Next Generation Resilient, Dynamic, Ultra-High Capacity Core Optical Network based on Optical Flow Switching
Hong Kong Principal Investigator : Prof Kwok Wai CHEUNG (The Chinese University of Hong Kong)
Mainland Principal Investigator : Prof Anshi XU (The Peking University)

This proposal investigates a new approach to build up the next generation of ultra-high capacity core optical network. Current networks employ huge electronic IP router switches that require tremendous energy to operate which is becoming a big problem for our environment. They are also susceptible to various kinds of malicious attacks. With the rapid increasing need for bandwidth capacity in the core network especially by the peer-to peer and video applications, the available capacities of service providers are quickly running out. In addition, there is a very visible paradigm shift in the usage pattern from telecommunication to data exchange. The percentage of data traffic in the core network increased from a few percents to over 97% over the last ten years. As data traffic has a less stringent delay latency requirement than the real-time continuous traffic in telecommunications, there is a great need for a new paradigm to engineer the next generation core network that is resilient, robust, dynamic, scalable, cost-effective, low in energy consumption, and supports hundreds of terabits per second of traffic.

The approach we propose to take is to combine the Generalized Survivable Network (GSN) concept recently proposed by the CUHK PI and the Time-Space Label Switching Protocol (TS-LSP) recently proposed by the Peking University PI onto a single framework to tackle the research issues for optical flow switching (OFS) networks. In OFS, only optical cross connects are used for the switching function like a piece of transparent glass with no electronic switching or processing. Thus the optical switch fabric can be very cost-effectively built, and the energy consumption can be extremely low. OFS is suitable for high bandwidth data applications such as HDTV, IPTV, video-on-demand, data center exchanges and various grid applications.

A very important advantage of combining OFS with GSN and TS-LSP is that the resource reservation problem becomes very straightforward because GSN guarantees that all dynamic traffic demands can be accommodated. There is no need to compute and locate where the resources are when a certain flow request comes in. This can significantly improve the overall throughput and reduce the delay. At the conclusion, we could better understand the feasibility of OFS for the next generation core network technology, resolving most problems involving the scalability, distributed routing protocol and resource contention resolution, signaling, survivability and service characteristics versus various applications. This research could lead to a revolutionary development for the next generation core network.

Computer-Assisted Precise Hepatectomy Research Based on Recurrent Neural Networks
Hong Kong Principal Investigator : Prof Pheng-Ann Heng (The Chinese University of Hong Kong)
Mainland Principal Investigator : Prof Yi Zhang (The Sichuan University )

China is one of the regions with the highest incidence of end-stage liver diseases in the world. Due to the enormous patients with chronic viral hepatitis B, liver cancer or other chronic liver diseases, a large amount of hepatectomy is required. With the development of medical imaging and 3D reconstruction techniques, precise hepatectomy is promising to offer new hopes for hepatectomy, especially for highly risky liver transplantation.

Precise hepatectomy requires accurate evaluation of the patient-specific liver images before surgery, which includes locating the position of lesions in the liver, understanding the tissues adjacency and vessel distributions, and estimating the volumes of removal and remnant liver segments. However, there are quite a number of challenges in precise liver image computing, e.g., the separation of liver parenchyma and adjacent tissues, the identification of liver lesions, the inter-subject variability in liver size and shape, the low contrast between vessels and surrounding tissues, and the presence of image noise. The existing shape-based and optimized intensity-based methods are not capable of precisely assessing liver images.

The development of recurrent neural networks (RNN) provides new approaches to solve the above-mentioned problems. With the properties such as multistability, continuous attractors, and permitted and forbidden sets, RNN has been found more powerful than the forward neural networks. In addition, as region features and edge features in medical images can be integrated by RNN through the principle of feature binding, RNN is a suitable technique to meet the challenges in processing complicated liver images. The image processing problem such as segmentation, registration and fusion can be transformed into an energy minimization problem in RNN. RNN can also be used to simulate the spread of potential energy among the mass points, which provides a new way for deformation simulation of soft tissues.

Based on the requirement of precise hepatectomy, at the same time taking the advantages of RNN in medical image processing, we would propose a series of novel research directions to solve the key issues in computer-assisted precise hepatectomy, such as liver tissue segmentation and precise volume measurement in CT and MRI images, atlas-based segmentation of hepatic vessels in CTA and MRI images, segmentation of liver lesions in CT and MRI images, registration and fusion of liver tissues, patient-specific liver modeling and segment-oriented liver resection, deformable simulation of soft tissues, and GPU-enabled parallel neural network implementation.

Research in Interactive Spoken Dialogs to Support Secondary Language Learning of English by Native Putonghua and Cantonese Speakers
Hong Kong Principal Investigator : Prof Helen Meng (The Chinese University of Hong Kong)
Mainland Principal Investigator : Prof Jia Liu (The Tsinghua University )

This project aims to develop an interactive dialog system to assist second language (L2) acquisition of English by adult learners whose primary language (L1) may be Putonghua or Cantonese (i.e. two major dialects of Chinese). English is the lingua franca of our world. It has been estimated that by 2010 there will be 2 billion English learners worldwide, and the proportion in China is approximately 380 million. It is of prime importance that we acquire communicative competence in English. Second language learning, specifically pronunciation learning, involves correct perception and production of sounds in the target language. The learning process tends to be influenced by the transfer of L1 features in L2 productions. Negative transfer leads to pronunciation inaccuracies and errors in the second language. These inaccuracies tend to fossilized with age and present specific challenges to adult L2 learners. Chinese has stark linguistic contrasts in comparison with English. We often observe negative transfer effects in the L2 English productions of L1 Chinese learners. Pronunciation improvement requires persistent practice in productive and perceptual training. In order to support productive training (i.e. eliciting speech from the learner for assessment), we propose to develop discriminative automatic speech recognition techniques that enable detection and diagnoses of targeted pronunciation inaccuracies (i.e. mispronunciations) due to negative language transfer effects. It is important to note that some segmental mispronunciations are more serious than others, from the communicative perspective. Mispronouncing the discriminating phone(s) in a minimal word pair will lead to word confusions. Contrarily, other phone confusions may not impede communication. Also, having a larger number of mispronounced phones in a word is generally a more serious error than having fewer. In general, an elementary learner tends to produce a higher number of pronunciation inaccuracies when compared with an advanced learner. Productive training should help elementary learners focus on more serious errors, rectify them and then move forward. This is preferable to flagging the high percentage of mispronunciations in their speech, which tends to discourage learning. Advanced learners tend to produce fewer mispronunciations and productive training should focus on common and subtle errors to help perfect their L2 pronunciations. To support the pedagogical needs of learners with different levels of proficiency, we propose to derive a gradation of mispronunciations (from subtle to serious) from a corpus-based study. The gradation will facilitate development of a dialog model that can adapt to pedagogical needs of learners at different proficiency levels. The dialog model will invoke generation of relevant, corrective feedback that is specific to the detected and diagnosed mispronunciations. In order to support perceptual training (i.e. developing the learners' skills to accurately discriminate among sounds of the target language), we propose to develop automatic response generation that provides multimodal visualization (e.g. through the use text-to-audiovisual speech synthesis) of the speech production process, intended as helpful instructions that guide pronunciation error correction and improvement. We believe that the development of such interactive spoken dialogs can effectively support Chinese learners of English, especially for certain (remote) regions in Greater China which lack qualified English teachers.

Privacy-Preserving Data Networks Publishing
Hong Kong Principal Investigator : Prof Lei Chen (The Hong Kong University of Science and Technology)
Mainland Principal Investigator : Prof Hong Mei (The Peking University )

Social network applications, such as Hi5 (hi5.com), Facebook (facebook.com), and Myspace (myspace.com), have become popular for sharing information. As a consequence, the amount of social network data has grown rapidly, and this offers rich opportunities for data mining and analysis, for example, to find community groups and their evolution [1, 2]. However, social network data usually contain users' private information; it is important to protect this information in any sharing and mining activities. Quite a few proposals have been proposed to protect privacy of published data networks. However, these works suffer from the following limitations : 1) No model has been proposed for modeling the background knowledge, which leads current proposals to address only one or two types of attack; 2) Simple combined solutions (combining methods proposed for attribute and structural attacks) to protect the privacy of labeled data networks lead published data networks to be less useful. Attribute attacks [16] refer to re-identifying individuals by matching the individual's publicly known attributes with the attributes of the anonymized table, and structural one [15] try to identify entities through learned network structural information; 3) No approach has been proposed to address privacy breaches in dynamic and distributed published data networks.

Thus, in this project, in order to utilize the information provided by the data networks, and meanwhile, to protect the privacy of these data, we plan to develop methods to realize privacy-preserving data network publishing. Specifically, first, we will model the background knowledge of adversaries by mining data networks and design a framework to guarantee privacy under any structural attack according to derived background knowledge. Second, we will develop a mechanism to protect the privacy of labeled data network, which considers possible privacy breaching caused by both attribute and structural attacks. Third, we will design methods to avoid possibly privacy breaching from dynamic or distributed published data networks. Finally, in addition to theoretical analysis, we will also build a test platform to study the performance of the proposed approached in publishing data networks. With our established expertise and extensive experience, we are confident that we can successfully complete this project and achieve the objectives.

Heterogeneous Transfer Learning with Applications to Web Data Mining
Hong Kong Principal Investigator : Prof Qiang Yang (The Hong Kong University of Science and Technology)
Mainland Principal Investigator : Prof Yong Yu (The Shanghai Jiao-Tong University )

In real-world machine learning and data mining, we often face the problem where the labeled data are scarce in their own feature space, whereas there are sufficient labeled data in other feature spaces. For example, there may be few labeled images available, but there are often plenty of labeled text documents on the Web (e.g., through the Open Directory Project, or ODP, (http://www.dmoz.org/). With the dramatic increase of the real-world data, heterogeneous data, such as text, multimedia, sensor signals and biological sequences, are also increasing at an amazing rate. It is becoming a necessary but relatively untouched area to apply machine learning models that are trained on heterogeneous data to domains where relatively little high-quality training data are available. In this area, we have identified three serious issues with existing machine learning algorithms on heterogeneous data. First, there is a lack of high-quality labeled data for training models in new domains, and data are expensive to obtain. Second, most current machine learning methods only work on homogeneous data but not on heterogeneous data between training and testing. Third, it is difficult to reuse existing data of one type to train models in others. We observe that this learning problem across heterogeneous data is difficult, but there are now growing resources available that can help bridge the gap between feature spaces. When the feature spaces are different, traditional learning methods can only achieve low performance. But as the social Web, such as Facebook and Flickr, are becoming increasingly popular, with mixed-feature pages embedded with images, text, and videos, these pages can be used to link different feature spaces.

In this project, we propose the following solutions for the above challenges. First, we will study the relationship between different domain data and uncover the transferable knowledge across heterogeneous data. Second, we will develop novel distance measures for guiding when to transfer knowledge between different feature spaces. Third, we will develop heterogeneous transfer learning algorithms to perform knowledge transfer between different domains, which can handle large-scale data as well. In addition to the theoretical analysis, we will also develop a prototype system and apply it to several important problems, including transferring knowledge between different vertical search engines, between sensors and Web spaces, and between different media data. With our established expertise and extensive experience, we are confident that we can successfully complete this project and achieve the objectives.

Cryptanalysis and Implementation of Cipher
Hong Kong Principal Investigator : Dr Hui, Lucas Chi-kwong (The University of Hong Kong)
Mainland Principal Investigator : Prof Wang, Xiao-yun (The Shandong University )

The main objective of this project is to cryptanalyze Hash Functions and block ciphers, which means to evaluate the security of various Hash Functions and block ciphers. Hash Functions and block ciphers play an important role in providing security services in e-commerce and e-government, including confidentiality, legality, integrity, non-repudiation and authenticity. In particular, Hash Functions ensure integrity, legality, and are widely used in digital signature; while block ciphers are the main technology to provide data confidentiality.

Research on cryptanalysis of Hash Functions and block ciphers is to discover potential weakness in the cipher design. This is to ensure that the widely used Hash Functions and block ciphers are secure. This is a very important contribution to modern e-commerce. The project team had extensive experience in this topic, and will continue to perform research on the cryptanalysis of Hash Functions and block ciphers in this project, and research the fast implementation algorithm for Hash Functions under different environment including the application of computer forensics and limited resource environment such as in a mobile device.

Research results of this project will show advancement in theory, as well as great impact to the deployment of Hash Functions and block ciphers globally.

PPARδ activation prevents endothelial dysfunction in diabetes: cellular mechanisms and clinical implications
Hong Kong Principal Investigator : Prof Huang Yu (The Chinese University of Hong Kong)
Mainland Principal Investigator : Prof Wang Nanping (The Peking University )

Type 2 diabetes mellitus and obesity represent a global health problem worldwide. Most diabetics die of cardiovascular and renal causes, thus increasing the urgency in developing strategies for improving cardiovascular outcomes, particularly in obesity-related diabetes.

Recent evidence highlights the therapeutic potential of peroxisome proliferators activated receptor (PPAR) agonists in improving insulin sensitivity during diabetes. While agonists of PPARα and PPARγ are clinically used, PPARδ is the remaining subtype that is not a target for current drugs. Most studies have focused on PPARδ-mediated metabolic effects but little is known about the implication of PPARδ in the regulation of cardiovascular function. Limited studies suggest a significance of PPARδ in vascular biology. Increased cyclooxygenase (COX) expression and activity participates in vascular inflammation and PPARδ activation inhibits COX-2 expression in endothelial cells.

The proposed research will focus on the pathophysiological role of COX-2 in endothelial dysfunction in diabetic db/db mice and will investigate endothelial cell protective effect of PPARδ agonists. We aim to elucidate molecular mechanisms underlying the cross-talk between COX-2 and PPARδ that modulate endothelial function under hyperglycemic conditions. The deeper understanding of PPARδ function in blood vessels is not only of basic interest but also has important clinical implications. The findings from this investigation shall provide novel mechanistic evidence in support for the PPARδ agonists as potentially effective drugs combating against diabetic vasculopathy.

BDNF-induced ionic plasticity of GABA signaling in the cerebellum: a mechanistic and functional study
Hong Kong Principal Investigator : Dr. Yung, Wing Ho (The Chinese University of Hong Kong)
Mainland Principal Investigator : Dr. Wang, Jian-Jun (The Nanjing University)

Neurons in the central nervous system function by integrating excitatory and inhibitory inputs, which are mediated mainly by the neurotransmitters glutamate and γ-aminobutyric acid (GABA) respectively. The integrated action on the excitability of neurons forms the basis of all brain functions like sensory perception, motor control, learning, memory and others. These synaptic inputs, however, are not static but change with time. By means of mechanisms which alter the strengths of these synaptic inputs in both short-term and long-term, the central nervous system can carry out different functions efficiently and also adapt to the changes of the body and environment. Currently, we still know relatively little about the diversity and mechanisms of synaptic plasticity, especially that of the inhibitory neurotransmission, and much remains to be studied.

One form of synaptic plasticity of GABA transmission, known as ionic plasticity, is expressed as a rapid change in the synaptic strength within minutes. It is an interesting and potentially important phenomenon as it is a unique property of GABAergic, but not glutamatergic, synapse and is often expressed in trauma and disease states of the central nervous system. However, the mechanisms of GABAergic ionic plasticity, especially in the mature nervous system, are still largely unknown. In this proposal, based on the results of our preliminary studies on the cerebellum, experiments are designed to investigate this form of synaptic plasticity in depth. Three key questions will be addressed. First, what is the nature of ionic plasticity induced by the neuromodulator brain-derived neurotrophic factor in the cerebellum? Second, what are the underlying mechanisms, especially the involvement of a neuron-specific chloride ion transporter KCC2? Finally, what is the functional significance and behavioral correlates of the ionic plasticity in the cerebellum?

It is known that plastic changes of the GABAergic system underlie many neuropathological conditions such as addiction, epilepsy, chronic pain and neurodegenerative diseases. Therefore, answering the questions of this proposal not only advances our basic understanding of the plasticity and functions of central inhibitory synapses, but will also have implications in the management of disorders which involve plastic changes in these synapses.

Generation of corneal epithelial stem cells from human induced pluripotent stem cells (iPSC) and application for corneal epithelial tissue engineering
Hong Kong Principal Investigator : Dr. Qizhou Lian (The University of Hong Kong)
Mainland Principal Investigator : Prof Zuguo Liu (The Xiamen University )

Corneal-limbal epithelium damage is one common disorder to result in vision loss or blindness. Corneal epithelial stem cell is the cell source for corneal-limbal epithelium regeneration. Transplantation of corneal epithelial stem cell is the effective strategy to restore vision in those patients with limbal stem cell deficiency (LSCD). However, the shortage of available donor tissues and immune rejection are the main obstacles to widely use corneal epithelial stem cell therapy. In the case of unilateral defects, although limbal grafts can be taken from the uninjured eye, this procedure required a large limbal tissue from the healthy eye. Hence a potential serious complication arising from this procedure is limbal deficiency in the donor eye. For bilateral LSCD, transplantation of allograft limbal stem cells resulted in very low success rate due to immune rejection. To overcome the shortage of corneal epithelial stem cell source and the risk of immune rejection, here we propose to use a novel embryonic stem cell-like, termed induced pluripotent stem cells or iPSCs, to generate patient-specific corneal epithelial stem cells. Compare to conventional source of corneal epithelial stem cells taken from donor tissues, there are great advantages in using iPSCs. First, iPSCs are customized and expandable infinitely ex vivo, therefore iPSCs can be used as an unlimited cell source to generate corneal epithelial stem cells. Second, iPSCs-derived corneal epithelial stem cells are patient specific therefore no need of immunosuppression in transplantations. Third, unlike human embryonic stem cells derived form early stage of embryos, iPSCs are derived from somatic cells (i.e., fibroblasts) therefore no concerns of ethical issues. In this proposal, firstly iPSCs will be generated from patient's ocular surface epithelial cells or fibroblasts by introducing four reprogramming genes (Oct4, Sox2, Nannog and Lin28). Secondly, corneal limbal epithelial stem cells will be generated from iPSCs by overexpression of p63 and/or Pax6 gene, co-culture with limbal stromal microenvironment, or limbal stromal cell culture conditioned medium. Finally, corneal limbal epithelial stem cells will be seeded on amniotic membrane to construct tissue-engineered corneal epithelium. Additionally, tissue-engineered corneal epithelium will be transplanted into mice to determine the proliferation and terminal differentiation of the epithelial cells. This project is the first time to generate unlimited autograft limbal stem cells by means of iPSC technology, and laying the ground work for the possibility of using iPSCs-derived corneal epithelium in clinical treatment in future.

Involvement of orphan nuclear receptor Nur77 in regulating β-catenin turnover and signaling
Hong Kong Principal Investigator : Dr. Alice Sze-Tsai Wong ( The University of Hong Kong)
Mainland Principal Investigator : Dr. Jin-Zhang Zeng (The Xiamen University )

Because β-catenin is both essential and central to the regulation of cell cycle and proliferation, the aberrant expression of β-catenin plays an important role in the development and progression of numerous cancer types in particular colon cancer. Since the canonical APC/Axin/GSK3β and p53/Siah-1/APC pathways for downregulation of β-catenin are often mutated in human cancers, identifying new mechanisms that regulate Wnt/β-catenin signaling will have great clinical significance. Recently, many members of nuclear receptor family have been demonstrated to be a potent modulator of Wnt/β-catenin signaling, which has attracted increasing attention. This cross-talk may be key to developing new therapeutics for cancer prevention and therapy.

Nur77 is an orphan nuclear receptor and an important molecular target in cancer for its critical role in regulation of proliferation and apoptosis. In our preliminary study, we show for the first time that in colon cancer, the expression of β-catenin can be regulated by Nur77. We observed that Nur77 was significantly lower and frequently cleaved into small fragments in most colon cancerous tissues when comparing to their adjacent normal colon tissues. Using transfection experiments, we found that overexpression of Nur77 could stimulate β-catenin degradation and inhibit β-catenin transcriptional activities. Intriguingly, certain components of bile acids could induce the expression of Nur77 and regulate the interaction of Nur77 with β-catenin. Based on these interesting findings, we will further investigate in this project the molecular mechanisms underlying the cross-talk between Nur77 and β-catenin and screen a chemical library for small molecules that may target at the Nur77/β-catenin pathway for the potential development of new chemotherapeutic agents. Because aberrant expression of β-catenin and Nur77 is associated with the pathogenic process and regulated by high fat diet and bile acids, further studies on the cross-talk between Nur77/β-catenin will be highly relevant to our understanding of the pathogenesis of colon cancer and should have significant impact on our development of new agents involving Nur77/β-catenin signaling for its prevention and therapy.

Identifying novel regulators in cADPR-mediated calcium signaling by combining approaches of synthetic organic chemistry and RNAi screen
Hong Kong Principal Investigator : Dr. Yue, Jianbo ( The University of Hong Kong)
Mainland Principal Investigator : Prof. Zhang, Liang-Ren (The Peking University )

Mobilization of intracellular Ca2+ stores is involved in many diverse cell functions, including cell proliferation, differentiation, fertilization, muscle contraction, secretion of neurotransmitters, hormones and enzymes, and lymphocyte activation and proliferation [1-5]. Cyclic adenosine diphosphoribose (cADPR) is an endogenous Ca2+ mobilizing nucleotide present in many cell types and different species, from plants to animals. cADPR is formed by ADP-ribosyl cyclases from nicotinamide adenine dinucleotide (NAD). It has been shown that many extracellular stimuli can induce cADPR production that leads to calcium release or influx, establishing cADPR as a second messenger [1, 6-8]. Although evidence indicates that cADPR elicits Ca2+ release via the ryanodine receptor, the molecular mechanisms regarding the cADPR-induced Ca2+ release remain elusive. We recently reported a novel strategy to chemically synthesize cell-permeant cADPR agonists [106]. Moreover, we found that one of the cADPR cell permeant agonists, cTDPRE, induced Ca2+ release in human Jurkat T cells, and knocking-down ryanodine receptor-3 (RyR-3) in Jurkat cells significantly inhibited the ability of cTDPRE to induce Ca2+ release. We hypothesize that novel signaling proteins are required for or can modify the ability of cADPR to induce Ca2+ release via ryanodine receptors. Therefore we plan to identify these novel regulators in cADPR-mediated Ca2+ signaling by combining approaches of synthetic organic chemistry and RNAi screen. We propose to: (1) chemically synthesize and pharmacologically characterize novel and potent cell-permeant cADPR agonists and (2) identify novel regulators in cADPR-mediated calcium signaling using RNAi screen. Given the pivotal role of cADPR-mediated calcium signaling pathway in a wide variety of cellular processes, understanding the molecular mechanisms involved in this novel signaling pathway is important not only for scientific reasons but also have clinical relevance.

Tracking Metals in Cells by Metallomics: Insights into the Metal-associated Pathophysiological Processes
Hong Kong Principal Investigator : Dr. Hongzhe SUN ( The University of Hong Kong)
Mainland Principal Investigator : Prof. Zhifang CHAI (The Chinese Academy of Sciences)

The homeostasis of metal ions is of critical importance in maintaining numerous biological and biochemical processes essential for life. Indeed, around one-third of all proteins require metal ions for activity. However, metal ions present a dilemma to cells; they are useful but can be toxic. Imbalances in the concentrations of metals may affect essential cellular functions (e.g. signal transduction pathways), promote oxidative stress and disrupt the proper folding of nascent proteins. Consequently, comprehensive and systematic investigations into metal (and metalloid) composition, speciation (i.e. metal-biomolecule interactions), localization and their uses within cells are urgently warranted. The entirety of metal content, species within a living organism (cell or tissue), their identity, quantity and localization is defined as the metallome, analogous to genome and proteome. Metallomics is the study of a metallome; the interactions and functional connections of metals with genes, proteins and metabolites.

Helicobacter pylori is a major human pathogen which can cause peptic ulcers, chronic gastritis and even cancer. The bacterium can survive under highly acidic conditions, relying on two important metalloenzymes: urease and [Ni,Fe]-hydrogenase. Bismuth compounds have long been used for the treatment of H. pylori infections and peptic ulcers. Despite the numerous essential roles played by metals within the bacterium, there is a paucity of literature reports detailing its metal contents.

The aim of the project is to conduct a detailed metallomic study to comprehensively analyze metals within cells using H. pylori as an example. Specifically, the PI and Co-I intend to:
(1) Quantify key metals (Fe, Zn and Ni) bound to proteome of H. pylori (and related pathogens) using one- and two-dimensional gel electrophoresis (2DE) coupled with advanced nuclear analytical techniques such as synchrotron radiation microbeam scanning X-ray fluorescence (SR-£gXRF);
(2) Identify and characterize important metal-binding 'environments' within proteins isolated by immobilized metal affinity chromatography (IMAC) approaches by X-ray absorption spectroscopy (XAS);
(3) Determine the structure and function of key metalloproteins involved in metal-associated pathophysiological processes by NMR, X-ray crystallography and XAS;
(4) Develop a fluorescence-based metal chelating conjugate for tagging metal-binding proteins within cells or 2-DE gels enabling them to be localized and quantified.

This study will greatly improve our knowledge of the localization and speciation of metals within this key bacterial pathogen, and will shed light on the unique roles of metals and metalloproteins in important pathophysiological processes. The long-term goal of this collaborative research program is to establish a state-of-the-art center for metallomics, particularly in the field of metal-related diseases.

Potential Impacts of Global Climate Change and Environmental Deterioration on Coral Reefs in South China Sea
Hong Kong Principal Investigator : Prof. Ang Put Jr. ( The Chinese University of Hong Kong)
Mainland Principal Investigator : Dr. Huang Hui (The Institute of Oceanology, Academia Sinica )

Corals reefs provide marine resources that are worth billions of US dollars per annum and support livelihood of millions of people. They are also important habitats that support some of the highest biodiversity in the world. However, corals and coral reefs around the world are under increasing threat from the effects of global climate change and environmental deterioration. The aim of this study is to examine the responses of select coral species from different sites in southern China and the South China Sea to changing environmental conditions, and to model the potential consequences of further environmental deterioration and global climate change on coral biodiversity and coral reef community structure. A pollution gradient can be drawn, with the highest pollution levels recorded in the Pearl River Delta, including Hong Kong, and decreasing towards the west to Sanya in Hainan Province and southwest in the Xisha Island group in the South China Sea. Experiments will be conducted to evaluate the responses of select coral species in these sites to combined changes in seawater temperature, levels of eutrophication and organic pollution. Based on information obtained from these experiments as well as on an in situ survey of the coral communities of the study sites, object-oriented computer modeling will be used to project the potential impact of global climate change and continued deterioration of environmental conditions on these corals and coral reefs in southern China and the South China Sea.

Atmospheric Halogenated Hydrocarbons in the Pearl River Delta Region
Hong Kong Principal Investigator : Dr. GUO Hai ( The Hong Kong Polytechnic University)
Mainland Principal Investigator : Dr. WANG Xinming (The Chinese Academy of Sciences)

The Pearl River Delta (PRD) region, as one of the most urbanized and industrialized city clusters in China, plays a vital role in the national use and emissions of halocarbons. The proposed project, by grid-sampling simultaneously at 45 sites over the whole PRD and long-term monitoring at 3 representative sites in the region, will study spatiotemporal patterns of atmospheric halocarbons and explore their variation trends in comparison with historical data obtained in the past. Based on the field measurement results, the project also aims to highlight hot spots of halocarbon emissions in the region with extensive meteorological analysis; to make source apportionment of major halocarbons using receptor models; and to estimate the emission amounts of halocarbons through their correlations with carbon monoxide or by inversion models.

Antifouling compounds (butenolides and indole alkaloids) derived from marine microbes: study of structure-function relationship and mode-of-action (at both gene and protein level)
Hong Kong Principal Investigator : Prof. Pei-Yuan QIAN (The Hong Kong University of Science and Technology)
Mainland Principal Investigator : Prof. Shu-Hua QI (The Chinese Academy of Science )

Marine biofouling refers to the undesirable buildup of marine organisms on man-made surfaces submerged in seawater and causes leads to more than US$6.0 billion of annual losses world-wide in the maritime industries. The recent complete ban of tributyltin (TBT)-based marine coatings request the introduction for non-toxic or less-toxic antifouling compounds for the marine coating industry. Although there have been substantial efforts in screening for non-toxic antifouling compounds (including work from the PI's lab), little information on the following topics has been revealed:
a) structure-function relationships of the identified non-toxic antifouling compounds, which makes it very difficult for us to design the compound structure to achieve high bioactivity and less toxicity,
b) mode-of-action of compounds on target organisms (larvae of major fouling organisms), which makes it very challenging to develop bioactive compounds into commercial products, and
c) genomics and proteomics information on target organisms, which makes it very difficult for us to study the mode-of-action of bioactive compounds.

In this project, based on our previous findings, we will select two groups of bioactive antifouling compounds and biomarkers to carry out in-depth studies of the antifouling mechanisms of those compounds. Specifically, we propose to:
1) isolate and identify antifouling compounds of butenolides and indole alkaloids from marine bacteria Streptomyces spp, Bacillus spp, using a bioassay-guided procedure;
2) study the structure-function relationship of two groups of bioactive compounds to identify the compounds with the highest antifouling activity against larval settlement of both Hydroides elegans and Balanus amphitrite;
3) study the effects of bioactive compounds on expression profiles of the p38 mitogen-actived protein kinase (MAPK) gene in settling larvae of H. elegans;
4) study the effects of bioactive compounds on the expression profile of selected proteins in B. amphitrite.

We expect to identify chemical structure(s) with high antifouling activities through structure-function analysis. More importantly, we hope to gain better insight onto the possible molecular mechanisms of how the selected bioactive compounds act on settling larvae of two major fouling organisms at both gene and protein levels. Finding(s) of this study may help us in developing some of bioactive antifouling compounds into commercial products for marine coating industry to address the urgent needs of the maritime industries.

A High-Throughput Platform for Investigation of Combination Effects of Environmental Pollutants
Hong Kong Principal Investigator : Prof. Hongkai Wu (The Hong Kong University of Science and Technology)
Mainland Principal Investigator : Prof. Guibin Jiang (The Chinese Academy of Science )

Being able to predict the toxicity of environment pollutants from their components is critical in environment science, because the pollutants generally appear in air, water, and soil as complex mixtures. It is also critical in the decision making process of government agencies to take appropriate preventive or damage-control measures against environmental contamination. However, it is not an easy task to define the relationship between the toxicities of a complex environmental mixture and its components. The difficulty arises from the fact that a single component in these complex mixtures may negate, augment, or induce the cytotoxicity of other components, causing antagonistic, synergistic, or potentiative combination effects. In addition, the multiplicity of polluting molecules in a typical environmental sample demands a vast number of bioassays, which is beyond the capability of traditional toxicity assessment methods, to define the combination effects. Here, we propose to develop a high throughput platform on the basis of modern microfluidic technologies to enable the investigation on the combination effects of the numerous toxic compounds in environmental pollutants. The high throughput will rely on a novel multi-channel pico-pipette that is integrated into the microarray system and capable of accurately delivering liquid of picoliter (10^-12 L) to nanoliters (10^-9 L). This platform will be evaluated and validated with standard compounds with known individual and combination toxicities. Moreover, key polluting components will be identified and isolated from real environmental samples and their combination effects in disrupting the estrogen balance in human will be determined with the microarray platform. This high throughput platform is targeted for development of new and accurate models to assess the health hazard of the complex environmental pollutants and to help government agencies make environmental policies on sound scientific basis.

An investigation of product recalls and supply chain quality management model
Hong Kong Principal Investigator : Prof. Xiande Zhao (The Chinese University of Hong Kong)
Mainland Principal Investigator : Prof. Xiucheng Fan (The Fudan University)

Product harm crises and subsequent product recalls are among a firm's worst nightmares, because product recalls can be very costly. These events not only result in huge losses to a country's economy, but have also caused injuries that require high levels of medical treatment and permanent disabilities and death.

Considering its impact on operations, a product harm crisis not only generates huge losses to manufacturers and overseas brand owners, but can also trigger a detrimental ripple effect throughout the entire supply chain. In other words, each party in the supply chain will suffer. It can distort long-standing brand equity, tarnish a company's reputation and cause major revenue and market-share losses. Product recalls, like other types of negative publicity, not only severely damage a corporate image and brands, but also entire companies.

It is in such a context that this study is motivated. We investigate how organizations should make use of different product recall strategies to mitigate the impact of a product harm crisis. We will provide a set of management guidelines for formulating product recall decisions and identify the components that comprise a good and workable product recall system by which organizations can efficiently and promptly identify, trace and purge risky products from a supply chain. However, to address product harm crises through proper product recall strategies is a short term remedy and is insufficient for firms to sustain their competitiveness in the market. Organizations should turn their eyes to setting up a Supply Chain Quality Management (SCQM) system to assure their success over time. Therefore, we will develop a holistic SCQM model that organizations can adopt, not only to prevent the recurrence of product harm crises, but also to improve and sustain their supply chain quality performance, both upstream and downstream.

This study will make significant contributions to academic literature by delineating the relationship between consumers' perceived degree of danger associated with product harm, their perceived attribution of responsibility of the company, the corporate social legitimacy, perceived risk/trust, customer satisfaction, repurchase intention and negative word of mouth. It will also reveal how product recall strategies affect these relationships. This study will be the first in the field to conceptualize and operationalize a SCQM model through a rigorous scale development process. We will synthesize the previous literature and propose factors that have been neglected in previous studies. Finally, we will scrutinize the model under different contexts and environmental conditions.

Multi-objective design, optimization and implementation of network congestion pricing schemes
Hong Kong Principal Investigator : Prof. Hai Yang (The Hong Kong University of Science and Technology)
Mainland Principal Investigator : Prof. Haijun Huang (The Beijing University of Aeronautics and Astronautics )

Road pricing as an effective means for both managing road traffic demand and raising additional revenue for road construction has been studied extensively by both transportation researchers and economists. Practical implementation has been progressing rapidly and electronic road pricing schemes have been proposed and tested worldwide (for example, the Singapore and Hong Kong electronic road pricing schemes and the recent London area-based congestion charging scheme). It is likely that over the next few years, with increasing public acceptability, there will be greater use of road pricing.

Despite the well-established economic marginal-cost pricing principle and rapidly emerging innovative road pricing technologies, the incontestable fact is that there is a great need for the development of efficient road-use pricing models in general networks. In particular, for a traffic network with multi-class, heterogeneous users in terms of their different values of time, the system performance can be measured either in time units by the total system travel time or in monetary units by the total system travel cost. Thus, we have two different objectives in network optimization, i.e., to minimize system time and to minimize system cost, which naturally gives rise to a bi-objective minimization problem. A Pareto optimum of this bi-objective optimization problem represents a bi-criteria system optimum for network optimization in the sense that, at each Pareto optimum, neither system time nor system cost can be further reduced without increasing the other one.

This project aims to tackle both theoretical and practical issues of the above multi-objective and multi-class network congestion pricing schemes. First, we consider the unconstrained Pareto system optimization and investigate whether or under what conditions a Pareto system optimum can be decentralized into a multi-class user equilibrium by positive link tolls (the first-best problem). Second, we consider practical road pricing and Pareto system optimization in a network subject to certain charging constraints (the second-best problem). The model and algorithm developed will be tested with test networks and case studies of Hong Kong and Beijing. The outcome of the study should be useful for our theoretical understanding and practical design and optimization of road pricing schemes from a multi-objective perspective.

Plasmon-enhanced, broad-spectrum-light-active photocatalysts based on hybrid
nanostructured materials made of noble metals, semiconductors, and rare earths
Hong Kong Principal Investigator : Prof. WANG Jianfang (The Chinese University of Hong Kong)
Mainland Principal Investigator : Prof. YAN Chunhua (The Peking University)

Semiconductor photocatalysts have proven highly effective in the remediation of contaminated water and air and solar energy conversion. Among various photocatalysts, TiO2 has attracted much attention owing to its photocatalytic activity, chemical stability, and photostability. However, anatase TiO2 has a bandgap energy of 3.2 eV. It absorbs only UV light with λ< 388 nm. The integrated solar photon flux between 280 and 400 nm amounts to only ~1.4% of the total solar photon flux. Efforts have therefore been made to find photocatalysts that are sensitive to the visible light, because the solar photon flux between 400 and 700 nm amounts to ~26.2%. While doping and dye sensitization have been employed to broaden the spectral response of TiO2, the photocatalytic activities of the resultant materials have generally remained low, due to their limited active spectral ranges, small light absorption coefficients, and instability under light irradiation.

We therefore propose to design highly efficient photocatalysts that are active from the UV to NIR region. Specifically, we are going to prepare hybrid nanostructures and thin films that are composed of semiconductors, noble metals, and rare earths. Rare earth elements that are doped or incorporated as nanocrystals will function as sensitizers to absorb the NIR light, upconvert low-energy photons into high-energy photons, and then transfer energy to UV- or visible-light-active semiconductor components for photocatalysis. The efficiency of photocatalysts will be significantly improved if they are made sensitive to the NIR light, because the integrated solar photon flux from 700 to 1500 nm amounts to ~47.4%. Moreover, noble metal nanocrystals exhibit rich plasmonic properties, with their plasmon resonance wavelengths tunable from ~400 to ~2000 nm. At their plasmon resonances, they can greatly enhance the light absorption of neighboring materials. Noble metal nanocrystals will therefore be incorporated into the hybrid materials to enhance their light absorption from the visible to NIR region. The presence of metal nanocrystals can also improve the charge separation in semiconductors. Both effects will increase the photocatalytic efficiencies of the hybrid materials. We will vary the geometrical configurations of the hybrid nanostructures and investigate the interactions among light absorption, upconversion, energy transfer, plasmon resonance, charge separation, and photocatalysis in order to obtain photocatalysts that are active under a broad spectrum of light and have very high efficiencies.

One-Dimensional Nanomaterials and Their Arrays; Synthesis, Characterization, and Applications in High Performance Lithium Ion Batteries
Hong Kong Principal Investigator :Prof. Shihe Yang (The Hong Kong University of Science and Technology)
Mainland Principal Investigator : Prof. Shi-Gang Sun (The Xiamen University )

Lithium-ion batteries (LIBs) are currently a dominant power source for portable appliances such as mobile phones and notebook computers. To meet the demand of the rapidly developing society for technologies in, for example, electric and hybrid electric vehicles (EVs and HEVs, respectively), developing next-generation LIBs with high energy density, high power, and high safety, has become a key issue. At present, traditional electrode materials such as graphite (anode), LiCoO2, LiMnO2 or LiFePO4 (cathode), cannot live up to the performance standard required for the next-generation LIBs, due to their relatively low specific capacity among other issues. Hence, there is now an urgent need for developing new electrode materials for the next-generation LIBs.

In the past decade, the Hong Kong team developed simple strategies for growing vertically aligned inorganic nanowire arrays directly from and on metal surfaces under mild conditions. Such one-dimensional (1D) nanostructured arrays could be the solution to many problems of the current LIBs because they can be directly used as electrodes for LIBs with conceivable virtues of low internal resistance for charge and mass transport, large specific surface area to sustain large current density, moderate inter-wire spaces for accommodating Li-insertion induced volume expansion, and no need for additives such as conducting agents and binders. Therefore, the ability to synthesize an extensive series of inorganic 1D nanostructured arrays on metal substrates will open exciting opportunities to develop high specific capacity, high rate, and long cycle lifetimes of LIBs. In parallel, the Mainland team broke new ground in understanding and controlling basic and applied electrochemical processes on electrode surfaces, which are important for developing LIBs. For example, by clever control of electrode surface kinetics, they were able to abate structural collapse induced capacity deterioration.

We seek to enhance the collaboration between our teams of HKUST and XMU on the development of 1D nanomaterials and their arrays for high performance LIBs by tackling technological challenges and the underlying fundamental issues. The two teams have a perfect combination of expertise on materials chemistry and electrochemistry to achieve the goals of this joint project with the common thread to exploit 1D nanomaterials for LIBs. Our goals are (1) to study in-situ synthesis and surface modifications of inorganic 1D nanomaterials and their arrays, (2) to delineate the structural, electronic and electrochemical characteristics of the 1D nanomaterials and their arrays, and (3) to develop applications of the 1D nanomaterials and their arrays in high performance LIBs.

Cement-based piezoelectric composites and transducers for concrete structure health monitoring
Hong Kong Principal Investigator : Prof. Zongjin Li (The Hong Kong University of Science and Technology)
Mainland Principal Investigator : Prof. Xin Cheng (The University of Jinan )

This project aims at developing cement-based piezoelectric composites and the corresponding transducers for health monitoring of intelligent civil engineering structures. To our knowledge, HKUST is the first institution in the world to start the research on cement-based smart materials and it is accumulated good experience in cement-based piezoelectric materials and sensor development. This project will focus on the enhancement of new piezoelectric composites and sensing systems. The sol-gel method will be employed to produce a new type of piezoelectric crystal on a finer sub-micrometer scale. The new crystal will be used with several different cements, including magnesium oxychloride cement, to manufacture cement-based piezoelectric composites as well as sensing elements. To induce chemical reactions between the matrix materials and piezoelectric crystals, some primers will be selected based on molecular dynamics simulation. Innovative cement-based sensors and transducers will be developed with these piezoelectric composites. With the incorporation of these sensors and transducers, the health condition of buildings and infrastructures can be monitored both passively and actively. Hence, protection measures can be proposed and taken based on the information provided by health monitoring sensors to ensure a safer and more comfortable living and working environment. This health monitoring system will not only improve the service ability of tall buildings and long-span bridges that are frequently encountered in Hong Kong and the Mainland, but it will also help position Hong Kong and Mainland China at the forefront of the health monitoring of civil engineering structures in the world.

Transport Property and Thermoelectric effect of iron-based superconductors
Hong Kong Principal Investigator : Prof. Fu-Chun Zhang (The University of Hong Kong)
Mainland Principal Investigator : Prof. Zhu-an Xu (The Zhejiang University )

Superconductivity is a remarkable quantum phenomenon, where the resistance disappears suddenly as temperature is lowered below to a critical temperature Tc. The transition temperature for superconductivity is however very low, which has limit its application. Recently discovered iron based superconductors are a class of new materials with highest transition temperatures only next to the copper oxides. There have been world-wide intense activities on this new class of materials and its physical properties. One of the keys in study of this material is the effect of chemical doping, which induces or affect the superconducting properties drastically. In this project, we propose to study three typical doped systems in the iron-based superconductors: 1) the electron doped material via substitution of in-plane Fe atoms with Co or Ni atoms, 2) the hole or electron-type systems with the heterovalent substitution outside the FeAs layers, and 3) the systems with chemical pressure induced by isovalent substitutions. We will systematically investigate the transport, magnetic and thermoelectric properties of the materials. Combined with the study of optical spectroscopy, we will explore the evolution of the ground state, especially the possible quantum phase transition and quantum critical phenomena in the systems, and try to obtain complete phase diagrams of these doped systems. With the combination of the theoretic and experimental studies, we will illustrate the evolution of various ordered phases which may further the understanding of the superconducting mechanism of the iron pnictide superconductors.

Fault-tolerant control of batch processes with hybrid nature
Hong Kong Principal Investigator : Prof. Furong Gao (The Hong Kong University of Science and Technology)
Mainland Principal Investigator : Prof. Donghua Zhou (The Tsinghua University)

Batch process is the preferred method for manufacturing high-value-added products. Control performance is critically important to the quality and consistency of batch processes. In comparison with a continuous process, a batch process has its own characteristics. Control system design and analysis should be, therefore, conducted in harmony with the features of this process for good performance. Both the Hong Kong applicant and the mainland applicants have advanced knowledge of batch processes. Batch process control design in the past has been predominately used to take advantage of its repetitive nature. In addition to the repetitive nature, a batch process has other important features, including (a) two dimensional time dynamics. It is easy to understand that a batch process has dynamics evolving with time within a batch; at the same time, it can also have significant batch-to-batch dynamics due to environmental factors, material variations, or employment of batch-to-batch control. (b) Hybrid nature. A batch process can typically have several operation or characteristic phases. A different control objective and strategy is typically employed for each phase in the so-called multi-phase control framework. The switching of control strategies from phase to phase can create significant control performance perturbation after the transition. This project proposes to treat the multi-phase control problem in a uniform framework as a hybrid paradigm for better performance and design. (c) Batch process may also be subject to faults as it needs to operate under challenging conditions; fault-tolerant control of batch processes has been neglected issue. This joint NSFC-RGC project will combine the expertise of the Hong Kong applicant with that of the mainland team to systematically study batch process control by fully exploring these features. The successful completion of this study will address these important academic issues and also provide a control method that is harmonious with batch process features. Control performance is expected to be increased significantly. Batch process is particularly important to Hong Kong as the industries remaining in Hong Kong are predominately batch processes industries.