaculture is an important food supplying industry in Guangdong-Hong Kong-Macao Greater Bay Area and worldwide. In order to achieve a higher yield, intensive cultivation is often adopted in the aquaculture industry. Under overcrowding stress, fish usually have weakened immune response. This greatly affects their growth performance and makes them to be more susceptible to different fatal infectious diseases. In order to stimulate the fish growth and reduce the fish loss, disinfectants such as formalin, and various antibiotics are often overused. This will lead to irreversible damage to the environmental ecosystems and development of super bacteria. In addition, the residue of the disinfectants and antibiotics in fish will threaten human health. Hence, there is an urge need to find a green fish supplement to reduce the use of disinfectants and antibiotics in freshwater cultivation.
Tonic Chinese medicine is a potential candidate due to its safety, low toxicity, limited impact on the environment and proven immunomodulatory effects. Recent research on the efficacy and mechanistic pathway of Chinese medicine on aquaculture is limited. Among different Chinese medicines, Astragalus membranaceus (AM) and Glycyrrhiza uralensis (GU), belonging to the “tonic” therapeutic group of Chinese medicine, appear to be the potential candidates for coping with the side effects caused by the overcrowding conditions on aquaculture. This may be due to their significant growth stimulating and immunomodulatory effects on aquaculture. This hypothesis also concurs with our preliminary findings.
Though some studies have shown that Chinese medicines have growth stimulating effect on aquaculture, there is lack of systematic study and common ground for comparison. The experimental conditions varied among different studies. Even if the same Chinese medicine was investigated, different preparation forms, such as the crude herb, the herbal extract or the major active compounds, of non-comparable dosage ranges, have been applied. Various fish models were investigated. More important, the underlying physiological and molecular mechanisms involved have not been studied thoroughly.
Without proper boiling, the therapeutic effects of Chinese medicine will be greatly affected and the related procedures are time consuming. At the same time, the price of extracted herbal active compounds is expensive. This greatly increases the feeding cost. Therefore, regardless of the efficacy of Chinese medicines, their application in the fish sector is very limited. Concentrated Chinese medicine granules (CCMG) are the ready-to-use form of Chinese medicine with proven efficacy and reasonable price. These unique characteristics flavor the use of CCMG in the fish industry. However, as far as we know, no studies of CCMG on edible fish have been done.
The aim of this project is to investigate the role of AM and GU, as a green fishery supplement, by analyzing the effect and mechanism involved in stimulating the growing, hormonal, metabolic, enzymatic, immunomodulatory and antibacterial responses in edible freshwater fish, Nile tilapia and mud carp, under normal and overcrowding conditions. The study will also compare the effect of different preparative forms including CCMG and the most abundant active compound, of the two herbal medicines. More important, the underlying mechanistic pathways will be explored using the advanced proteomic and transcriptomic approaches. The findings generated from the current study will promote the translation of Chinese medicines from human system to fish aquaculture.
Project Reference No.: UGC/FDS14/E09/23
Project Title: Ensuring the operational resilience of shipping by integrating vessel slot allocation and container supply planning with uncertain demand for yield management and order fulfilment in the global supply chain
Principal Investigator: Dr WONG Eugene Yin-cheung (HSUHK)
Abstract
The global supply chain has been increasingly disrupted by cargo demand fluctuations, trade policy restrictions, pandemic measures, price competitions, factory order changes, and other economic factors, leading to frequent mismatch of the shipment demand against vessel and container supplies. This disruption has evolved from an industrial problem to a threat to economic stability. Ship liners strive to ensure smooth cargo delivery by seeking more agile management of the vessel slots, shipping network, container supply, repositioning strategies, and slot exchange to lower operation costs and improve shipping yield and profitability. Slot allocation is typically based on the demand prospects from the traffic control regions in which cargoes are directly loaded or transhipped at the port calls of the service routes served by ship liners. In our previously completed research project, a novel three-echelon collaborative slot allocation planning model for dynamics among local, regional, and global scales, with containers loading and discharge at multiple vessels and service loops, was developed. The simulation results demonstrated that this model could assist trade and traffic planners in maximising their slot usage.
Notably, vessel slot allocation must be optimised considering the demand dynamics of the traffic loading regions. With the demand and supply are frequently mismatched, it is necessary to realise consistent and synchronised planning considering the container supply, empty container repositioning, and demand prospects of the cargo control party. Optimal slot allocation for laden and empty containers on own and alliance partner vessel spaces is vital. The existing slot allocation planning strategies based on container supply planning methods, such as container repositioning, new container purchasing, lease-in, one-way free use, disposal control, and slot exchange, are implemented over separate entities in a sequential manner. The slot allocation operations are simulated without considering and integrating the changes in all the container supplies as well as cargo demand from cargo control parties. These deficiencies in slot allocation operations are expected to result in container supply shortage, booking shortfall, low vessel utilisation, and order fulfilment failure. These problems have been reviewed, and ideas have been exchanged with ship liners, shippers and scholars.
In the proposed research, we will establish a novel integrated framework for slot allocation with comprehensive consideration of the container supply planning and slot exchange under cargo demand uncertainty to improve the container usage, vessel utilisation, shipment order fulfilment, slot exchange, and global supply chain stability. A recurrent neural network-based algorithm will be developed for container demand prediction for supply planning, and mixed integer programming will be applied for adaptive slot allocation optimisation for laden and empty containers. A novel column-generation-based algorithm will be further developed for slot allocation, integrating the cargo demand dynamics and container supplies along the planning time horizon. The proposed model will assist trade traffic planners and equipment flow managers in ensuring optimal and integrated slot allocation and implementing the container supply plan to maximise the equipment utilisation and yield. This model can serve as a best practice for minimising the mismatch between the cargo demand and equipment supply and stabilising the global supply chain. Through global slot allocation planning with advanced simulations, Hong Kong, which provides high–value-added maritime services as a global logistics hub, can promote the sustainable growth of the global supply chain. In particular, vessel allocation tools can ensure that the cargo dimensions and weight satisfy the cargo payload capacity and verified gross mass requirements, thereby preventing vessels from sailing with excessive weight and limiting the fuel usage and greenhouse gases emissions. The project deliverables on vessel allocation planning will be incorporated into teaching modules to enhance students’ knowledge regarding complex maritime operations and advanced simulation tools.
Project Reference No.: UGC/FDS25/E05/23
Project Title: A Study of Temperature and Relative Humidity Effect on Strength Development of Low Carbon Concrete
Principal Investigator: Dr WONG Ho-fai (THEi)
Abstract
Hong Kong is a modern city with enormous amount of construction work for building and infrastructure every year. Reinforced concrete is traditionally the most commonly adopted construction materials for construction works. The quality assurance of concrete is, therefore, an essential component to guarantee the achievement of design strength and stiffness of structures. According to the Construction standard CS1, the strength of concrete is currently determined by concrete cube (100mm or 150mm), which are cast and cured in curing tank with a temperature maintained at 27± 3°C. In construction site, however, it is obvious that the curing temperature and surrounding relative humidity (RH) are largely different from those for standard curing. According to the record of Hong Kong observatory, the average ambient temperature (2021) in Hong Kong varied from 13.8°C (in winter) to 32.8°C (in summer) while the average relative humidity changed from 67% (in winter) to 83% (in summer). Many researchers have indicated that the early age strength development of concrete is largely dependent on the curing temperature and relative humidity of the environment. According to the Code of Practice for Structural Use of Concrete 2013, “formwork supporting cast insitu concrete in flexure may normally be struck when the strength of the concrete in the element is 10 N/mm2 or twice the stress to which it will be subjected, whichever is the greater…”. Without other information, the recommended minimum period striking formwork and falsework for concrete vary from 12 hours to 14 days, dependent on the type of structural element. Low curing temperature and relative humidity may induce low early concrete strength which may put removal of formwork at risk.
Currently, many prediction models, e.g. maturity model, have been proposed to predict the strength development of concrete. However, most of them are related to curing temperature only but not relative humidity. In addition, since late 1980’s, supplementary cementitious materials (SCM) like Pulverized Fuel Ash (PFA) and Ground Granulated Blastfurnace Slag (GGBS) were widely used in combination with Portland Cement (PC) to produce low carbon concrete for sustainable development. Their applications become more important. However, the strength prediction model for concrete and especially low carbon concrete under the effect of curing temperature and relative humidity are not well developed.
This research study aims to study the strength development of low carbon concrete under the coupling effect of curing temperature and relative humidity in Hong Kong and create a RH-modified maturity model which can monitor and record the real time concrete strength development at construction site. Stakeholders of the construction works can monitor the real time concrete strength development which can shorten the construction work cycles and enhance the construction safety e.g. early removal of formwork and falsework. Engineers and workers can design and install temporary works based on the concrete strength developed.
Project Reference No.: UGC/FDS14/B03/23
Project Title: Distinct effects of intrinsic and extrinsic rewards on employee creativity and routine performance: The dual role of mindfulness
Principal Investigator: Dr WONG Noel Yuen-shan (HSUHK)
Abstract
In an increasingly globalized environment, organizations have to adapt quickly to the environmental changes and development and develop creative products and business models to maintain competitiveness. Competition for talent has become one of the critical challenges faced by organizations. In the post-pandemic era, employees are more concerned about work-life balance. The traditional carrot-and-stick approach to motivation is inadequate in attracting talent, therefore, organizations have to provide an all-rounded incentive scheme to attract and retain talents. Furthermore, employees’ individual differences should be considered in reward system.
In recent years, mindfulness has been widely promoted in different sectors (e.g., education and workplace), considering the empirical support of its health-enhancing role. By integrating the motivation and cognitive processing perspectives, we propose that mindfulness has a dual role in moderating the reward-motivation-performance process in employee creativity. First, mindfulness will moderate the reward-motivation link. Mindful individuals are aware of their present experience and likely to engage in self-directed and self-determined behaviors which may affect their responses to different types of rewards. Second, mindfulness will moderate the motivation-performance link. Motivation can be translated into different types of performance for people with different levels of mindfulness. Mindful employees have a wide attentional span and cognitive flexibility, which are important for generating new ideas. On the other hand, less-mindful employees are more likely to translate their motivation into less creative and routine work performances.
The present project has several theoretical and practical contributions. First, it contributes to the literature on motivation and creativity by using mindfulness to explain the mixed findings in the related fields. Second, it provides implications for organizations and managers on the design of reward programs that fit with the needs of employees and strategies of the organizations. Lastly, it explores the potential downside of employee mindfulness.
Project Reference No.: UGC/FDS16/P02/23
Project Title: Metal-Catalyzed C–H Bond Functionalization of (Hetero)arenes by using alterable P,O or P,N hemilabile benzimidazolyl phosphine ligands
Principal Investigator: Dr WONG Shun-man (HKMU)
Abstract
Bi(hetero)aryl and heterocyclic derivatives are common structural motifs present in the core structures of natural products, biological, and pharmaceutical materials, in which are crucial and useful in human being. Synthesis of these useful compounds is important all the time in pharmaceutical industries. Hence, development of mild and simple synthetic methodologies has become a great attention from scientist in order to produce wide range of biaryl derivatives. However, in different traditional transition metal-catalyzed reactions, such as Hiyama, Kumada, Negishi, Stille and Suzuki-Miyaura coupling reactions, preparation of pre-functionalized (hetero)arene coupling partners is needed to connect two (hetero)arenes. In the recent decade, different bi(hetero)aryl derivatives have been made by employing remarkable C–H functionalization as synthetic strategy to overcome pre-functionalization steps due to the simplicity.
In the previous protocols, researchers developed different transition-metal-ligand catalytic system to catalyze C–H functionalization. However, to the best of our knowledge, there is no example that by using a easily assessable P,O and P,N hemilabile benzimidazolyl phosphines in the metal-catalyzed C–H functionalization catalytic system. This family of benzimidazolyl phosphine ligands can be easily assessed by a modular one-pot assembly approach across three readily or commercially available starting materials, including (1) benzimidazole scaffold, (2) acid chlorides/sulfonyl chlorides, and (3) chlorophosphines, to achieve a high level of diversity. Notably, this ligand family has a unique coordination feature that containing both soft (phosphorus atom) and hard (oxygen and nitrogen atoms) donor atoms. During the coordination of the metal center, the hard donor provides a weak coordination to protect the metal center during the harsh reaction conditions and provides a vacant site for oxidative addition of the substrate by dissociation of oxygen/nitrogen atom. It can offer an enhancement of catalyst longevity. Furthermore, this ligand family has a high tunability as it obtains tunable phosphine group and additional tunable group in the benzimidazole scaffold. Those features allow the ligand family to have a considerable potential towards different type of catalytic coupling reactions. In addition, this ligand family showed highly active towards palladium catalyzed Suzuki-Miyaura coupling and Buchwald-Hartwig amination of aryl chlorides in previous studies. Indeed, with this ligand family in hand, a wide array of ligand-structure towards specific C–H functionalization can be studied, and the general picture gained from this proposal will be highly useful and important in the area of cutting-edge C–H functionalization.
In this proposal, a family of benzimidazolyl phosphine ligands bearing unique alterable P,O or P,N-coordination will be used to investigate the potential effectiveness in C–H functionalization for synthesizing different useful structure. In addition, different transition metal (palladium and nickel) and coupling partners will be examined in order to provide wider scope of synthetic protocols.
The research findings of this project will promote the application, improvement and development of efficient and convenient methodologies for organic synthesis of biaryl products. Hence, in the related research fields, academic, education and industries will be benefited.
Project Reference No.: UGC/FDS16/M21/23
Project Title: Oxidative activity of mangrove leaf phenolics and their inhibition on growth and physiology of harmful algal species, and the underlying molecular mechanism
Principal Investigator: Dr XU Jingliang (HKMU)
Abstract
The rapid growth of microalgae known as harmful algal blooms (HABs) often results in massive and acute death of aquatic organisms. Most coastal areas of the world are adversely affected by HABs, causing severe economic losses in fisheries and threatening human health. The control of the growth and spread of algal cells after algal blooms has become an important topic of governments and research institutions in various countries. Chemical, physical and biological methods are commonly used to control HABs. Chemical methods use herbicides or metals, while physical methods use ultraviolet or ultrasonic treatment. However, these two methods usually cause secondary pollution or consume a lot of energy. In contrast, the use of biological methods has the advantages of environmental friendliness and energy saving. One such biological approach is the use of anti-algal chemicals (i.e., allelochemicals) produced by macroscopic aquatic plants. In freshwater systems, numerous studies have been conducted to explore the use of allelochemicals, especially phenolic compounds, to control the growth and reproduction of microalgae and HABs. Oxidative stress, indicated by excessive induction of reactive oxygen species (ROS) in microalgal cells, is generally considered to be the main mechanism of allelopathic action. The oxidation activity of phenolic compounds is affected by water chemistry, such as pH and metal ion concentration. However, similar studies on marine systems are relatively limited. Mangrove plants dominate tropical and subtropical coastlines. They are rich in tannins, a type of phenolic. But whether these allelochemicals can inhibit the growth of marine microalga is seldom known. Limited studies show that the frequency of HAB occurring in coastal areas with mangroves is relatively less than in those without mangroves. Our previous experiments showed that leaf aqueous extracts of six mangrove species had different effects on the growth of five HAB species, most of which were inhibitory. Although there was no apparent direct relationship between the concentration of phenolics in mangrove leaf extracts and their growth-inhibitory effects on microalgae, the effects appeared to vary between different types of phenolics. Different mangrove species contain various major types of phenolics, including condensed tannins (CT), gallotannins (GT), ellagitannins (ET) and low-molecular-weight polyphenols (LMWP). Various compositions of phenolics in different mangrove species may be a key factor affecting the degree of the inhibition on microalgae; changes in water chemistry (e.g., pH level and metal ion concentration) during algal blooms, as well as the presence of other non-HAB species, may also affect the inhibitory effect on the HAB species. Therefore, this study aims to determine the relationship between the oxidative activities of different types of phenolics in the leaf extracts of mangrove species and their growth inhibitory effects on harmful algae. Six mangrove species with various types of dominant phenolics will be selected. The total phenolics (TPs), CTs, GTs, ETs and LMWPs in their senescent leaves will be extracted and quantified. Their oxidative activity will also be determined. The most effective phenolics of each mangrove species will be identified by comparing their growth inhibitory rates on five HAB species. The growth and physiological effects of the identified phenolics on five HAB and two non-HAB species will then be compared, including antioxidant response, photosynthesis and nutrient uptake efficiency, cell membrane integrity and programmed cell death. The effects of the identified phenolics on the growth and physiology of harmful algae will be assessed under varying exposure conditions including changes in pH, ionic concentration, co-cultivation of microalgae, and dosage frequency. Eventually, the underlying mechanism will be elucidated by proteomic analysis. It is expected that the results of this study would provide an alternative control method for HABs by using allelochemicals in mangrove plants. Another aspect of the ecological function of mangroves will be further verified, which will enhance the conservation value of mangrove ecosystem.
Project Reference No.: UGC/FDS14/H09/23
Project Title: Future Museum: Review of immersive technology applications and curatorial strategy for exhibitions and art education in museums and galleries
Principal Investigator: Dr YANG Rochelle Yi-hsuan (HSUHK)
Abstract
Museums and galleries have always been significant centres for culture and art education. As museums and art centres are slowly recovering from the impact of COVID-19, they continue to face the challenges and uncertainty of operating in a post-pandemic future. This new environment may herald innovative curatorial strategies and technologies that change the way exhibition display designs are conceptualised. For them to stay relevant, museums or art organizations may have to re-evaluate how they serve the public and provide memorable learning experiences to their audiences. Knowing how to face these curatorial challenges will undoubtedly become a highly-valued expertise in the near future.
Thanks to the advances in immersive media such as borderless spatial immersion, virtual reality (VR), augmented reality (AR), mixed reality (MR), and extended reality (XR) which give exhibition institutions hope for the future. Museums are able to adapt the way they interact with the public by integrating physical and digital reality. In recent years, as technologies transformed the ways in which visitors engage and interact with display items, art museums have reached a new era of art education.
This study aims to investigate the curatorial strategy of adapting immersive technology to art exhibitions. It is particularly concerned with the strategy’s effectiveness in achieving educational and communicative goals.
Key Words: Immersive technology, museum, art education, communication, display design, interaction
Project Reference No.: UGC/FDS14/B01/23
Project Title: SOE State Audit in China
Principal Investigator: Dr YAU Belinda Ling-na (HSUHK)
Abstract
Institutional feature has been a key role in many China-related studies. The Chinese government controls a significant amount of critical resources such as banking and energy through state-owned enterprises (SOEs). In terms of monitoring SOEs, the Chinese government has emphasized the significance of state audit conducted by the National Audit Office. State audit aims not only to increase the asset value and competitiveness of SOEs, but also prevent corruption and state-asset tunneling. While understanding more about state audit shall deepen our understanding of China’s institution, there has been limited research in this area. This proposed research project aims to provide descriptive evidence of SOE state audit and examine how such government monitoring affects firm performance and corporate governance.
First, by hand-collecting state audit report data and manually categorizing issues noted by state auditor, we attempt to understand more about state auditor’s concerns, which may alter the firm behavior. Second, we will examine whether state audit report carries any information to investors. Specifically, we will focus on: (i) the 3-day cumulative abnormal return centered on the state audit report release date, (ii) the absolute magnitude of the 3-day cumulative abnormal return, (iii) the abnormal 3-day trading volume, and (iv) the abnormal 3-day stock return volatility. Third, we will consider the deterrent effect of state audit. We will investigate earnings manipulation behavior (via real activities or accrual management) surrounding the time of state audit. In addition, we will examine whether SOE mangers are penalized and face worse career prospect when state auditor identifies irregularities. Finally, we will consider the effect of state audit on the external auditor selection choice and audit fee.
Overall, this study aims to provide evidence on SOE state audit, which is a research area that promises fruitful and insightful opportunities. State audit on SOEs is one of the most important monitoring instruments by the Chinese government. It affects both firm performance and corporate governance of SOEs. Future research can be extended to political connections of state auditors or SOE managers, as well as social connections between state auditors and external auditors. Another research avenue is to study the state audit effect on listed SOEs versus non-listed SOEs, or centrally-administrated SOEs versus locally-administrated SOEs.
Project Reference No.: UGC/FDS17/H01/23
Project Title: Cardioprotective and mental health benefits of the Mediterranean-DASH Intervention for Neurodegenerative Delay (MIND) diet combined with forest bathing (FB) among adults with elevated blood cholesterol: A randomized controlled trial
Principal Investigator: Dr YAU Ka-yin (TWC)
Abstract
Introduction: Elevated low-density lipoprotein cholesterol (LDL-C) and persistent stress are proven risk factors associated with a higher risk of cardiovascular diseases (CVDs). Some studies have suggested that the Mediterranean Dietary Approaches to Stop Hypertension (DASH) Intervention for Neurodegenerative Delay (MIND) diet is associated with reductions in body weight, metabolic syndrome, and depression. Forest bathing (FB) helps to reduce blood pressure and anxiety and promotes a positive mood. Because they downregulate the proinflammatory response, the MIND diet and FB have been separately reported to potentially moderate CVD risk factors (e.g., LDL-C, blood pressure, stress) among adults. However, it is unknown whether FB enhances the effects of the MIND diet to further reduce CVD risk.
Objectives: The proposed study will investigate the combined therapeutic effect of the MIND diet and FB in reducing LDL-C and other risk factors of CVDs among Chinese adults with elevated LDL-C.
Methods: A single-blind three-arm randomised controlled trial is proposed. A total of 273 participants aged 50–75 years old with elevated LDL-C will be recruited from local community centres. Eligible participants will be randomly assigned to either the MIND plus FB (MIND-FB), MIND or routine care groups by a research assistant based on a predetermined random allocation list generated by a computer program. The MIND group will receive four nutrition sessions on four consecutive weekdays and will be instructed to adhere to the MIND diet for 12 weeks. The MIND-FB group will receive the same intervention as the MIND group but will also undertake 2 h FB sessions on four consecutive weekends plus 2 h of self-practice FB at weeks 8 and 12 in a country park. The routine care group will be instructed to perform daily activities as usual and will attend a health talk and receive pamphlets provided by the Department of Health/Hospital Authority about the protective and risk factors of CVDs. The primary outcome (change in LDL-C at 12 weeks) and the secondary outcomes (change in LDL-C at 4 weeks; change in total cholesterol, high-density lipoprotein cholesterol (HDL-C), triglycerides, glucose, systolic blood pressure (SBP), waist-to-hip ratio, body mass index, anxiety level, and emotional state at 4 and 12 weeks) will be recorded at baseline (T0) and at the designated time points at the participating community centres. The missing data will be analysed according to the intention-to-treat principle at follow-ups. The chi-square test and analysis of variance (ANOVA) will be used to assess and compare the baseline demographic, socioeconomic and lifestyle characteristics, cardiovascular medical history and medication use among the three groups. Generalised estimating equation (GEE) models will be used to examine change in the LDL-C concentration and the secondary outcomes at 4 and 12 weeks. The models will be adjusted using the random effect of the community centres. Secondary analysis will be conducted to adjust the GEE models for other potential confounders, e.g., Separate GEE models will be applied to examine dietary changes among the three groups at 4 and 12 weeks.
Conclusion: This proposed study will investigate the cardiac and mental health benefits of the MIND diet combined with FB in Chinese adults with elevated LDL-C. Our findings may provide evidence for the combined effect of the MIND diet and FB in reducing CVD-related risk factors.
Project Reference No.: UGC/FDS14/P05/23
Project Title: Nonconvex Structured Optimization: Theory and Applications
Principal Investigator: Dr YU Carisa Kwok-wai (HSUHK)
Abstract
Big-data analysis techniques have become increasingly important in many disciplines. By virtue of certain structures of the data in applications, structured optimization problems have been extensively applied in various fields such as compressive sensing, data analysis, machine learning and statistics. In recent decades, extensive successful applications have revealed that the structured optimization problems could provide an accurate representation of reality and enlighten a class of fast numerical algorithms benefiting from the exploiture of certain structures, such as the well-known Lasso for the sparse structure and Robust Principal Component Analysis (PCA) for the low rank structure. However, the structured optimization problem is usually nonsmooth and nonconvex; hence, it is of great demand but challenging to deliver fast algorithms and develop the convergence theory for the structured optimization problems.
In this project, we will address a class of nonconvex structured optimization problems, which consists of a nondifferentiable convex function and a composition of a possibly nondifferentiable convex function and a possibly nonconvex smooth function. This class of structured optimization problems is usually nonsmooth and nonconvex, and covers many optimization models in data analysis, machine learning, statistics and other areas of current interests. For this class of structured optimization problems, we will propose two fast first-order iterative algorithms by virtue of the structure of the problem and then explore their convergence theory. For this purpose, we will investigate the weak sharp minima property or error bounds property of the structured optimization problems under some assumptions on each component functions and by virtue of the structure of the problem. By using the weak sharp minima property, we will establish the convergence rates of the proposed first-order iterative algorithms. This convergence theory will cover and improve the convergence results of several popular numerical algorithms. In the application aspect, we will apply our proposed algorithms to improve the quantitative and automatic analysis of differential optical absorption spectroscopy. The successful application to differential optical absorption spectroscopy will help physicist to promote the computational optics technology and to comply with the development of optical spectrometer.
Project Reference No.: UGC/FDS24/E08/23
Project Title: CFD-based Autonomous System of Health Estimation and Maintenance Scheme for Electrical Apparatus of Power Substations in Hong Kong
Principal Investigator: Dr ZHANG Hao (PolyU SPEED)
Abstract
The power substations are essential to delivering reliable power to end customers at lower voltages. The operation of substations is constantly facing challenges from the growing demand for electricity, limited budget, high requirements for reliable electricity supply, as well as the safety of personnel. Their reliability and continued performance are the keys to power distribution. However, on June 21, 2022, a cable bridge in Yuen Long caught fire at 7:10 p.m. The fire has caused damage to three groups of high-voltage cables. Households in Yuen Long, Tuen Mun, and Tin Shui Wai were cut off. It is estimated that 160,000 households will be affected by the blackout. Although the accident is still under investigation and the cause have not yet been made public, aging of the apparatus is considered as one of the plausible causes. Aging is a common problem that many of the power substations are facing in Hong Kong. This aging problem is caused by a variety of reasons, including unfavorable hot, humid, and salty air, as well as the long-term heavy workload during the summer in Hong Kong. Meanwhile, in the traditional design, data from the fixed temperature sensors cannot illustrate the real indoor temperature distribution as there are usually some heat accumulation points. Undetected overheating in these heat accumulation points will significantly degrade the life expectancy of the electrical apparatus or even cause an unexpected failure. However, the replacement of all the aging electrical apparatus is not an option as the power supplier cannot afford the cost. For each power apparatus in a power substation, the conditions should be regularly checked, quantified, and categorized based on the understanding of the devices themselves and the degradation process. Nevertheless, this human-based time-based traditional inspection method is cost-inefficient and labour intensive, which will no longer fit the future need.
Hong Kong needs a multi-dimensional upgrade to the power substations to increase their intelligent ingredient and ensure the reliability level; this situation faced by Hong Kong serves as a strong motivation to propose this project. In the project, an integrated system for upgrading power substations will be developed to observe the operation status, monitor the remaining service life, execute the emergency responses, and make productive maintenance and repair predictions. Firstly, an experimental platform replicating a real power substation will be designed and set up in order to facilitate the constructions of modules to be developed. The experiment setup will contain a transformer, switchgears, a compartment with a ventilation system simulating a real power substation, as well as a variety of necessary sensors. Secondly, a CFD-based thermal sensing, mapping, and analysis method will be developed. With the help of CFD simulation, the heat accumulation points can be detected according to the latest inputs from a limited number of sensors, the operation condition of the electrical apparatus, and the indoor design of the station. In addition, CFD will provide assessments to optimize the installing locations of sensors, such that the minimum number of sensors can be used to reduce the upgrade costs of substations. Thirdly, by using the data obtained from the sensors and CFD simulation, the health index (HI) for the electrical apparatus will be calculated based on the reinforcement learning method. The simulation environment to facilitate reinforcement learning will be constructed based on the experimental platform; the CFD results will provide abundant data support for RL simulations; and the reward function will be adjusted according to the RL simulation results. The HI is the major criterion for the proposed system to provide life expectance, maintenance predictions, and early-stage warning of unexpectancy. Lastly, a remaining lifetime estimation algorithm as well as an autonomous maintenance suggestion scheme will be developed. By integrating all the aforementioned submodules into a complete intelligent system, it is expected that, the system can increase the reliability, achieve intelligent management, and reduce the operating cost of power substations in Hong Kong.
Project Reference No.: UGC/FDS15/B02/23
Project Title: Does Corporate Sociopolitical Activism Oriented Market Exits Benefit Corporate Brand Equity and Firm Profitability? Evidence from Corporate Exodus from Russia amidst the 2022 Russo-Ukrainian War.
Principal Investigator: Dr ZHANG Xiao (Shue Yan)
Abstract
With the rise of political consumerism, consumers and other stakeholders often assert pressures on firms to voice their stance on sociopolitical issues (Jungblut & Johnen, 2021). In response to these pressures, multinational corporations (MNCs) have been increasingly engaging in corporate social activism (CSA), which refers to a public statement or action taken by a firm to express its support for or opposition to a sociopolitical issue (Bhagwat et al., 2020).
In the recent Russo–Ukrainian War, various MNCs (e.g., Apple, Nike and McDonald’s.) announced their withdrawal from the Russian market to demonstrate the stance against the war. These CSA-oriented market-exit decisions are substantially different from the traditional view of market exits with the economic logic of abandoning underperformed markets (Karakaya, 2000). Firms engaging in CSA can satisfy target stakeholders, which may benefit the corporate brand. Nevertheless, market exits implicate resources and cost commitments, making the financial return of the CSA-oriented market exits questionable.
Taking the MNCs’ market exit from Russia amidst the 2022 Russo–Ukrainian War as the research context, this project aims at exploring whether a CSA-driven market exit benefits firm performance in terms of (a) corporate brand equity and (b) profitability. This project will conduct a natural experiment to examine the treatment effect of CSA-oriented market exits on corporate brand equity and profitability. The analysis aims to quantify the impacts of CSA- oriented market exits by capturing significant abnormal changes in corporate brand equity and profitability after the decisions were made.
In addition, this project explores the factors that affect the benefits of CSA-oriented market exits. Through the theoretical lens of instrumental stakeholder theory, this project will examine the moderating roles of target market structure and media reaction in the abnormal performance changes caused by CSA-oriented market exits.
The outcome of this project is expected to provide theoretical contributions. Specifically, this project can contribute to the CSA literature by providing empirical evidence concerning the long-term impacts of CSA on brand and profit performance of firms. Furthermore, this project will provide evidence to demonstrate the importance of major markets’ support and positive media reaction to the benefits of engaging in CSA. In addition, this project can also enrich the literature on market exits by incorporating the lens of CSA in investigating the impacts of market exits.
Practically, this project will provide references for MNCs deeply involved in global markets in terms of how CSA-oriented market exits may affect their brands and profits. This project will also provide guidance on managing market exits in a more effective manner in the current international markets involving complex sociopolitical issues. The project outcomes can also inform governments of the quantitative effects of CSA-oriented market exits and the potential impacts on the economy, which can be important references for future policy making.
Project Reference No.: UGC/FDS24/B10/23
Project Title: Towards a Resilient and Sustainable Supply Chain under Industry 4.0 Perspective: Lessons Learned from the Tourism Industry for Mitigating the Pandemic Disruptions
Principal Investigator: Dr ZHANG Xinyan (PolyU SPEED)
Abstract
Supply chain resilience (SCR) is a desirable capability of a supply chain to restore operations in the event of a misfortune such as the Coronavirus disease (COVID-19). The new global reality that has emerged after the outbreak of COVID-19 has exerted an unprecedented influence on technology adoption in business processes. Industry 4.0 technology (I4T) is characterized by technological innovations that connect a firm’s supply chain network electronically and intelligently. Owing to the globalization phenomenon, many tourism companies have adopted the strategy of supply chain management (SCM). The COVID-19 disruption may generate unprecedented chances to develop a measurement scale for tourism supply chain (TSC) resilience and acquire insights on how TSCs can be more resilient. This project will grab this opportunity to investigate the resilience of TSCs in Hong Kong amid the COVID-19 pandemic crisis and discover insights into how TSCs can be more resilient.
Both SCR and supply chain sustainability (SCS) are critical issues of SCM. Literature shows that SCR and SCS can influence each other. The core emphasis of SCR is to sustain competition positions in the long term through effective business model development. This may imply a positive relationship between a resilient strategy and sustainability. The primary purpose of this project is therefore to investigate the relationship among I4T implementation, SCR and SCS in the tourism industry and to discover in-depth insights into I4T-driven SCM strategies for developing resilient and sustainable TSCs. This project will assist various TSC practitioners in developing SCR and thus enhancing Hong Kong tourism industry competence and responsiveness.