Competitive Research Funding Schemes for the Local Self-financing Degree Sector:
Children and youth are a sizable group in Hong Kong. They do not live in isolation but are affected by various personal factors and social systems. Due to the interdependent nature of the systems and the depth and breadth of the factors involved, we cannot analyse them in isolation through one single research project, but require a wide range of projects to address the multifaceted issues under the coordination of a research centre. Against this backdrop, the Children and Youth Research Centre at the Caritas Institute of Higher Education serves as a focal point for dialogue, consultation, training, education, research and international cooperation. The Centre aims to achieve the following objectives: (1) identify the risk and protective factors in relation to children and youth development and tackle them through positive education, career development, community involvement, counselling and family support; (2) support children and youth in exercising greater control over their lives, handling the challenges they face and promoting their social well-being in a responsible way; (3) strengthen their resilience against radical ideologies and develop their rational thinking to address the potential harm done to themselves and society, and (4) evaluate and improve current children and youth work practices and strategies, enhance their service impacts, and if necessary, revamp the current service models. In the long run, the Centre will benefit children, youth, family, service providers, practitioners, and professionals, and assist the government and NGOs in making informed policy and practice decisions.
Data science has been defined as the ability to gather, manage, translate, and communicate data in an effective way (Berkeley School of Information, 2020). Or simply, data science is the study of extracting value from data. With the great advances in computing power, storage and networking, numerous facets of social, business and industrial activities are moving on-line. This has ignited, and provided rich ground for data science related research, often interdisciplinary in nature, bringing together data scientists and domain experts to collaborate. The purpose of setting up a Data Science Research Centre for Social Policies and Services is to promote and support data science related social sciences research in our Institute.
While a lot of the successes with data science research are in the e-commerce and social media areas (e.g. recommendation systems), considering the strengths of our Institute as the largest provider of social work degree and sub-degree programmes and in accessing the extensive services of our partners and sponsoring body, which is one of the largest NGOs in Hong Kong, we propose to focus (at least initially) on supporting research in the areas of social services and policies. This focus can be considered part of the "Smart City" that governments worldwide are driving for. City services, whether carried out by government, NGOs or commercial entities, are increasingly supported by digital platforms that can systematically collect data, which can be used to enhance these services and relevant policies. The term "Smart X" is often used to describe a service (X) that smartly applies data science in its operation. Our goal is to do research to help make many of the social services "smart".
To achieve these objectives, the Data Science Research Centre for Social Policies and Services aims to provide the following resources: (a) Centralised computing resources that can be used for research projects, and storage of data that can be made available with controlled access, (b) Data Engineers who will help manage the data, clean the data for sharing, and offer some data science programming training on need basis, and (c) Organize seminars and workshops, to inspire data science based research, and enable our researchers to share their experiences. More importantly perhaps, the Research Centre will help get research projects started by liaising between researchers and data providers, and providing the often necessary consulting and data engineering support. The Research Centre also aims to provide some suitable topics and data for undergraduate student Capstone projects, involving outside social service workers as appropriate. This can greatly help students get prepared for the workforce after they graduate, and the higher calibre students can even be hired by the Research Centre as interns, when the needs arise and funding is available.
The success of Data Science Research Centre for Social Policies and Services in connecting the interesting practical research problems with the data needed for addressing these problems will attract research collaboration with academics in other tertiary institutions locally or beyond. Such collaboration will be very helpful for improving the research capacity of self-financing institutions by generating a larger collective critical mass of research.
The Techno-Humanities Research Centre (the Centre) at Caritas Institute of Higher Education is established to promote the use and development of technology in humanities research, provide technological facilities to colleagues to conduct research on techno-humanities, provide a platform for research collaboration on techno-humanities among tertiary institutions in Hong Kong, and create a network of techno-humanities scholars to advance academic study of this field.
To achieve the above objectives, the Centre will form a number of research units to conduct projects on the humanities, acquire digital resources for research, publish a refereed journal entitled Journal of Techno-Humanities, run a book series Routledge Studies in Techno-Humanities, hold international conferences and organize workshops and seminars to promote the academic study of techno-humanities in Hong Kong, and develop technologies for research in the humanities.
The objective of this project is to create an automatic football commentary system with the use of imaging, pattern recognition and tracking technologies, image to text conversion, text to voice output functions, and machine learning (especially deep learning) techniques. Results of the study will give a real-time automatic football commentary system with some original contributions in terms of good research papers in international archival academic journals and conferences. The targeted voice output is in Cantonese, however the investigated technologies should be useful for any language. The research work can also be extended to the use of IoT sensor in 5G for training football players.
The development and deployment of artificial intelligence is progressing rapidly throughout the world. Because AI is increasingly affecting many aspects of human life and will impact many professions, interdisciplinary research in AI is therefore needed. An AI laboratory is to be established to enable different schools to conduct research with deep learning and with possibility of using distributed learning technology.
Hong Kong is now facing a serious problem of fast ageing population with increased life expectancy. Caritas - Hong Kong, the sponsoring body of Caritas Institute of Higher Education (the Institute), devotes to resolving this problem through its Medical Services Division and Social Services Division which are two of its four major servicing areas. The Institute shares its mission and vision as our sponsoring body and has identified healthcare as its strategic area for development. The School of Health Sciences (the School) was established to focus on the education of healthcare professionals.
The Institute proposes to establish a Healthcare Research Centre (the Centre) as a flagship research centre at the Institute. The Centre aims to provide a powerful research infrastructure in healthcare and related areas. The Centre will promote and coordinate the research activities in this strategic area to produce synergy effect among all its research projects. It is expected that this area of research will enhance development and delivery of healthcare programmes in the School.
The Institute has planned the following Programme on Research at the Centre.
Summing up the above, the Healthcare Research Centre will play a major role in facilitating and overseeing all research activities in healthcare and related areas. We expect that these research activities will have a positive impact on the teaching of healthcare programmes in the Institute. As a long term prospect, the Centre will also seek to contribute to the efforts of the Hong Kong community in addressing various issues arising from its ageing population.
Central to the proposed Programme on Research is the setting up of a research office with its primary aim of providing a powerful research infrastructure to render support to staff members in their research, to ensure robust research standards and effective management of all research work through the following planned activities:
The Research Office will play a major role in facilitating and overseeing all research activities in the Institute.
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.
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.
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.
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.
Tax revenue is a critical financial resource for governments to carry out their governance functions. Tax avoidance threatens a government's ability to collect proper amounts of tax revenue according to tax laws. Therefore, the question of how to better protect and collect tax revenue is an important challenge that needs to be addressed continuously by governments.
Governments can use both carrots and sticks in revenue collection. Attention from professional practice and academic literature has focused mostly on the use of tax audits and penalties to enforce tax compliance. Not much attention has been devoted to the use of incentives to entice compliance. The main objective of the proposed research is to shed light on the efficacy of an incentive-based scheme coordinated and implemented by the Chinese tax authorities intended to reduce tax avoidance.
Specifically, we propose to investigate whether a tax credit rating system that classifies companies based on a variety of factors related to tax payments (with A-rated firms being the best compliant) and has the ratings publicized can effectively reduce corporate tax avoidance. We also investigate whether non-A-rated firms lose out resources such as the ability to obtain bank loans and issue corporate bonds because of the inter-agency cooperative agreements orchestrated by the Chinese tax authorities. Such a loss of resources can motivate non-A-rated firms to reduce their tax avoidance for an improved rating.
We posit that the publicized rating for A-rated firms and the loss of resources for non-A-rated firms provide incentives for non-A-rated firms to reduce their tax avoidance in an effort to gain a better compliance rating. If so, we should observe a spillover effect created by A-rated firms for peer firms in the same industry to reduce tax avoidance.
The proposed research studies a new development in tax administration practices. Furthering our understanding of this new development provides opportunities for future administrative innovations. The research results should have significant policy implications for tax authorities, corporate management, investors, auditors, and relevant public policy makers. As this incentive system represents an important new development in tax administration, if proven effective, it can also serve as a useful model for tax authorities internationally, particularly those in developing economies, to enhance corporate tax compliance.
We propose a safe technology of Chaotized Plasma to disinfect suspended viruses because plasma air ionization was proven to reduce Coronavirus Surrogate MS2 Bacteriophage by 99% in Independent Spanish Testing and many other places. UK Health and Safety Executive said, "Do not rely on recirculation units which just move COVID virus around". Omicron viruses were found in the air purifiers at Silka Seaview Hotel on 18/1/2022 and at Hakka Cuisine in the Lion Rock Shopping Centre on 31/1/2022 and people were infected just under the air purifiers. Just air exchanges as mandated by the Hong Kong Government in designated places are not scientific as viruses are still active and infecting the air purifiers. Existing disinfection technologies, e.g. nano-coating, silver ions, spirit, and UVC(Ultraviolet-C light) are for landed viruses only. However, over 90% of COVID infection is through the air. The European Commission has published its decision to classify nano-coating as a category 2 carcinogen.
Evidence from the HK Airport Toilets and Moon Palace Restaurant shows that Omicron is not a landed virus and is transmitted mainly through the air. On 2/1/2022, the Health Protection Centre inspected the Moon Palace Restaurant in Festival Walk and collected over 80 environmental samples with no landed viruses found. On 4/2/2022, 5 people were infected in Caritas Hospital but the ventilation was good. Therefore, good ventilation may not be able to avoid infection. The main cause of the disease is airborne. Air disinfection is necessary.
The PI used chaotized bleach water in 2003 to disable viruses from patients coughing at an airspeed of 22m/s (funded by the Research Grant Council) and then chaotized ozone water in 2020 for negative pressure isolation rooms (funded by ITF). The former devices were installed in the Prince of Wales Hospital and the latter in Pok Oi Hospital and Society for the Blind Care Homes successfully. Now, instead of vapor, we use chaotized plasma, the first time in the world. Plasma is produced by elevating the usual 220V voltage to 5000V and discharging between positive and negative poles. For example, O2 is split into O+ and O- ions, both will neutralize back to O2 releasing tiny energy to disable microbe's reproduction. However, ions and viruses are extremely small (nanometres) that they have little chance to meet in a finite space. Chaotic mixing of ions and virus is created inside the invented use of Chaosors so that, in a trial run using Eilat coronavirus, 99% efficacy of disinfection within 0.1 seconds can be achieved. It was found also that the efficacy is reduced to less than 10% when plasma was not chaotized. Plasma can also decompose big air molecules like formaldehyde and H2S. Therefore, chaotized plasma can deodorize.
The project's deliverables include (1) chaotic simulation of viruses and plasma ions by computational fluid dynamics; (2) hardware: air disinfection devices for use in institutions; (3) Indoor Air Quality monitoring devices in smart and healthy city applications; (4) student learning materials in engineering, biophysics, and microbe disinfection; (5) patents, seminar and conference presentations and paper publications. New knowledge for teaching and learning in hygiene practices and air disinfection will be created in the post-COVID19 healthy-smart city. The device will be installed at the inlet of an Air Handling Unit to clean air almost costless. It will also be tested in a model elevator. The piston effect of the lift car movement provides the airflow and cleans the adjacent air at merely 0.04% additional energy cost by chaotized plasma. The PI will use four of his invention patents in this project. Extensive public engagement activities are arranged for wide application and commercialization.
Clustering is a fundamental task in data analysis, which partitions a set of samples into homogeneous groups. Over the past few years, graph-based clustering algorithms are becoming very popular and widely used in various applications due to the high efficiency and simplicity in implementation, i.e., graph-based clustering only requires a typical pairwise similarity matrix of samples (a.k.a. a graph) as input and performs spectral decomposition on that matrix to generate the clustering result. Despite its popularity, some fundamental issues still exist, which greatly affect its performance.
First, the performance of graph-based clustering algorithms is highly dependent on the quality of the input similarity matrix. Therefore, how to construct a high-quality graph which can best capture the underlying structures of data is the core of this kind of method. Besides, graph learning is also a hot research topic in signal processing. However, it is a challenging task. The traditional methods usually use the Euclidean distance of samples to construct the pairwise similarity matrix, which is very sensitive to the noise and usually suffers from high computational complexity. Besides, in practice, there is some side information available that may provide valuable information to promote graph-based clustering, such as the available supervisory information, which may be overlooked or not fully exploited to some extent. We will propose a unified theoretical framework capable of fully exploring the available side information, including the raw features and some week supervisory information, to learn a more robust and informative similarity matrix for graph-based clustering.
Second, most of the traditional graph-based clustering methods adopt linear models to cluster the data, but in practice, the data is not necessarily located in the linear subspaces, which may significantly affect the performance of graph-based clustering algorithms. It becomes insufficient and unreliable to only use the linear models in graph-based clustering. In addition, inspired by the powerful representation capability of the deep learning framework to model the nonlinear system, in this project, we will develop a novel deep neural network architecture to learn a nonlinear mapping of the data to well adapt to the subspace clustering.
With our solid backgrounds and the promising verifications achieved, it is highly expected that our investigations will provide more efficient and effective graph-based clustering algorithms for the large-scale dataset. Clustering is a fundamental and universal problem related to many topics, including low-rank matrix/tensor approximation, low-dimensional embedding, semi-supervised learning, noise estimation, deep learning, etc. In addition, the constructed high-quality graph can be widely used in various tasks in data analysis rather than just limited to clustering. We believe that beyond the three years envisioned for this work, the scientific findings of this project will continuously motivate the research on a wide range of research communities and benefit various real-world applications.
Calligraphy is a popular visual art form that intends to produce pleasant writing, especially in East Asia. However, it is not easy for amateurs to write calligraphy due to the delicate writing trajectory and brush pressure control. To help beginners in calligraphic writing, calligraphic animation videos are extremely helpful to show the writing trajectories and brush pressures at all writing positions. Besides, calligraphic animations are frequently used in different digital media as a better way to present and animate the text, such as in movies, dramas, cartoons, and advertisements. While there are tools available for manually producing calligraphic animation, the process is tedious and requiring professional skills. An automatic calligraphic animation generation system would help much in this tedious and time-consuming calligraphic animation making process. What's more, the generated animations with writing trajectories and brush pressures can be directly used in robot writing as well.
An automatic calligraphic animation generation system includes three parts: segmenting the calligraphic image into individual strokes, identifying trajectory and brush pressure for each stroke, and generating animations to all strokes. There are several key challenges to be solved. Firstly, the writing topology of a calligraphic character may not be the same with the formally written one (Figure 1). So, even though the writing order of the strokes for formally written characters is well defined, the calligraphic character may not follow the defined writing order. So, it is not realist to hardcode the writing order for animation generation. Secondly, the writing habits are very different for modern and traditional Chinese. The modern Chinese is customed to write from first left to right and then top to down, while the traditional Chinese is customed to write from first top to down and then right to left (Figure 2). It is needed to identify the writing order before animating the characters. Finally, the generated writing trajectory and brush pressures should be precise and smooth to simulate the real writing.
Despite the usefulness of the system in education, entertainment, and automation, the existing methods generally only focus on one part of the system. Without taking consideration of the whole system, the existing stroke segmentation methods usually only generate low-resolution output, so the segmented individual strokes can hardly be directly used for animation generation. Besides, the state-of-the-art stroke generation methods are usually trained on a specific font style and are error-prone to even minor font changes (Figure 3). On the other hand, the existing text animation generation methods generally rely on hardcode writing topology of each character, the character writing order and calligraphic characters that do not obey the defined writing topology are still lack of discussions.
In this project, we propose a novel system which consists of a novel learning-based stroke segmentation module, a contour-based trajectory and brush pressure identification module, and a learning-based stroke animation generation module. The key idea of the stroke segmentation module is an iteratively refined stroke segmentation network to gradually adaptive the network to different calligraphic fonts. The key to the trajectory identification module is to obtain the precise trajectories with brush pressure based on the contour information. The key to the stroke animation generation module is to first identify the writing order of the corresponding calligraphic image and then generate the ordering and animations of the strokes based on the identified writing order.
With the proposed system, we believe the tangible outcomes, e.g., publications and algorithms, should directly benefit the industry and the research society. The research project would also provide invaluable chance in developing the skills of the teachers and the students in artificial intelligence and digital entertainment technology, which is also part of the curriculums and programmes offered by the school and the institute.
Nowadays, a low resolution image gives us the impression of low-tech, old fashion, awkward style and uncertainty. This is particularly true if the image is used for advertisement, surveillance, medical diagnosis or object recognition. The low quality face image might devalue a person's beauty and faith. In order to relieve the problem we may turn the image into higher resolution, with super-resolution technology. Image super-resolution (SR) usually refers to an increase of the resolution of a single low-resolution (LR) image by up-sampling, deblurring and denoising, while the resultant high-resolution (HR) image should preserve the characteristics of the natural image, such as sharp edges and rich texture. This proposal is on a study of large-scale face super-resolution of small images, or even incomplete images. In this proposal, an incomplete image refers to an image which is so small that it is faint and unclear, and part of the details might be missing. This is a very difficult and ill-posed problem, since a number of unknown pixels have to be inferred from very limited information, say in the case of 16x super-resolution, or even higher.
We propose to investigate large-scale face super-resolution not only because we have fruitful experiences in face recognition and regular face super-resolution, but enlarging a face with good quality is always demanding, and viewers can easily appreciate the effect of good quality. We must stress that results of our investigation should be useful for many applications, such as on-line teaching, video conferencing, remote medical diagnosis, remote operation monitoring, computational photography, video surveillance, multi-media amusement and metaverse. Actually, the resultant techniques can also be generalized (not limited to faces), applying to medical imaging, the development of low cost electronic microscope, etc.
We have a strong background in digital signal processing, imaging and video technology, pattern recognition, machine learning and deep learning, which are particularly suitable for this proposal. We start to propose a new and effective deep learning structure with back projection framework for the face super-resolution. Novel arrangement and new algorithms will be proposed with some new techniques in deep learning, including the generation of a latent edge component, back-project strategy, smart arrangement of StyleGAN for high-resolution face generation, attention mechanism for global structure acquiring, etc. The following are some more brief points.
(i) We start with the super-resolution of 64x64 face images to 1024x1024 images.
(ii) Edge quality is important to an image, hence we form latent vector structure which is generated with the assistance of the edge image generated from the original LR image.
(iii) Pre-trained StyleGAN which can generate sharp image is used to generate the high-resolution image, with the guidance of the edge image.
(iv) A new back-projection strategy is designed for the overall structure of deep learning network.
(v) Our initial test has verified that the above arrangement is good, and seems more attractive compared with conventional approaches.
Inadequate retirement protection has long been a pressing problem for the elderly in Hong Kong. The major retirement pillars for the general public are the Mandatory Provident Fund (MPF) introduced in December, 2000, and the social security system for the low-income group. After the recent reform of the retirement system in 2017, half of the elderly persons are now supported by the means-tested social security system, which provides about 20% of the median employment income. The rest of the population have seen no substantial improvement for them. Those who earn higher than the median income will not be eligible for social security upon retirement and probably many years afterwards. Many will find it hard to maintain their standard of living if they retire at the normal retirement age of 65, and so must work afterwards. In the 3rd quarter of 2021, around 26% of those aged between 65 and 69 were still in the labour force, with a higher labour force participation rate among the male population (36.3%).
This study will focus on those aged between 65 and 69 who were employed at 60-64 with an employment income fell into the third income quartile regardless whether they are in the labour force or not at the time of interview. In the third quarter of 2021, this income quartile was between $20,000 and $33,000. The income range varies depending on their age during the interview. This proposed study will investigate their retirement and employment arrangements, namely, whether they are not-retired, i.e. having either a full-time or part-time employment, temporarily retired but having an intention to seek job, or completely retired. For those who are temporarily or completely retired, they will be further classifed as voluntary or involuntary in such arrangements based on whether they are dismissed or displaced from their job, and whether they retire completely because they believe that they cannot find suitable job in the labour market even if they want to work.
It will also study the factors associated with such arrangements, as well as the challenges they faced as elderly job-seekers and employees. Specifically, these factors will include financial readiness, caring responsibilities, and perceived age discrimination practices. A number of personal factors, such as gender, property ownership, health conditions, occupation, and family composition will be collected and controlled to identify the contribution and relative importance of these factors in their emplyoment and retirement arrange.
A representative sample of the elderly persons aged 65-69 meeting the income criteria when they were 60-64 will be identified through randomized telephone contacts, followed by a survey with a structured questionnaire. The effective sample size will be 500. Follow-up in-depth interviews will be conducted with 18 respondents based on the result of the survey to gain more insights into the trajectory of their retirement and employment arrangement, and the process and mechanisms they made their decisions.
The study will generate useful information to make policy suggestions to improve the employment environment to allow them to have more and better choices after normal retirement age.
Health literacy (HL) means much more than just the ability to read health-related information or make clinic or hospital appointments. For patients with chronic illness, like chronic kidney disease (CKD), HL is a personal attribute which relates to the achievement of a level of knowledge that can help to prevent the development of health-related problems and protect health. In 2022, the focus of World Kidney Day will be 'Bridge the knowledge gap to better kidney care', emphasizing the importance of information needs and HL in CKD.
Globally, CKD is becoming increasing common because of the growing prevalence of diabetes mellitus, hypertension, obesity, and ageing. Left untreated, CKD is likely to progress to end-stage kidney disease, with higher risk of mortality. Fundamentally, it is the patient who is tasked with understanding, implementing, and maintaining the medical recommendations for CKD self-management. As such, HL is of great relevance in vulnerable patient groups, like immunocompromised patients with CKD. Currently it is estimated that over a quarter of CKD patients have limited HL, however the full extent of the problem remains unknown in Hong Kong.
Increasingly CKD management involves self-management activities, therefore, as well as HL it is crucial that patients have good levels of self-efficacy so that they can gain access to, understand, and utilize health-related information. Inadequate levels of HL constitute a risk factor for low health outcomes and poor treatment compliance. In addition, limited health literacy is associated with poor control of disease, greater risk of cardiovascular disease, more missed treatment appointments and higher rates of hospitalization. The current situation is made even more challenging for those with CKD due to COVID-19 related issues.
In face of these clinical challenges, HL-sensitive forms of communication and educational support packages may play an important role in successful disease management, slowing down the progression of this chronic disease.
It is proposed that greater understanding of information needs, and HL could increase local patients' ability to self-manage their disease more successfully and slow down the progression of their chronic condition. Despite the proposed influence of HL, research into this important ability remains lacking in Hong Kong.
In this three-phase mixed method study, we aim to address this issue. From a quantitative perspective we plan to explore levels of general, COVID-19 related health literacy and levels of self-efficacy in individuals with CKD who attend one renal self-help group in Hong Kong with validated questionnaires, to address adequacy of levels of HL. Then from a qualitative viewpoint, focus groups will be used to identify information priorities and needs to ascertain barriers to understanding and acting on health-related information in this patient's group. Based on the information gained in the second phase of our study, a tailormade online health literacy support package for CKD patients will be developed in the third phase.
In this study we hope to address an important current knowledge gap, reducing potential inequalities of care and having a positive impact upon the health of CKD patients in Hong Kong and beyond. Ultimately, improving levels of HL in CKD patients may help to slow down disease progression, promoting more effective levels of self-efficacy, self- management, and optimal health outcomes.
The introduction of glucagon-like peptide-1 (GLP-1) receptor agonists changed the treatment strategies of type 2 diabetes mellitus (T2DM) as GLP-1 receptor agonists not only provide glycaemic control but have additional body weight control and cardiovascular benefits. However, the use of GLP-1 based therapy can be associated with gastrointestinal side effects, including nausea and emesis, limiting the doses that can be used. The mechanisms controlling emesis and the sensation of nausea and those involved in appetite are subtly different and is less well understood. Research strategies elucidating emesis and anti-emetic drugs have focused on the brainstem vomiting centre. The potential involvement of the hypothalamus, which is integral to autonomic control, has been overlooked. This project is based on our own original findings that intracerebral paraventricular hypothalamic (iPVH) administration of exendin-4 inhibited significantly food and water intake and induced emesis in a dose-dependent manner in Suncus murinus. Exendin-4 appeared more potent in inducing emesis following iPVH compared to intracerebroventricular administrations. We also showed that subcutaneous administration of exendin-4 induced emesis but not the associated inhibition of feeding was antagonized by iPVH administration of the GLP-1 receptor antagonist, exendin (9-39). In addition, the emetic effect of exendin-4 was dissociated from its anorectic effect. These findings suggest that hypothalamic GLP-1 receptors may be at least partially involved in mechanism of nausea and emesis. The parabrachial nucleus receives reciprocal inputs from the hypothalamus, amygdala and limbic system and sends projections to the nucleus tractus solitaries. We hypothesize that GLP-1 receptor activation in the PVH may modulate local release of GABA and/or glutamate and other transmitters which act as key modulators in these brain areas and are involved in mechanism of emesis control as well as nausea which is a subjective feeling involving higher brain functions.
In the present project, we aim to determine the role of the hypothalamic GLP-1 receptor system in feeding and emesis and the possible underlying mechanism of the signalling pathway. Animal experiments will be performed using standard behavioural testing and established radiotelemetric techniques to evaluate physiological changes indicative of nausea (PCIN) coupled with c-Fos immunohistochemistry analysis of brain function. Changes in brain neurotransmitters will be monitored using brain microdialysis. Our studies will uncover a novel mechanism of nausea and emesis. Data obtained from studying the emetic mechanisms of GLP-1 receptor system may not only enable a more improved management of diabetes and obesity, but may also lead to the discovery of new target for anti-emetic development and improve the quality of life of cancer patient receiving chemotherapy.
Comic is a world-wide popular art form enjoyed by people of all ages. However, creating a comic book is time-consuming. It generally takes a professional comic artist one or several days to finish a single comic page. Usually, creating a comic page takes three steps: sketching, inking, and screening. The artist will first convey the idea and concept by quickly sketching a rough drawing using freehand strokes and then trace a clean inked line drawing. Finally, screentones will be applied to the inked line drawing to enrich the visual content (Figure 1). While the screening process is highly tedious and repetitive, there still lacks automatic tools to release the artist from this tedious process. So, comic companies and famous comic artists generally hire several assistants to help with the tedious and repetitive screening task. An automatic comic screening system is highly desired in the comic industry for saving time and labor.
However, automatic screening a comic line drawing is extremely challenging. There are many different screentones of various shapes and tones. Finding the proper screentones to fill the white areas in a line drawing in non-trivial. Firstly, not necessarily all areas need to be filled with screentones, e.g., human faces and hands are usually screentone-free. Secondly, the tone of the screentone should correctly express the object's color and shades, e.g., a dark object should be filled with a low-tone screentone, and vice versa. Thirdly, screentones might be affected by semantic usage, e.g., stripes are commonly used to express the shades. Lastly, areas belonging to the same object should be filled with the same screentone, e.g., the two sleeves of a shirt should be filled with the same screentone. But all the above characteristics of screentones are highly related with the high-level semantics, which is extremely difficult to be analyzed, especially for line drawings where most of the pixels are either black or white and exhibit similar local charactertistics.
Recently, the deep learning technologies have greatly advanced the development of various image analysis and image synthesis applications. Nevertheless, the existing deep learning solutions face significant problems when dealing with line drawings and screentones. Firstly, the input line drawing is relatively simple and uninformative, which only contains a solid white background and some sparse black lines. In the meantime, the target comic image is much more complex that contains color, material, and lighting information presented by screentones of different appearances and densities. The existing methods generally fail to resolve this large information gap. Secondly, screentones are composed of high-frequency elements such as lines and dots. However, the generation of the output image is interpolative with deep neural networks, so the existing methods generally generate blurry outputs (Figure 2 and 3).
Despite the difficulties of generating screentones with intensive strokes from simple line drawings, we propose a novel learning-based solution to automatic generate the screentones for an inked line drawing. The key idea of our design is to split the difficult line-to-screentone process into two easier-to-solve problems by taking the tones as an intermediate representation between line drawings and screened comics. So, the line-to-screentone task now consists of two processes, a line-to-tone process and a tone-to-screentone process. While the information gap between line drawing and screened comic is large, the information gap between line and tone or tone and screetone is much smaller, which could be potentially learned and predicted by the deep convolutional neural networks.
With the proposed system, we believe the tangible outcomes, e.g., publications and algorithms, should directly benefit the comic industry, the research society, and comic customers. Moreover, the research project would also provide invaluable chance in developing the skills of the teachers and the students related to digital entertainment and artificial intelligence technologies, which is also part of the curriculums and programmes offered by the school and the institute.
Cartoon is one of the most widely used and popular non-photorealistic media forms. Starting from the 19th century, cartoon has been gaining wide popularity due to its iconic exaggerated drawing style and its freedom in storytelling and expressing the artists' ideas. Recently with the development of the mobile web and social media, people tend to share their thoughts and ideas with more creativity and uniqueness. For example, after Spiderman's movie release, the audiences may want to represent their own life as comic strips and wish to have their own life stories illustrated in the style of the Spiderman comic by converting their captured photos and videos into comic frames. On the other side, professional cartoon and animation productions also have a specific need to create cartoon frames based on existing captured photorealistic scenes and characters. Unfortunately, current approaches to render natural photos and videos into cartoon styles usually require expertise in computer graphics pipelines, knowledge of art, and intensive human labor, which limits the further growth of these cartoon-related applications.
This project intends to propose a fully automatic framework to tackle the cartoon stylization task. The proposed framework can apply visually aesthetic cartoon styles to a diverse set of user-specified input, including both photographs and videos. The realization of the framework meets several challenges: firstly, because of the free-from sketch basis and the arbitrary shading styles in cartoon production, the characters, objects, and scenes must be rendered with the unique exaggerated and abstract cartoons styles. However, simple image filtering cannot achieve this task because it can hardly understand the image content and apply cartoon-style image editing. Besides, the conversion should still retain the original image content so that people can still recognize the content after stylization. Yet, unfortunately, no existing style transfer nor conditional image generation frameworks can solve this challenging problem. Finally, when processing video contents, the framework should also take care of the temporal coherence to make the result looks smooth and pleasant during playback.
Our preliminary approach to achieve cartoon stylization is constructing an unsupervised cross-modal understanding and synthesis between natural contents and cartoons with an instance-level awareness through deep learning techniques. Specifically, the term instance means the stylization is on an object-level basis. The instance-wise conversion also matches the real-life cartoon stylization pipeline by the artists. To build up our framework, we first extract object instances from raw photo images and cartoons and facilitate a feature alignment network to learn a cross-modal feature alignment with an awareness of the underlying object category (e.g., humans, cats, moving objects, etc.) After the input instance-level features are aligned, we propose to synthesize the features into stylized cartoon instances with a cartoon synthesis network. Afterward, we introduce a canvas refinement network to composite the individual objects into a single canvas. The network learns to enhance visual harmony among these converted instances and (optionally) proceeds to output temporal-smooth video sequences.
After the framework development, we shall evaluate the expressivity and the visual quality of our outputs by a comprehensive subjective study from both casual users and experienced artists. We shall build software toolkits and API interfaces for tighter integration with existing creative software studios and mobile applications. We believe the proposed framework helps provide a more effortless and innovative way for the public to express their inspirations and lifestyles. Moreover, we also expect the proposed framework to benefit the whole cartoon and animation industry by cutting off the tedious visual design and asset preparation procedures. This project's potential outcome, including the deep learning algorithms, interactive software, and academic publications, will help motivate the computer graphics and deep learning community. At the same time, we hope that this project can continually contribute to developing students' and teachers' skills in digital entertainment and AI technologies of the institute.
Algorithmic game theory is a research field that integrates game theory and algorithm design. The major target is to design good algorithms in strategic environments. In this project, we are going to study a well-established problem in algorithmic game theory called Facility Location Games. In the most classic setting, the government (service provider) selects locations to place its facilities on a street where some strategic agents (consumers) with private information live. Agents want to be as close as possible to the facility but they may report wrong information to the service providers in order to get more benefits. The objective of the service provider is to collect the agents' information and use a deterministic or randomized mechanism to decide where to build the facility so that certain objective values are approximately optimized and the agents will not gain benefits by reporting false information.
In recent years, many extensions to the original facility location games have been proposed and studied by researchers. Most of them assumed that agents had uniform preferences, which cannot model the complex scenario in real application. In this proposal, we plan to study more diversified preferences of the agents, and the utility or cost of an agent with preference might be determined by a single facility or by many facilities. Suppose that the government plans to open two schools (or kindergartens) in a district, where one school pays more attention to music education, and the other school mainly focuses on physical education. The parents, who plan to send their children to school, have their own perspectives on education, and thus have different preferences. The parents make decisions (on which school to choose) based on their preferences and the locations of the two schools: if their favorite school is much farther away from home compared to the less preferred one, they would choose the less preferred one instead, which implies a tradeoff between the distance and the preference. This is an example of single-facility-dominated facility location game, where the cost or utility of the parents depends on the single facility (the school) they choose. Consider another case that the government plans to build a hospital and a clinic in a district. If a patient needs to visit the doctor in the hospital twice a week, and visit the clinic once a week, then we can model the case as an instance of many-facility-dominated facility location game. According to the frequency they visit to hospital and clinic, we could assume that the patient prefers 67% on the hospital and 33% on the clinic. The utility of this patient should be a weighted sum over the utilities received from each facility, which is determined by many facilities. This is known as fractional preference. We extend to the following scenario where the government has a budget of money to assign to the hospital and the clinic, which can be used to maintain or improve the facilities. An immediate question is how to assign the budget in order to better serve the patients.
In this project, we study the facility location games with ordinal and cardinal preferences. For ordinal preference case, the agents are required to report a ranking over the facilities, and their costs are decided by the facility that provides the least cost; for cardinal preference case, the agents are required to report a fractional (or proportional) for each facility and the government needs to decide a budget allocation to these facilities. Under well-defined system objectives, we aim to design deterministic and randomized strategy-proof mechanisms with good approximation ratios, and provide lower bounds. In addition, we study both cardinal and ordinal preferences when the government needs to both locate the facilities and allocate the budget. Our study of these new preference dimensions will enrich our understanding of possibilities that can arise with agents' preference matrix with respect to facilities. Therefore, the study could provide a more complete picture of facility location games with the existence of diverse preferences. We also define unified cost for agents with cardinal preferences and agents with ordinal preferences, so that it is possible to allow these two types of agents to co-exist, further elevating the problem scope to a higher level of generality.
The design of novel functional nanomaterials is increasingly dependent on the creation of new complex architectures and, more importantly, the understanding of structure-performance relationships at the nanoscale. Numerous novel strategies for production of functional nanomaterials, including those for biomedical applications and controlled release drug delivery have been investigated based on the complex nanostructures such as core-shell, side-by-side (often termed as Janus) and their combinations. However, how to robustly produce these complex poses a big challenge to the researchers in the fields of physics, chemistry and material engineering.
In our recent work, we successfully developed several kinds of multifluid electrospinning processes to generate complex nanostructures, allowing for tailoring the functional performances of the resultant nanofibers. We have also developed modified coaxial and modified tri-axial electrospinning to create hierarchy nanostructures from un-spinnable working fluids.
Building on this success we are proposing the preparation of new tri-layer Janus nanofibers using a new 3-fluid side-by-side electrospinning process, aiming to figure out two key fundamental issues. One is the nanofabrication mechanism of tri-layer Janus nanofibers using the tri-fluid side-by-side electrospinning process, particularly the influence of un-spinnable fluid on the formation of an integrated tri-layer Janus structure. The other is the structure-performance relationship based on the tri-layer Janus nanofibers. The knowledge gained from these two issues should benefit the developments of new multifluid electrospinning processes and provides more strategies for conceiving multiple-functional applied nanoproducts.
The main objective of the proposed research is to investigate and compare, with reference to the book-tax tradeoff theory, the relationship between taxable and accounting earnings immediately before and after a firm's Initial Public Offering (IPO). IPO is an important event for a firm and its shareholders. The Shanghai and Shenzhen Stock Exchanges are two of the leading players in the global IPO market. Between 2016 and 2019, the Shanghai Stock Exchange ranked as one of the Top 5 worldwide in terms of IPO fund raised. The Shenzhen Stock Exchange also ranked as one of the Top 5 in 2017. Given the large number of foreign and institutional investors involved in IPO, our research findings from China should also be relevant for economies beyond China. In particular, the Shanghai and Shenzhen-Hong Kong Stock Connect facilitates Hong Kong investors to invest in the Mainland Chinese market.
If the framework of a sound theory can help establish a significant relationship between corporate tax avoidance and IPO, then the proposed research will provide important implications for investors (domestic and foreign), tax authorities, regulators, auditors, and other capital market stakeholders. Our research results can help various stakeholders make more informed decisions and formulate policies and regulations to address tax avoidance during the pre- and post-IPO periods.
Hong Kong is facing the challenges of population ageing. Along with this is a serious gap between the supply of long-term care services and their demand. In response, the case management model has been recommended as a new service delivery mode for community care services in Hong Kong. The literature on case management has been critically reviewed, and it is found that most of the studies focus on evaluating the outcomes of case management, and only a few have examined the process. In Hong Kong, none of the research on case management has focused on the social welfare sector. To address this knowledge gap, the study will conduct a process evaluation of case management in community care services in Hong Kong, utilizing a qualitative research approach.
The goal of this research is to obtain evidence that will allow policymakers understand when and how case management will be effective, how to institutionalize such a system, and what particular forms of case management would be recommendable.
The proposed study is the first study in Hong Kong to examine the perceived academic stress of students exhibited in their placement. The relationship between the background of students, their level of perceived academic stress, and the response of students will be examined. It will also study the impact of the perceived academic stress upon students' satisfaction with their placement, their well-being and professional identity. A mixed methodology approach will be adopted and all local recognized social work degree and post-degree students having their fieldwork placement will be invited to participate in this study. The level of academic stress students perceived, their coping methods, and their level of satisfaction towards their placement, well-being and professional identity will be measured by quantitative methods while the explanation of the relationship of the variables will be explored through in-depth qualitative interviews of a small group of voluntary informants. The findings will be useful to develop strategies that can better prepare students for their practice learning.
This study aims to study (1) opinions of students and teachers on the implementation of school-based sexuality education; (2) appraisal of sexuality topics (e.g., menstruation, puberty, self-image, sex harassment, wet dreams, etc.) by students and teachers; (3) students' and teachers' sexual attitudes; and (4) their attitudes towards school-based sexuality education. The influence of sociodemographic characteristics, religiosity, and spirituality on the above four aspects; similarities and differences between students and teachers with regard to sexual attitudes and factors associated with would be examined.
Data would be collected from 1800 students and 300 teachers of secondary schools from four major districts of Hong Kong, i.e., Hong Kong Island, Kowloon, New Territories East, and New Territories West. The findings would contribute to the development of curriculum sexuality education and teacher-training course on sexuality education, which respectively meets the needs of students and teachers. With a well-planned school-based sexuality education programme delivered by well-trained teachers, students would not only acquire valid sexual information that helps avoid negative sexual outcomes, but also develop healthy sexual attitude and prepare themselves for a healthier sexual life..
Aim: The aim of the study is to examine current practices and attitudes of staff towards physical restraint and restraint-free care for older people with and without dementia and residing in residential care homes.
Design: A descriptive qualitative approach and critical incident technique.
Settings: Two non-government organizations and a private aged home will provide residential care home service for the elderly in Hong Kong.
Participants: Purposive sampling will be used to recruit staff in residential care homes. Selection criteria include those providing physical or social care to the residents with and without dementia; at least one staff from each grade; has worked in this home for more than 3 months; and have used restraints before. It is estimated to recruit around 70 staff.
Methods: Participant observations and audio-taped interviews will be used to collect the data. An interview guide will be piloted before collecting data. Staff will be invited to provide information about their own experience of using restraints in order to understand the incidents. Content analysis will be used to analyze the data.
Results and Conclusion: The findings will provide data to understand the current practices adopted and attitudes of staff towards the use of physical restraints of older people with and without dementia in institutionalized settings. Situations warranting restraint-use and situations when restraint-free care can be considered will be identified and compared. This study will provide a basis for further research and teaching into proper restraint use and restraint reduction for older people in Hong Kong.
Since the theoretical work of DeAngelo (1981), audit firm size has been widely used as a surrogate for audit quality in accounting and auditing literature as large audit firms with a greater number of clients have more to lose in case of audit failure. Other reasons including better training programs and ability to recruit better quality graduates by large audit firms to support the use of audit firm size as a fundamental proxy of audit quality. Nevertheless, we observe that large audit firms had not taken good care of many of their clients, resulting in numerous audit failures and legal actions over the years. In contrast, many small and medium audit firms are headed by prominent partners and have worked hard to provide the best service they can, especially for their important clients. This study investigates whether smaller audit firms provide better quality audits for their listed clients compared with Big 4 firms or Top 10 large firms for audit of clients of similar size and characteristics in China. We note that clients of similar size should normally be more important for a smaller audit firm than a large audit firm. We test whether the effect of audit firm size may be offset by the effect of client importance.
Cocaine- and amphetamine-regulated transcript (CART) peptides are important modulators that act in the brain and contribute to a range of physiological and behavioural processes including feeding, stress, energy expenditure and body weight control. In rodents, CART peptides and their mRNAs are found in many brain regions and in peripheral tissues that are involved in reward/ enforcement, feeding and emesis, and there is strong anatomical association between CART and various orexigenic and anorexigenic neuropeptides. Acute administration of d-amphetamine, which upregulates CART mRNA in rat striatum, is anti-emetic against apomorphine-induced emesis and motion sickness. Central administration of CART peptides in rodents inhibits food intake and induces c-Fos expression in the hypothalamic neuroendocrine neurons, nucleus tractus solitarius (NTS) and area postrema (AP). Recent evidence indicates that alternations in CART have been associated with reduced metabolic rate, obesity and increased risk of type 2 diabetes, making it a potential target for anti-obesity drug development. Nevertheless, the potential involvement of CART system in emesis control is poorly understood because common laboratory animals (e.g. rat and mouse) are incapable of emesis. In this proposal, we will use Suncus murinus, a species with proven translational value in anti-emetic research in our studies. Our recent data identified that the CART protein in S. murinus exists in the short form of 117 amino acids. In the C-terminal region of the protein, the region that shows physiological activity, there are only two amino acids difference between human, rat, mouse and S. murinus forms. In conscious animals, central administration of CART (55-102) produced a complete block of cisplatin-induced emesis in 83% and 67% animals in the first 60 and 90 min, resulting in an overall reduction in vomits by 89% and 79%, respectively, compared to the saline-treated group. CART (55-102) also produced a complete inhibition of cyclophosphamide-induced emesis in 5 out of 6 animals tested over a 4-h observation period, without an effect on food intake. In the current proposal, therefore, we will identify the anatomical distribution of CART peptide and CART mRNA in S. murinus using immunohistochemistry and in situ hybridization, respectively. We will elucidate the mechanism of action of the CART system in emesis control. Experiments will be performed using standard behavioural testing and established surgical and radiotelemetric techniques to track a cluster of physiological changes indicative of nausea (PCIN). c-Fos immunohistochemistry will be performed to gain more insight into the neuro-anatomical signaling following various nauseagenic treatments. Changes in brain neurotransmitters will be determined using LC-MS analysis. Our findings will unlock the role of CART peptides in the mechanism of emesis and also provide insight on whether targeting CART system could be a potential therapeutic strategy for anti-emetic development.
Wireless networked control system (WNCS) is a control system wherein the control loops are closed through a wireless network. The defining feature of a WNCS is that different system components (including sensors, actuators, and controllers) exchange the control signals and feedback signals among them in the form of information packages via the wireless network. There are many sensing and control applications like self-driving vehicles, where the environmental data is collected periodically and control decisions are made by the control algorithms and delivered to actuators afterward. These tasks must satisfy the real-time constraints. The sensors and actuators use wireless network to communicate with the controllers, while the controllers use CPU to handle the data and calculate the best decision.
To measure the Quality of Service (QoS) offered by a WNCS, we usually count how many tasks executed satisfy the end-to-end time constraints.
If we take the self-driving vehicle as an example, the typical tasks work as follows: the sensors (including cameras, radars and lasers) first perform the critical function of detecting the dynamic conditions surrounding the vehicle, and send the data to the controllers. The controllers (processors) handle data received from the vehicles' sensors and calculate the best decision. Finally, the decision (control signal) is sent to the actuators which physically perform actions such as changing gears, applying brakes, steering, and so on.
How to schedule these real-time tasks in WNCS to make every task satisfy the time constraints as well as to reduce the energy consumption is a key issue in WNCS. There are two kinds of resources involved: the network resource and the computing resource. Traditional approaches schedule the two resources separately. They either assume that the execution time of the control algorithm is negligible or consider it as a constant. Therefore, most of the previous works only focus on how to model the constraints of transmission conflicts, but do not consider the computation time of the control algorithms while enforcing the end-to-end deadlines.
This motivates us to study the Joint Resource Scheduling (JRS) model where two resources are allocated to the tasks, and the three segments of each task have precedence constraints. In this project, we model each real-time task as a three-phase job that consists of three segments: a sensing segment which transmits sensed data to the controller, a computing segment which computes the control decisions, and an actuating segment which transmits the control signals to the actuators. For each three-phase job, its computing segment cannot start until the sensing segment completes, while the actuating segment cannot start until the computing segment completes. Both sensing and actuating segments are processed on the network resource while the computing segments are executed on the computing resource. If all the three segments have arbitrary processing times, it is hard to get the optimal schedule in polynomial time. However, the processing times of the sensing and actuating segments may have some patterns because the amount of data sent to the controllers from sensors of the same type is the same which implies that these sensing segments have the same processing time. Similar argument goes for actuators with the same type. If the patterns of sensing and actuating segments are known, it is possible to design polynomial time optimal algorithms. Our goal is to decide whether all of these tasks can be scheduled while all the constraints are satisfied. If yes, then we want to find an optimal schedule to complete the tasks as soon as possible. Otherwise, we need to adjust the speed of the processors so that they can calculate faster but then the power consumption will be larger. How to balance the processor speed-up and power consumption is another key objective in our project. The proposed scheduling problem is one of the fundamental problems not only in WNCS, but also in other systems that present similar patterns. Any improvement of the joint scheduling method could increase the utilization rate of the system and therefore improve the system performance.
The main objective of the proposed research is to investigate whether tax avoidance based on the book-tax tradeoff theory is associated with and influences the choice of corporate acquisition payment methods in Merger and Acquisition (M&A) transactions. Specifically, we test the applicability of the book-tax tradeoff theory in a new context, that is, whether it can explain the relationship between tax avoidance and the choice of corporate acquisition payment methods. The proposed research will be the first attempt to explore this relationship.
Corporate acquisitions are often extremely large and important transactions. While research has shown that tax can be an important determinant of M&A activities, no research has been identified that investigates whether the level of tax avoidance as explained by the book-tax tradeoff theory relates to the choice of an acquisition payment method. Our proposed research intends to fill this gap in the literature. If a significant relationship between corporate tax avoidance and the choice of acquisition payment method can be established through the framework of a sound theory, the proposed research will have important implications for shareholders to assess their wealth, for corporations to allocate their resources and for tax authorities and public policy makers to formulate policies that address M&A transactions.
Since the launch of the Belt and Road Initiative in 2013 which necessitated business expansions and the industrial upgrading of Chinese companies, M&A activities have tripled in China. These activities involve mainly domestic entities, but the number of outbound M&As is increasing, and these can involve many Hong Kong entities. While the majority of the acquisitions involve non-SOEs from numerous industries, SOEs have also become more actively engaged in such activities in recent years. The increase in M&A activities by Chinese firms in overseas markets suggests that our research findings from China are relevant to other economies.
From the 18th to the mid-20th century, Chinese Pidgin English (CPE) was used to enable communication between Chinese and foreigners at different parts of China and Hong Kong. Despite its historical and linguistic significance, systematic research on CPE is still lacking. This project aims to investigate the origins of the grammar of CPE and construct a corpus comprising of data from diverse historical sources. The corpus will be the first of its kind, offering CPE data written in both the Chinese and English language. The quantity and diversity of sources make the corpus a valuable resource for generating new ideas on CPE as well as contact languages in general. This research focuses on the development and characteristics of two grammatical constructions in CPE: the light verb makee construction as in (1) and for as a complementation marker as in (2).
(1) make trade 'to trade'
makee die 'to die'
(2) must make go chop chop for see dat doctor man. (Tilden 1831-32: 764-65)
'(You) must go to see the doctor quickly.'
Both the light verb makee and the complementation marker for are common in pidgins and creoles; however, their origins still require more investigation, especially the light verb makee. This project addresses their grammatical characteristics and the languages and factors involved in their formation. It is hoped that results from this project can generate new discussions on the genesis of pidgins and creoles. Chinese Pidgin English is one form of the outcome of Chinese-English bilingualism. Results obtained from this project have potential to explain similar contact situations, such as Cantonese-English code-switching and contemporary English variety spoken in Hong Kong. This impact of this project can be manifested in different ways. The CPE corpus serves as an important resource for systematic documentation of the pidgin, paving the way for future projects. The availability of an electronic corpus also allows other researchers to access CPE data. The project could also cast new light on different facets of Chinese-English bilingualism.
At present, research studies on organizational effort in handling turnover problems and in retaining social workers have been lacking in Mainland China. The high social worker turnover is caused by the low social status and poor professional identity of social workers, poor management in social work organizations, and an imperfect social welfare system. Organizational career management in social work agencies, however, remains a significant yet little understood area. Organizational career management (OCM) refers to various activities employed by an organization to promote the career development of employees, develop staff's potential, and retain employees. OCM consists of different dimensions such as the provision of training courses, mentoring, supervision, job rotation, line management support, performance appraisal, and career planning. Therefore, this study is to explore how organizational career management can be theorized further as to how it can be extended or adapted for the social work context in Mainland China. A mixed-method approach will be adopted includes quantitative survey, in-depth interviews, document review, and site visits. The theoretical contribution will be identifying the missing components of organizational career management in the social work sector and suggesting the development of a theoretical framework for social work organizations.
The aim of this study is to explore the experiences of care providers in managing challenging behaviours of persons with intellectual disability and who are living in residential care services. A descriptive qualitative approach based on the use of non-structured direct observations and vignette interviews. Conveniently-selected residential care services for persons with disabilities from four non-government organizations will be approached. Maximum variation convenience sampling will be used to recruit care providers from the residential care services. The final sample size will depend on the numbers required to reach data saturation. Direct unstructured observations for each residential care services, and audio-taped recorded vignette-based interviews are the data collection methods to be used. Latent content analysis will be used to analyze the observation and interview data. Dedoose data analysis software will be used. The findings will provide data on the types, events and interactions that can trigger off challenging behaviours for persons with intellectual disability and residential care services. Responses of the staff and their management of different types of challenging incidents and events will also be described. This study acknowledges that care providers are an asset to any organizations and focused attention should be given to address their needs and concerns when they have to confront and handle a range of challenging behaviours of persons with intellectual disability.
The Four Masterworks of the Ming Novel (四大奇書, abbr. the Four Masterworks): “Romance of the Three Kingdoms”(“三國演義”), “Outlaws of the Marsh”(“水滸傳”), “Journey to the West”(“西遊記”), “Golden Lotus”(“金瓶梅”) were favourably praised as "Qi Shu"(奇書) for their “uniqueness”. The academia is not on the same page regarding the standard of their “uniqueness”, but as a whole the uniqueness is constructed by the common basis of the form (rhetoric) and the content (philosophical thought). In this regard, our project is planned to focus on the Four Masterworks’ content (philosophical thought) to elucidate their uniqueness and complexity which were unattended from previous studies. From a broader perspective, there is a relationship of mutual influence, development and construction between literature and philosophy, rather than literature is interpreted as reflecting the society, culture and thought. Therefore, this study will focus on the variation, breakthrough and construction of the philosophical thoughts being embodied in the Four Masterworks. Our study will further explore its continued development and influence on the philosophy of the Qing Dynasty.
In the past, the study on the philosophical thought of the Four Masterworks always emphasized the guidance or exhortation of Confucianism, Taoism and Buddhism, such as the concept of loyalty and righteousness(忠義), sunyata and rupa(色/空), transcendental and secular(出世/入世) as well as “Universal peace must be preceded by national harmony; national harmony, by familial solidarity; and familial solidarity, by personal integrity” (修身齊家治國平天下). In the other words, most of the studies in the past have been to reflect the traditional norms and values (“theory of reflection”) to conscribe the Four Masterworks - even in the form of “irony” or “sense of humor”. However, by scrutiny of the texts, we will find that the author(s) attempted to challenge the traditional values in order to subvert the tradition, anti-society and even anti-fate as a main axis through the works of the characters, the plot and the theme. It is not difficult to find the author(s) of the Four Masterworks to try to explore and construct values that are distinct from traditional ideas and values.
Although the main theme of the Four Masterworks are different from each other and discussed with the predecessors, we believe that the Four Masterworks could be penetrated by an axis, that is, anti-tradition. The content of anti-tradition is the “exploration of human desire and its moral signification” (“Satisfaction of human desire”), and this theme also continued and resulted in a great impact on philosophy and literature of the Qing Dynasty. Therefore, this study will employ this axis as a starting point to launch an in-depth and specific analysis of the following questions:
This proposal would gather, collate and analyze relevant texts such as original texts, commentaries and related research materials. The research results would be published in the form of papers and participated in academic conferences to exchange point of views with scholars. Finally, various papers would be published in a book to share the research achievements.
Nesfatin-1 is a newly discovered 82-amino acid anorectic peptide derived from nucleobindin2 (NUCB2). NUCB2/Nesfatin-1 immunoreactivity is highly expressed in hypothalamic as well as medullary nuclei such as the nucleus tractus solitarius (NTS) and dorsal motor nucleus of the vagus (DMNV), brain areas implicated in feeding and emesis control. Over the past decade, nesfatin-1 has been extensively studied in common laboratory animals because it is generally believed that the peptide not only reduces food intake and gastric emptying but is also involved in the long term regulation of body weight, making it a promising target for anti-obesity therapeutic drugs. Nevertheless, we know relatively little concerning the potential involvement of nesfatin-1 in emesis control because common laboratory animals (e.g. rat and mouse) are incapable of emesis and therefore the role of nesfatin-1 as a transmitter linking the forebrain and hindbrain being involved in nausea and emesis and feeding control has been overlooked.
The aim of the project is to use Suncus murinus (house musk shrew) to investigate if NUCB2/nesfatin-1 is a key transmitter in emesis and feeding control. We identified that S. murinus nesfatin-1 shares >85% amino acid sequence similarity with human and mouse. Our pilot study showed that central administration of NUCB2/nesfatin-1 potently reduced food intake and induced emesis in S. murinus. In the current proposal, therefore, we will identify the anatomical distribution of NUCB2/nesfatin-1 protein and mRNA using immunohistochemistry and in situ hybridization, respectively. We will also examine plasma and tissue expression of NUCB2/nesfatin-1 in the fasted and fed states. We will elucidate the role of NUCB2/nesfatin-1 in mechanisms of feeding, biomarkers of nausea, and mechanisms of emesis in S. murinus. Experiments will be performed using standard behavioural testing and established surgical and radiotelemetric techniques combined with measurement of brain monoamine neurotransmitters and immunohistochemistry. The studies will use a range of classical and novel anti-emetic drugs to determine the mechanism of action of nesfatin-1 to reveal the differential effect of nesfatin-1 in the regulation of feeding and emesis. Our findings will yield novel mechanistic insights into forebrain-brainstem communication.
Aim: The aim of the study is to explore the styles of leadership and management adopted by management staff, and how these styles influence front-line care providers' perceptions when they provide care and support to the older residents in residential care homes.
Design: A multi-site exploratory descriptive qualitative cum focused ethnographic study.
Settings: Conveniently-selected residential care homes for the elderly (RCHEs) from three distinctly different non-government organizations (NGOs) and independent private aged homes will be approached. Three care-and-attention homes, two contract/self-financing homes, one nursing home and two private aged homes will be selected, giving rise to eight homes that differ in terms of types of homes, resident capacity, staffing numbers and religious affiliation.
Participants: Maximum variation sampling will be adopted to recruit the staff members at two levels. Level one will be the management staff (senior staff) appointed to perform leadership and managerial duties and responsibilities. These include the superintendents and assistant superintendents or equivalent grades, social work assistant or equivalent, registered nurses, and health workers or equivalent (if in private aged homes). Level two will be the care providers (front-line staff) performing the daily care for the residents. These include enrolled nurses, health workers and personal care workers. The selection criteria for staff refers to the professional (social and health care personnel) and non-professional care staff who provide or influence the physical, psychosocial and/or spiritual care of residents; at least one staff representative from each rank; and, a willingness to engage in an individual interview with the researcher. Twelve staff members from each type of home will be recruited. Depending on the capacity of the eight homes, the preliminary calculation of the sample size will be 96 staff members. The final sample size will depend on the numbers required to reach data saturation.
Methods: Individual audiotaped semi-structured interviews and non-structured participant observations will be adopted for each home. Field notes will be collected to supplement the analytical descriptions of the interview data. Constant comparative analysis methods will be used to analyze the observational and interview data. Once the coding scheme is confirmed, a final analytical framework will result to describe the leadership and management styles adopted by management staff from across the homes, and how these styles influence the care providers' perceptions when they provide care and support to the older residents.
Results and Conclusion: The preliminary framework will delineate the leadership and management styles perceived by management staff to be appropriate and relevant for Hong Kong RCHEs. Findings will also examine how the management staffs' perceptions of leadership and management styles are translated into the work practices of care providers. Factors underpinning the practice of leadership and management styles, and how they influence the attitudes, behaviours and views of care providers looking after residents with different needs will be revealed. Specifically, findings will increase our understanding of Chinese styles of improving organizational effectiveness and staff relationships in RCHEs. Methodologically, interviews and participant observations are used to generate understanding of the phenomenon. The study also provides a glimpse of the situation in the NGO and private sectors by including all types of RCHEs in Hong Kong. Recommendations hope to identify ways to prepare people to be leaders and managers and hold supervisory/visionary and/or managerial positions in RCHEs; and, strategies to equip them and the staff to work creatively to achieve common goals in the provision of high standards of resident-family care. The ultimate output is to equip the staff working in the homes to operate effectively to satisfy the multiple needs of the residents, family and staff.
Upon taking office in 2013, President Xi Jinping called for a revamp of the Communist Party's approach to fight corruption and launched an anti-corruption campaign to pursue officials of all rankings within the Party. In China, it is a common practice for officials to sit on the board of companies. Taking advantage of this political connectedness, and together with corruption practices in some cases, these firms very often enjoy preferential business privileges over their competitors. We conjecture that the launch of the anti-corruption campaign will have direct negative impact on these political-connected firms, resulting in a much tougher operating environment for them. However, management's possible response, in terms of accounting practices, to the changed operating environment is subject to debate. One line of argument is that they will refrain from managing earnings in order to access external debt finance or avoid further political scrutiny. Another line of argument is that they will window-dress their financials to meet investors' or analysts' expectations. Our findings will help to address this issue and highlight the possibility that the anti-corrupting campaign may result in an unintended scenario of higher level of earnings management for these firms.
Childhood sexual abuse (CSA) refers to the involvement of a child in sexual activity to which the child is unable to give informed consent (Hong Kong Social Welfare Department, 1998). This study adopts a mixed method research with concurrent triangulation design to investigate the relationship of CSA with sexual self-concept among female adult survivors of CSA. Sexual self-concept generally refers to how individuals perceive themselves as sexual beings. It also refers to individuals' evaluation of their sexual feelings and actions (Buzwell & Rosenthal, 1996; O'Ssulivan et al., 2006).
This study aims to (1) present a comprehensive understanding of the sexual self-concept of female adult survivors; (2) describe the current level of psychological functioning (self-evaluation, social relationship, and psychological well-being) of female adult survivors of CSA; (3) examine the nature of the relationship between CSA, sexual self-concept, and psychological functioning of female adult survivors; (4) explore intrapersonal, interpersonal, and sociocultural factors affecting sexual self-concept and to investigate if there are any factors that act to intervene the effects of CSA.
The qualitative study of the mixed method research involves conducting in-depth interviews with 20 female adults who reported CSA experience. Participants of the interviews would be recruited through referrals from NGOs which provide services for female adult survivors; and through convenience sampling by posting recruitment note on campus, in church, in Facebook, and in emails. Semi-structured interview guide would be used to conduct the in-depth interview, which lasts for about 2-3 hours. The interviews would be recorded, transcribed, and analyzed by Fraser's method of narrative analysis. Results of the qualitative study would be integrated with that of the quantitative study.
A questionnaire would be designed for collecting quantitative data. It consists of objective self-report inventories of self-evaluation, social relationship, psychological well-being, and eight dimensions of sexual self-concept. The eight dimensions of sexual self-concept (viz., sexual self-esteem, sexual self-efficacy, sexual satisfaction, sexual anxiety, sexual depression, sexual fear, sexual motivation, and sexual-consciousness) are selected from a multidimensional sexual self-concept questionnaire. The selection was based on extensive literature review of female adult survivors of CSA, feedback from counsellors of CSA survivors, PI's previous research on female sexuality, and PI's pilot studies on female survivors of CSA. A targeted sample of minimum 80 female adult survivors of CSA would be gathered over a consecutive period of 1? years from sources similar to those of the qualitative study. All those who participate in in-depth interview would also be invited to fill in the questionnaire.
Qualitative and quantitative data are gathered separately yet concurrently. Data collected would be integrated in the interpretative phase of the study. However, throughout the research process, investigators of qualitative and quantitative study are in continual dialogue to ensure that data collection procedure is properly done and cross-validated.
Results of this study would present a comprehensive understanding of the sexual self-concept of female adult survivors of CSA and factors associated with CSA. The findings should have significant implications for training of professionals and for intervention of CSA. For examples, knowledge generated may enrich the current social work training curriculum content relevant to CSA. Information gathered is useful for enhancing frontline intervention strategies for CSA adult survivor. Findings would also help advocating prevention of CSA and promotion of well-being of survivors.
Recently, with the rapid development of computer techniques, computational studies have become an indispensable alternative in bioinformatics, healthcare informatics and cancer research. Among these studies, binding-affinity analysis for a receptor-ligand system, which closely relates to enzyme inhibitor potency estimation, is an appealing and important representative. However, current studies mostly focus on reversible binding systems and lack investigations of irreversible or hybrid ones. Major reasons include the evolved binding mechanism and increased modeling complexity of such systems. Besides, the higher computational complexity in system dynamics simulations and affinity calculations can also be a practical difficulty in these studies.
To tackle the difficulties regarding computational modeling, characterization and analysis of new receptor-ligand system structures, we initiate this project. In this project, we will rise to the challenges including simplification of evolved binding mechanisms, reasonable modeling and characterization of new systems, and efficient calculation and analysis of molecular binding affinities.
Our straightforward idea for system simplification and modeling is outlining a structure based on its atomic information and using building blocks to represent the irreversible binding-site variants. Computational characterization of the binding mode in a new receptor-ligand system will be subsequently implemented. In this step, we will develop novel feature-extraction methods relying on free energy calculation and decomposition, interface shape/volume estimation and binding-site motion analysis. Examples of such computational characterizations will be shown in this project, and these characterization alternatives will lead to efficient calculations of binding affinity in a new receptor-ligand system. We will combine the alternatives into an integrated software/program suite, which can encourage users to select different models based on specific requirements or within various application scenarios. Finally, a post-analysis will be conducted for a situation where a group or crowd of system variants are involved. We will utilize well-validated methods for supervised or unsupervised learning of these system samples, and will thus make predictions of newcomers or variants. This analysis will also join the software/program suite for multi-purpose analyses or studies. Moreover, the whole software/program suite will be adapted at the end of this project, to provide compatible interfaces with other broadly-used software of molecular modeling, molecular dynamics simulations and structural analysis.
With the success of tackling above challenges, the proposed project will encourage the development of cancer studies and structure-based drug design. Also, it will output useful publications and software/program suites that can benefit researchers and fresh users. In addition, this research project will provide valuable and practical experiences for undergraduates to improve their analytical/programming skills and to apply them to a specialized project or field.
Sentiment analysis aims to extract sentiment-related information from text automatically. Although most sentiment analysis is concerned with the detection of opinions from reviews, there is increasing interest in the affective dimension of the social network. The abundant social network data include comments on media published in YouTube, Flickr and Last. FM, in addition to discussions of politics, sport and the news in blogs. Analysing sentiment in this broad class of text is valuable because it can aid the discovery of sentiment-related patterns, such as gender differences and successful communication strategies. However, analysing social network data using traditional sentiment analysis methods is problematic for several reasons.
The first issue is the preprocess of short text which is prevalent in social networks. The second issue is how to deal with noisy labels. Due to the fact that it is hard to control the authenticity and quality on the social network, spotting potentially malicious users and detecting noisy labels are critical to sentiment analysis. The third issue is that emotion perception is very personal and could vary from one person to another, that is, different users may have different attentions, perspectives or abilities when labeling documents emotionally. Thus, sentiment analysis is not only text-related but also user-dependent. The fourth issue is that the sentiment embedded in social networks often has intrinsic dynamics. Capturing such dynamic characteristics of sentiments is critically important for the successful development of various social services, such as public opinion monitoring and social event detection. Last but not the least, the application of sentiment analysis in recommendation systems, stock price prediction and other domains deserves further research.
In light of these considerations, we firstly expand short documents in social networks by extracting the most similar words from the whole corpus. Secondly, we develop de-noising models to identify noisy sentimental labels. We consider the following two constraints in our models: (i) the simplicity of assigning the sentimental category to a document by users under any contexts, and (ii) the authority of a user in assigning sentimental categories to documents across various domains. Thirdly, we extract the sentiments from the social network comprehensively and accurately by jointly modeling the users, text and labels in the latent topic model. Fourthly, due to that the spread and evolution of an event will affect the general public's sentimental responses (including the category and intensity of sentiments), we propose an event-based framework of dynamic sentiment analysis. Finally, we explore the application of sentiment analysis in other domains.
This project is concerned with sentiment analysis based on multi-source social network data, which will not only tackle the challenging research problems triggered by noisy labels, varied user emotion perception and sentiment evolution in multi-source social network data, but also shed light on computational social science, natural language processing and other areas. To evaluate the effectiveness of our model, algorithm and framework, we plan to collect diverse social network datasets from Twitter, Sina Weibo, MySpace, BBC Forum and so forth. Based on sufficient social network data, user opinions could be sensed, evaluated, and even predicted in a certain circumstance.
The aim of the project is to use Suncus murinus to investigate the mechanisms of the differential roles of centrally located glucagon-like peptide-1 (GLP-1) receptors in emesis and feeding control.
There is strong evidence from clinic studies and from published pre-clinical studies using common laboratory animals that GLP-1 receptor systems are involved in nausea and emesis and in mechanisms regulating feeding. Thus, GLP-1 receptor agonists can be associated with nausea, emesis and reduced appetite in man. Common laboratory animals (e.g. rat and mouse) are incapable of emesis and therefore the link between a novel GLP-1 receptor being involved in nausea and emesis and feeding control has been missed. This project is based on our own original observations that administration of the GLP-1 receptor agonist, exendin-4, prevents feeding at low doses and at higher doses affects glucose homeostasis and induces emesis in both Suncus murinus and ferrets. We also showed that the GLP-1 receptor antagonist, exendin (9-39), antagonised emesis but it was ineffective in reversing the exendin-4-induced inhibition of food intake. In the ferret, exendin-4 also increases blood pressure and heart rate and decreases heart rate variability (HRV) and the dominant frequency of gastric myoelectrical activity (GMA) without affecting body temperature. The differential effects of exendin (9-39) on emesis and feeding suggest that exendin-4 may act via more than one type of receptor, or a "non-classical" GLP-1 receptor. If this is the case, the "non-classical" receptor may be responsible for the side effect profile of exendin-4, or other GLP-1 receptor agonists; the receptor may also be involved in mechanisms of nausea that are dissociated from those regulating the control of emesis.
In the current proposal, therefore, we will elucidate the role of exendin (9-39)-insensitive GLP-1 receptors (non-classical GLP-1 receptors) in mechanisms of feeding, biomarkers of nausea, and mechanisms of emesis in Suncus murinus. The studies will be performed using standard behavioural testing and established surgical and radio-telemetric techniques coupled with immunohistochemistry and optical imaging of a fluorescent imaging probe to assess effects on gastric emptying. The studies will determine if another classical GLP-1 receptor agonist, GLP-1 (7-36) amide, shares similar pattern of action with exendin-4 to induce emesis and inhibit food intake. The studies will determine the potential site of action of GLP-1 receptor agonists and GLP-1 receptor antagonists to modulate feeding and emesis relative to brain areas regulating vasopressin release and changes in gastric myoelectric activity, which are known to be altered during nausea and emesis. The In-vivo Xtreme imaging system, with a novel near-infrared (NIR) fluorescent imaging agent GastrosenseTM 750, will be used to permit a monitoring and quantification of gastric emptying rates in vivo and in real time.
As an Asian world city, Hong Kong is proud of being a multicultural society and officially describes itself as "open, tolerant and pluralistic". (Information Services Department 2010:29). However, this claim and self-image has not been fully echoed by many studies (Ku et al. 2004; Hewison 2004; Crabtree & Wong 2012; Law & Lee 2012) on immigrant integration in Hong Kong. Multiculturalism in Hong Kong has remained a descriptive term rather than a benchmark of achieving equal opportunity of rights and resources. With 6.4% of non-Chinese in the total population (Census and Statistics Department 2012), Hong Kong accommodates migrants of various ethnic origins and backgrounds. In recent years, we have witnessed the emergence of clusters such as "Little Thailand" in Kowloon City and "Little Indonesia" in Causeway Bay, with restaurants, fast-food shops, grocery shops, hair salons and remittance banks as some typical examples of ethnic economic activities. These activities have not merely enriched the city ethnoscape with a multicultural touch, but also suggested significant implications for integration of ethnic minority migrants. However, these implications have not yet been well-explored in Hong Kong.
This study is an exploratory research on this less touched upon dimension of social integration of ethnic minorities, namely the ethnic economy. It aims to understand the experiences of ethnic minority groups in business, and in particular, the significance of ethnic economy in the long-term social integration process of ethnic minorities in Hong Kong. More concretely, the study endeavors to obtain a general mapping of ethnic economies by identifying their patterns, distribution and tendencies. Secondly, it explores and analyzes experiences of ethnic entrepreneurs in Hong Kong. How do they make use of various individual and group resources for achieving economic goals? How do they experience empowerment and disempowerment in the process of overcoming structural constraints? Thirdly, it identifies areas for improvements in policy making and service provision. Emphasis will be put on the experiences of the disadvantaged ethnic minority groups (South Asians, South East Asians and other groups). Conceptually, it takes reference to the literature of ethnic economy, multiculturalism and immigrant integration, and adopts mixed embeddedness (Kloosterman, Van der Leun & Rath 1999) as its analytical framework. Methodologically, this study combines a baseline study with qualitative methods including field investigation, key informant interviews, semi-structured in-depth interviews and analyses of media materials and documents.
Contributions of this study will be twofold. Academically, by examining ethnic economy in Hong Kong in the regional and international contexts, this study contributes to the conceptual discussions on ethnic business in general, and in particular, ethnic business as a potential means to social integration in advanced Asian societies, which is still an under- researched area. Empirically, findings garnered will enrich our knowledge and understandings of ethnic minorities' life patterns and needs in Hong Kong. This will enhance the awareness of policy makers and practitioners on the significance of ethnic economy. Moreover, towards the goal of achieving a more inclusive multicultural society, the study will shed light on how policy makers and practitioners can improve by formulating measures and appropriate services for ethnic minority entrepreneurs in order to facilitate their social integration in the long run.
Level of educational attainment of the population is an important indicator of human capital development of modern society especially in what we called the 'knowledge-based' one. Education is also commonly assumed to be a legitimate and significant key to advancing one's social position. As an important social investment that is positively associated with the city's competitiveness, the Hong Kong Government has been spending considerable amount of money in education, accounting for an average of around 20% of annual total public expenditure. Amid the flourishing of degree and sub-degree programs all around the world, Hong Kong's post-secondary education has expanded much more dramatically. Among the senior secondary school leavers, the government aimed at increasing their post-secondary education opportunities from 33% in 2000 up to 60% in 2010. This target has achieved eventually and mainly, however, only by the introduction of two-year post-secondary sub-degree education, including the higher diploma (HD) programs and the newly developed associated degree (AD) programs, offered by self-financing higher education institutions in Hong Kong. With the increasing chances of educational attainment, social mobility of young people seems to be promising. Has it really been the case? What are the concrete developmental paths of the sub-degree program graduates grown up in a period of 'credential inflation' or 'diploma disease'? Is the self-financing sub-degree education in Hong Kong articulated to degree education, higher employability and income, and thus upward social mobility? What is/are the directions and possible intermediated factors affecting their social mobility (if any) after student's graduation? Do professionally accredited sub-degree programs contribute to better prospect of graduates than those graduated from programs without professional accreditations? What are the actual personal aspirations, family environment, mobility barriers/generators and school life experiences of the students studying sub-degree programs?
This research will be the first of its kind to examine the impact of self-financing sub-degree education on social mobility of Hong Kong's youth. Both quantitative and qualitative research methods will be adopted. Focus group method and research methodology is informed by the critical consciousness raising pedagogy suggested by Paulo Freire and the related critical narrative analysis. Questionnaire surveys and focus group interviews will be conducted with current students and graduates of self-financing sub-degree programs with and without professional accreditation offered by Caritas Institute of Higher Education. The research will access the actual impacts of the self-financing sub-degree education on social mobility with the considerations of different underlying and possible intermediated factors including credential inflation, social class and professional qualification. It will bring about theoretical concerns and teaching reflections on the expected role of educational attainment in the pursuit of a more open and equal society particularly in the years witnessing youth activism in Hong Kong.
Aim: The aim of the study is to explore family carers' experiences of current caring needs, the support they have received and their plans to continue to provide care for community-dwelling persons with mild and moderate intellectual disability (ID) as carers themselves are approaching old age.
Design: A grounded theory methodology using a constructivist approach.
Settings: The settings for this study will be conveniently-selected sheltered workshops from Caritas-Hong Kong and Yan Chai Hospital. The sheltered workshops will be selected from the Hong Kong Social Welfare Department's list of services provided to persons with ID. These sheltered workshops will be the sites used to recruit those family carers who look after persons with mild to moderate grades of ID. For Caritas-Hong Kong, two sheltered workshops providing services for persons with ID will be approached. These are Caritas Lok Hang SW (capacity 150) and Caritas Lok Kin Workshop (capacity 126). For Yan Chai Hospital, there is only one sheltered workshop named 'Yan Chai Hospital Madam Lo Lee Pui Ching Memorial Workshop' (capacity 150). Therefore, there will be a total of three sites, giving a capacity of 426 potential clients with ID in which to recruit their family carers.
Participants: Purposive sampling will be used to begin the recruitment of family carers, followed by theoretical sampling. Based on the principal investigator's prior work in conducting qualitative research in multi-sites, fifteen family carers are planned to represent each site, giving rise to a total of 45 family carers. At this point without undertaking the study, it is not possible to know whether the capacity of the sheltered workshops may influence the phenomenon under study - the family caregiving experiences. Indeed, these sites are larger in capacity than the sites used in the principal investigator's prior study to estimate the sample size. Furthermore, as the actual number of clients with ID who are mild and moderate cannot be determined at this stage this will therefore influence the number of family carers that will be recruited. As a contingency measure, this project will target at sampling 20 family carers per site, giving an estimated number of 60 family carers. Therefore, the final sample size of this project will depend on the numbers required to reach data saturation.
Methods: Individual audiotaped semi-structured interviews will be adopted. An interview schedule composing of broad questions has been developed based on the search of the literature and the research team's prior knowledge and expertise of conducting research with persons with ID and on family carers. Field notes will be collected to supplement the interview data. Constant comparative analysis methods will be undertaken to generate the concepts and to develop the theory of family caregiving experiences, and to identify the processes of the ageing carers' planning for continuation of care for their family members with ID.
Results and Conclusion: The theory will capture family caregiving experiences and the processes of family carers in addressing caring needs, the support received and the family carers plans to continue to provide care for the community-dwelling persons with ID in their later life. New understanding and insights into emerging issues, needs and plights of family carers will be made available to inform the policies and practices to improve the care of the persons with ID living in the community, as well as to provide better support for the family carers who are ageing themselves. The theoretical framework that will be generated will be highly practical and useful in generating knowledge about factors that influence the caregiving processes; and, tracking the caregiving journey at different time-points to clearly delineate areas to implement practice changes. In this way, the theoretical framework will be highly useful in guiding timely and appropriate interventions to target at the actual needs of family carers as they themselves are ageing and will need to continue to take care of their family members with ID in the community.
Applying game concept and Virtual Reality (VR) technology for healthcare education and medical training is becoming popular and widely accepted in practice. These medical related serious games are attractive and prove to be cost effective in motivating learning. Many healthcare training systems share a common goal in requiring trainees to accomplish specific tasks with strong two-hand interaction and hand-eye coordination. These tasks usually look simple, but repeated practices are necessary in order to achieve certain skill level. Otherwise, it is difficult for the trainee to perform the task properly, and this can easily lead to risk in hygiene or even injury. Existing VR-based training systems require tailored devices to provide realistic interaction with the virtual environment, or some others may require wearing of gloves for recognizing hand gestures. These solutions commonly incur high development cost, and the inconvenience in wearing skin contacting devices. For many of the training applications, a fast vision based non-contact hand recognition solution will be more suitable and welcomed, as the solution provides a more comfortable setting to users and require shorter setup time.
Therefore, in this project, our major objective is to enable the development of healthcare related serious game systems with the use of vision based techniques, so that tasks requiring two hands cooperation and interaction skills can be automatically and properly evaluated. A wide range of related healthcare skill training applications can take advantages from the proposed recognition system including training on personal care, rehabilitation, massage, and exercise. Specifically, these training systems can correct and guide the trainees with proper hand gestures when performing certain tasks, and keep practising without the involvement of human coach. Moreover, with the elimination of skin contacting VR devices from the recognition procedure, the concerns for hygiene and long setup time are also reduced.
Our preliminary idea to the solution of the two-hand recognition is to rely on multiview vision based tracking techniques. In recent years, vision based tracking technology is becoming popular. Many attempts try to recognize hand gestures for simple control purposes, such as pointing and dragging operations in the graphical user interface (GUI). Many of them were based on the conventional real-time depth sensors to produce depth map of hand, followed by analyzing hand structure for fast and robust tracking of finger and palm postures. However, it is still challenging in handling highly coupled two-hand interactions. The major difficulties lie in the occlusion problem between two hands, especially when they are close or even touching each other.
To reduce the effect of occlusion, multiple depth cameras are being developed to cover a wider viewing directions. A fast and tailored algorithm in processing the several obtained depth maps or point clouds will also be developed. In brief, the point clouds are first unified by hardware accelerated registration methods. Then, a quick hand parts segmentation will be proposed for the extraction of features used for identifying palm and finger gestures. Finally, the recognition is accomplished by the use of classification or machine learning algorithms on a large set of sample hand gestures obtained from volunteer subjects in laboratory.
At the end of our project, we will demonstrate the capability of the hand gesture recognition system with a serious game in the theme of healthcare. The application will provide repetitive training and practises on tasks requiring skills using both hands which yield instant feedback and suggestion to improve the related skills.
Teaching and learning of computer programming in beginners' classes are known to be difficult. Students have to do a lot of exercises to practise their programming and debugging skills, and they need feedback on the correctness of their programs. However, assessing students' programming work manually is tedious, time-consuming and error-prone. With the large class and diverse background of students nowadays, it is increasingly challenging for instructors to provide students with adequate appropriate exercises and individual feedback, which are vital to the effectiveness of students' learning and sustenance of their learning motivation. In response, universities worldwide have developed automated program assessment systems (APASs), which free up instructors' time for other un-automated educational tasks. Moreover, APASs are found to be of tremendous benefit in many other aspects, such as facilitating the design of effective pedagogy, provision of instant and personalized feedback to both instructors and students, and enhancing students' learning motivation.
One core function of APASs is to assess the correctness of students' programs, typically by automatically executing them against a suite of pre-defined test cases and comparing the programs' actual outputs with the instructor's expected outputs. The latter task requires a test oracle, that is, a mechanism for determining the correctness of program outputs. In the field of software testing, the general problem of test oracle automation is well known to be challenging. Implementation of test oracles in existing APASs is often too simplistic, rigid and incapable of being tailored to support the intended educational outcomes of the exercises. For example, multiple correct (or admissible) programming solutions to an exercise may produce different outputs (called output variants). A program which the human instructor accepts to be correct (or admissible) could be inappropriately rejected by a rigid test oracle in an APAS. This technical limitation is common and has been a root cause of many educationally undesirable effects on teaching and learning that can substantially compromise the benefits of an APAS in practice. There is a clear and pressing need to address this limitation, which is said in a recent literature review to be the main disadvantage of using APASs.
In this project, we will develop an automatic program assessment framework which can recognize a variety of admissible output variants so that both instructors and students can focus on the essentials of the exercises instead of minor output deviations. The framework will satisfactorily handle a much broader range of programming exercises to suit different educational needs. We will build an online APAS platform to implement our framework that highly automates the program assessment process for use in real programming courses. The platform will also be instrumental for empirical evaluation and validation of our research. To fully utilize our expertise and collaboration, we will conduct cross-institutional experiments for higher reliability and validity.
Our study is significant in perfecting the state-of-the-art APAS technologies and the advancement of the teaching and learning environment for programming classes, as what the work of ours and others has previously contributed. The research is expected to benefit the learning of hundreds of students each year with visible impact in the form of their improved programming skills. This project will advance the research capability of the investigators so that they can transfer research experiences and new knowledge into teaching and learning in their institutions, which is also the main objective of this Faculty Development Scheme (FDS). Since the newly developed online APAS platform can be used by all academic staff of the institutions, the new knowledge and tangible outcomes arisen from this project will benefit the teaching and learning of all three institutions. Finally, we will actively publicize our research findings both locally and internationally so that their potential benefits and values can be deployed by the global computer education community.
Worldwide smartphone and tablet shipments will surpass two billion in year 2015. Powered by batteries, these mobile devices are in desperate need of higher power efficiency to extend usage time. The display is often the most power-consuming component of a mobile device. Traditionally, the majority of power consumption can be attributed to a Liquid Crystal Display (LCD) panel, which can take up about 60% of the total power consumed by a mobile device. Recently, Organic Light Emitting Diode (OLED) has emerged as the choice of display, given its advantages of higher emitting efficiency, bigger contrast ratio, brighter colors, and the possibility of being built on a flexible and transparent substrate. Although OLED is more power efficient than LCD, it still consumes quite a lot of power in mobile devices. We tested the OLED display power of a Samsung Galaxy S4, which in standby mode occupies approximately 43% of the total power consumption. One unique property of OLED that differentiates it from LCD is that OLED power consumption is highly color-dependent. This property will be exploited in our project to significantly improve OLED display power efficiency. Specifically, a hierarchical collaborative optimization framework across three layers will be developed.
First, at the device-level, this project will develop a unified power model for OLED cells capable of capturing both color-dependent power consumption and dynamic power from color transformations between adjacent frames.
Second, at the hardware-level, as OLED cells allow trade-offs between power and color, visual-quality aware dynamic voltage scaling (DVS) will be developed for different applications.
Third, at the software-level, as different colors consume different amounts of power, the dynamic tone mapping (DTM) approach will be studied for videos to save display power through video classification and color transformation.
Through exploiting hardware-level power trace information, a novel software-level online video classification technique will be investigated. The most challenging aspect of this project is to adaptively conduct the proposed techniques in a time-conscious and effective manner considering stringent timing requirements for applications such as video streaming and games.
The key is the power model, which integrates the hardware-layer and software-layer to overcome this challenge. Preliminary results re-confirm the huge potential of OLED display power reduction delivered by tri-layer optimizations of OREO.
The principles of power saving for OLED display that will be obtained from this project will also benefit the software development training for undergraduate students whose majors are digital entertainments. They can embed the power saving idea in the design of mobile applications, and eventually build power conscious applications.
With the success of this project, we aim to cut OLED display power by half, which will help reduce power consumption of mobile devices, extend usage time, and contribute to a better user experience, as well as a greener planet.
Helping language teachers to improve their quality and standard of marking has been a perennial problem confronting trainers of student teachers for many years. Besides attending to heavy work duties, language teachers are faced with the burden of rating and marking student assignments and papers which is an added pressure to an already stringent schedule. This also affects the quality of marking and students become the ultimate losers in the long run.
The writer proposes to explore the composition marking behavior and uniqueness of Chinese language teachers in two secondary schools in Hong Kong. The research team will be stationed in the school during the research period to examine the school culture and group communities (e.g. school policy on marking; views of the Principal and the Panel Heads on teachers' attitude towards marking compositions; students and their parents' opinions on who is responsible for marking) and their influence on teachers' marking behavior through onsite observations of these teacher subjects from classrooms, staff rooms, conference rooms through to informal occasions with respect to their marking behavior, interaction and mutual influence processes; and through multiple, in-depth interviews to understand how they formulate and mode their individual marking principles and strategies under the effects and challenges of community group expectations and extant school culture. Think-aloud protocols, though time-consuming by nature, will be used to elicit teachers' verbalized judgment on student writings instead of relying solely on the textual analysis of composition written by students which can be seen as the most remarkable feature which distinguishes the current project from earlier works of similar nature using the product-oriented approach.
This project will be the first ethnographic study of the marking behavior of Chinese Language teachers in the Hong Kong classroom context from the perspective of contextual analysis. Results of the project will help better understand the judgment of frontline Chinese language teachers in determining what to use or not to use in their marking strategies, how they arrive at the final score and factors influencing their marking behavior and scoring decision. Possible suggestions will be provided on how the marking quality of Chinese language teachers could be further improved and upgraded.
3D visualization and printing technologies has emerged for decades. The advantages of employing 3D technologies for all kinds of business and industries including e-commerce, e-marketing, healthcare, education, and manufacturing are obvious. However, the popularity of 3D content on the web is still low nowadays. One of the major reasons is the lack of low-cost solution to construct high quality 3D contents. Moreover, high development cost for 3D visualization and manipulation platform also affects the popularity of developing 3D related website. All these factors prohibit normal web users to create 3D content in a DIY (Do It Yourself) manner.
Therefore, in this project, we will tackle the difficulties related to the cost and quality of 3D model creation, modification and customization on the web environment. The project will address three challenging areas of problems; they are the low-cost depth camera-based 3D data acquisition, fast and high quality 3D mesh denoising, and collaborative 3D content manipulation and customization under the web environment.
Our preliminary idea for 3D data acquisition assumes that consumer level depth camera is employed for capturing 3D surface points of an object, as cost for these devices are becoming more affordable right now. A post-processing step will be introduced for combining 3D points from different views to form a resultant 3D surface model. However, these reconstructed models commonly suffered from serious artifact or noise. As a result, our project will invent novel algorithms for fast 3D mesh denoising, so as to maintain high quality ready-to-use 3D contents for various web-based 3D applications. Finally, to fulfill the needs from na?ve users on modifying and customizing 3D contents, we will develop useful and user-friendly operations for manipulating 3D models on the web. Popular advanced operations like cutting, union, and deformation will be improved with intuitive user interface design and able to work collaboratively with other users under a distributed environment. Some visualization or 3D printing specific operations will also be proposed and included in our project.
To facilitate the evaluation of our proposed methods, we will implement and modulize these methods into software development packages at the end of the project. Moreover, we prepare to demonstrate the practical value and potential benefits of our solutions with a web-based 3D application. One of the possible applications is the development of a comprehensive platform for 3D printing objects with the capabilities of searching, sharing and customization on the web. By solving the above mentioned challenges, we believe the tangible outcomes (publications, algorithms and software packages) from the proposed project should be directly applicable to the community to develop web-based 3D applications with minimal effort. Moreover, the research project will provide invaluable chance in developing teachers and students the necessary skills and capabilities related to the latest 3D graphics technologies which is included in the curriculums and programmes offered by the department and the institute.
Advances in Massive Online Open Course (MOOC), web 2.0 communities, social media, mobile and sensor technologies have been phenomenal in recent years. Worldwide, there is a significant proliferation of learning resources with multimodalities such as online course platforms, lecture videos, and learning materials (e.g., web pages, animations and documents). Meanwhile, the number of people who have access to mobile technologies has also grown exponentially. These developments provide people with more opportunities which help them not only learn new knowledge and skills but also communicate and discuss with other users, sometimes even with instructors and tutors of the courses.
These emerging technologies for on-line learning, on one hand, bring more fruitful learning resources and human interactions. Yet on the other hand, they make more difficult for learners to find their desired learning resources effectively and efficiently when confronting with such a large volume of learning data. To assist learners to find their desired learning materials and suitable virtual classes, it is essential to manage and organize information about learners as well as various learning resources. A main stream solution is to construct user- and resource- profiles, e.g., a bag-of-categories (BoC) or the category tree (Tree-based), so as to facilitate personalized learning services. However, the static nature of categories in profile brings forth two problems to be resolved in e-learning applications. First, it is not unusual that existing data are insufficient to construct a powerful learner profile, since only those categories used by the learner are taken into consideration during the process of profiling her/him. Second, not all categories in learner profiles are useful or necessary to be taken into consideration in different contexts.
To tackle these problems, it is paramount to understand well not only learner behaviors but also the hidden relations among their interested resources and current contexts. As revealed by recent social network studies, user behaviors are highly influenced by his/her neighbors who tend to share similar patterns in behaviors and common opinions. Therefore, we plan to propose a hybrid profiling approach to aggregating the multiple hidden relations, such as pre-requisite relations, content relations and social relations, so as to enrich the learner profile with valuable information from his/her neighbors or some potential interested learning resources. Furthermore, in this project, we also plan to devise an explicit context model to fully exploit the value of learner profile so that the problem of static learner profile can be resolved and context-aware learning services and applications can be supported. Specifically, we will classify contexts into three different layers, which are personal-level, group-level and class-level learning contexts. Each layer of contexts may have different effects on the learner profile. To demonstrate the validity and evaluate the effectiveness of the proposed framework, we plan to apply the hybrid learner profile and context model to a suite of applications in three different layers (i.e., personalization, groupization and classization) for the e-learning systems, including personalized courseware recommendation, group member discovery, and class content pruning (or augmenting).
The proposed project aims at advancing policy formulation, research, teaching and learning, as well as service development, to support informal caregivers for frail elders and persons with disabilities in Hong Kong. Informal caregivers are usually family members, relatives or friends, who provide some form of ongoing and unpaid care to another person who requires assistance in daily tasks due to their frailty, cognitive and/or physical impairments. The Labour and Welfare Bureau (LWB) has commissioned a consulting study to explore the needs and services required of these informal carers, and the government has indicated its readiness to implement the recommendations. Hong Kong does not have a caregiver policy per se, thus leading to a lack of systematic planning, as well as the review and allocation of relevant resources in caregiver support. At the same time, additional analyses are needed to identify the needs of carers from various backgrounds (e.g., those in different age groups or varying stages of life span development, or who serve people with different impairments, etc.,) and to assess the effectiveness of various initiatives. At present, there are quite a number of initiatives regarding caregiver support financed by charity foundations and carried out by non-governmental organizations (NGOs). However, they are usually on a small scale and focus on a particular district or caregiver group.
This project will thus organize an international conference on caregiving policy and practice from international perspectives, two experience-sharing workshops (including one on the theme of at-risk caregivers and the other on mobilizing family and community support) and a seminar on sharing research projects in addition to research idea incubation for staff of the Caritas Institute of Higher Education (CIHE) and Hong Kong Shue Yan University (HKSYU), as well as consultation sessions to be provided by invited experts via Zoom after their departure from Hong Kong.
Presentations by overseas experts on exemplar caregiver policies and practices in their countries or in the global context as well as experts in economics and health sector will provide important insights for the Hong Kong Special Administrative Region (SAR) government in reviewing the need for policy formulation, as well as for academics, researchers and other stakeholders contributing to caregiver support through research and education. At the same time, this project will also invite the sharing of innovative initiatives in other economies and those financed by charity trusts and carried out by NGOs in Hong Kong, so as to provide inspiration to the welfare sector and human service professionals (such as social workers and nurses) to consider upscaling and enhancing these initiatives. Scholarly exchanges with international and local experts will encourage academics to consider how caregiving issues can be incorporated into the training of human service professionals such as social workers, nurses, occupational therapists and physiotherapists.
All these activities will contribute to promoting caregiving support through policy formulation, research studies, education and training, as well as service provision. At the same time, researchers, educators and students in higher education in Hong Kong in general, and those at CIHE and HKSYU in particular, will be inspired to incorporate the theme of caregiving support into their various scholarly activities.
Toy computing is a study on design, computation, prototyping and integration of various toys using computing techniques. Recently, toys are becoming more intelligent and penetrating to gadgets such as smart phones, digital cameras, and motion sensing devices. The player can communicate, interact, or collaborate with them to achieve some tasks. Also, modern toys are closely related digital entertainment. Emerging computing techniques such as Augmented Reality (AR) has become popular for developing creative toys. To promote the studies of Digital Entertainment and Toy Computing to the Hong Kong public, we propose to organize a technical event consisted of a series of seminars (public lectures with panel discussions) and hands-on workshops covering a great variety of creative software and hardware techniques include mobile app development, and creative electronic design, etc. The target participants are toy developers, toy manufacturers, academics in the related fields, safety experts, and the students who are seeking jobs in toy/entertainment industries, etc. After the proposed activities, we expect the participants could gain basic knowledge and self-learning resources for producing creative toys and digital entertainment outputs.
Our project benefits toy manufacturers and developers and digital entertainment producers. In recent years, the toy industry is turning from pure manufacturing to creative design marketing. According to the Hong Kong governments' fact sheet, creative industries (include toy manufacturing) have more than 192,000 practitioners and they contributed nearly $90 billion or 4.7% of Hong Kong's Gross Domestic Product (GDP) annually by November 2013. It is expected to have more people joining this industry because many high-ended toy manufacturers are moving their design base or even the production line back to Hong Kong. However, there is a lack of talented and skilful young people joining this industry. In order to provide enough human resources to local toy manufacturers, the proposed event aims at enriching the toy developers' skills and guiding them to apply creative ideas to toys through computing techniques. The event will serve as a platform for academics and manufacturers to interact and initiate new collaborations. Moreover, it benefits to safety experts because they will learn the state-of-art trends of cutting-edge computing technologies that are useful for them to examine whether these new toys are safe enough for users, especially small children.
Our project also benefits academics and the students who want to join in the toy industry. The government intended to nurture Design and Digital Entertainment experts; so in the 2014 Policy Address, the Chief Executive introduced a new Theme-based Subsidy Scheme (TBSS) (now it is a.k.a. Study Subsidy Scheme for Designated Professions/Sectors (SSSDP)) to subsidize up to 1,000 students per cohort to pursue designated full- time locally-accredited self- financing undergraduate programmes such as in this discipline, starting from the 2015/16 academic year. The theme of this seminar and workshop series is aligned with the government's policy, and the outcomes of the seminars/workshops and research experiences can be transferred into teaching and learning of the newly introduced BSc. (Honours) Digital Entertainment programme at Caritas Institute of Higher Education (CIHE). The academic staff of different institutes could meet at the event and have a chance to initiate new research collaborations. The event will have a Project Competition for students (mainly undergraduate) to demonstrate their Research & Development outputs to the public, and they may receive useful suggestions from technical/design experts or meet potential employers. In the long run, we intend to consolidate students, researchers (especially the computing and creative media experts), and local inventors to form a community, and promote the innovation and R&D for the toy and digital entertainment industries in Hong Kong. Also, we will make use of the experience to support and enhance the teaching and research in CIHE and OUHK, and strengthen the research collaboration between different tertiary institutes.