Web applications are one of the fastest growing types of software systems today. Structurally, they are composed out of two parts: the server-side, used for data-access and business logic, and the client-side used as a user-interface. In recent years, thanks to fast, modern web browsers and advanced scripting techniques, developers are building complex interfaces, and the client-side is playing an increasingly important role. From the user's perspective, the client-side offers a number of features. A feature is an abstract notion representing a distinguishable part of the system behavior. Similar features are often used in a large number of web applications, and facilitating their reuse would offer considerable benefits. However, the client-side technology stack does not offer any widely used structured reuse method, and code responsible for a feature is usually copy-pasted to the new application. Copy-paste reuse can be complex and error prone - usually it is hard to identify exactly the code responsible for a certain feature and introduce it into the new application without errors. The primary focus of the research described in this PhD thesis is to provide methods and tools for automatizing reuse in client-side web application development. This overarching problem leads to a number of sub-problems: i) how to identify code responsible for a particular feature; ii) how to include the code that implements a feature into an already existing application without breaking neither the code of the feature nor of the application; and iii) how to automatically generate sequences of user actions that accurately capture the behavior of a feature? In order to tackle these problems we have made the following contributions: i) a client-side dependency graph that is capable of capturing dependencies that exist in client-side web applications, ii) a method capable of identifying the exact code and resources that implement a particular feature, iii) a method that can introduce code from one application into another without introducing errors, and iv) a method for generating usage scenarios that cause the manifestation of a feature. Each contribution was evaluated a suite of web applications, and the evaluations have shown that each method is capable of performing its intended purpose.

As a more efficient means of supplying computing resources in the form of a utility, cloud computing platforms have been increasingly used to meet the insatiable demand from mobile applications. The research problems we study in this thesis are in the general research area of mobile cloud computing, as we seek to design and implement new algorithms and protocols that straddle the boundary between mobile applications and cloud computing systems, so that their performance can be jointly optimized to provide the best possible user experience, yet operating within the constraints of available resources and operational costs. From the perspective of mobile applications, we show that interactive applications have the need to stream multi-touch gestures among multiple users. Tailored to the nature of multi-touch gesture streams, we propose a new protocol that uses inter-session network coding to reduce the gesture recognizing delays. Towards supporting mobile applications using the cloud resources, we believe that multi-party video conferencing service can benefit from the inter-datacenter networks in the cloud. We apply intra-session network coding to design a new protocol to maximize the total throughput of all conferencing sessions in the cloud, subject to a latency constraint imposed by the nature of video conferencing. Our real-world experiments have shown that the inter-datacenter networks help to achieve substantially improved throughput, with very similar delays compared to traditional peer-to-peer solutions. From the perspective of cloud service providers, we study the challenges involved when resource utilization is to be maximized and when operational costs are to be minimized. To maximize resource utilization, we propose a virtual machine (VM) migration algorithm based on Nash bargaining solutions. To minimize operational costs, we present optimal routing and flow assignment algorithms for the inter-datacenter traffic, with and without store-and-forward capabilities in intermediate datacenters. With efficient and cost-effective utilization of resources in the cloud, and by designing new protocols that are applicable to both mobile applications and cloud computing systems, achieving an optimized level of user experience with interactive mobile applications will become a reality.

As a more efficient means of supplying computing resources in the form of a utility, cloud computing platforms have been increasingly used to meet the insatiable demand from mobile applications. The research problems we study in this thesis are in the general research area of mobile cloud computing, as we seek to design and implement new algorithms and protocols that straddle the boundary between mobile applications and cloud computing systems, so that their performance can be jointly optimized to provide the best possible user experience, yet operating within the constraints of available resources and operational costs. From the perspective of mobile applications, we show that interactive applications have the need to stream multi-touch gestures among multiple users. Tailored to the nature of multi-touch gesture streams, we propose a new protocol that uses inter-session network coding to reduce the gesture recognizing delays. Towards supporting mobile applications using the cloud resources, we believe that multi-party video conferencing service can benefit from the inter-datacenter networks in the cloud. We apply intra-session network coding to design a new protocol to maximize the total throughput of all conferencing sessions in the cloud, subject to a latency constraint imposed by the nature of video conferencing. Our real-world experiments have shown that the inter-datacenter networks help to achieve substantially improved throughput, with very similar delays compared to traditional peer-to-peer solutions. From the perspective of cloud service providers, we study the challenges involved when resource utilization is to be maximized and when operational costs are to be minimized. To maximize resource utilization, we propose a virtual machine (VM) migration algorithm based on Nash bargaining solutions. To minimize operational costs, we present optimal routing and flow assignment algorithms for the inter-datacenter traffic, with and without store-and-forward capabilities in intermediate datacenters. With efficient and cost-effective utilization of resources in the cloud, and by designing new protocols that are applicable to both mobile applications and cloud computing systems, achieving an optimized level of user experience with interactive mobile applications will become a reality.

This thesis examines the issues related to calibrating digital cameras and lenses, which is an essential prerequisite for the extraction of precise and reliable 3D metric information from 2D images. The techniques used to calibrate a Canon PowerShot A70 camera with 5.4 mm zoom lens and a professional single lens reflex camera Canon EOS 1Ds Mark II with 35 mm, 85 mm, 135 mm and 200 mm prime lenses are described. The test results have demonstrated that a high correlation exists among some interior and exterior orientation parameters. The correlations are dependent on the parameters being adjusted and the network configuration. Not all of the 11 interior orientation parameters are significant for modelling the camera and lens behaviour. The first two coefficients K1, K2 would be sufficient to describe the radial distortion effect for most digital cameras. Furthermore, the interior orientation parameters of a digital camera and lens from different calibration tests can change. This work has demonstrated that given a functional model that represents physical effects, a reasonably large number of 3D targets that are well distributed in three-dimensional space, and a highly convergent imaging network, all of the usual parameters can be estimated to reasonable values. / Mining Engineering

Microbial nitrile hydratases are important industrial enzymes that catalyse the conversion of nitriles to the corresponding amides. A thermostable, cobalt-type Bacillus sp. RAPc8 microbial nitrile hydratase was cloned and expressed in E.coli. In this study the primary aim was to determine the molecular structure of Bacillus sp. RAPc8 microbial nitrile hydratase.

<p>The nitrilases are an important class of industrial enzymes that are found in all phyla. These enzymes are expressed widely in prokaryotes and eukaryotes. Nitrilases convert nitriles to corresponding acids and ammonia. They are used in industry as biocatalysts because of their specificity and enantioselectivity. These enzymes belong to the nitrilase superfamily in which members share a common &alpha / &beta / &beta / &alpha / structural fold and a unique cys, glu,lys catalytic triad with divergent N- and C-terminals.</p> <p>There are four atomic structures of distant homologues in the superfamily, namely 1ems, 1erz, 1f89 and 1j31. All structures have two-fold symmetry which conserves the &alpha / &beta / &beta / &alpha / -&alpha / &beta / &beta / &alpha / fold across the dimer interface known as the A surface. The construction of a 3D model based on the solved structures revealed the enzyme has two significant insertions in its sequence relative to the solved structures, which possibly correspond to the C surface. In addition there are intermolecular interactions in a region of a conserved helix, called the D surface. These surfaces contribute additional interactions responsible for spiral formation and are absent in the atomic resolution homologues.</p> <p>The recombinant enzyme from R.rhodochrous J1 was expressed in E. coli BL21 cells and eluted by gel filtration chromatography as an active 480 kDa oligomer and an inactive 80 kDa dimer in the absence of benzonitrile. This contradicts previous observations, which reported the native enzyme exists as an inactive dimer and elutes as a decamer in the presence benzonitrile. Reducing SDS-PAGE showed a subunit atomic mass of ~40 kDa. EM and image analysis revealed single particles of various shapes and sizes, including c-shaped particles, which could not form spirals due to steric hindrances in its C terminal.</p> <p>Chromatographic re-elution of an active fraction of 1-month old J1 nitrilase enabled us to identify an active form with a mass greater than 1.5 MDa. Reducing SDS-PAGE, N-terminal sequencing and mass spectroscopy showed the molecular weight was ~36.5 kDa as result of specific proteolysis in its C terminal. EM revealed the enzyme forms regular long fibres. Micrographs (109) were recorded on film using a JEOL 1200EXII operating at 120 kV at 50K magnification. Two independent 3D reconstructions were generated using the IHRSR algorithm executed in SPIDER. These converged to the same structure and the resolution using the FSC 0.5 criterion was 1.7 nm.</p> <p>The helix structure has a diameter of 13nm with ~5 dimers per turn in a pitch of 77.23 &Aring / . Homology modeling and subsequent fitting into the EM map has revealed the helix is built primarily from dimers, which interact via the C and D surfaces. The residues, which potentially interact across the D surface, have been identified and these confer stability to the helix. The conservation of the insertions and the possibility of salt bridge formation on the D surface suggest that spiral formation is common among microbial nitrilases. Furthermore, the presence of the C terminal domain in J1 nitrilase creates a steric hindrance that prevents spiral formation. When this is lost &ndash / either by specific proteolysis or autolysis - an active helix is formed.</p>

<p>The nitrilases are an important class of industrial enzymes that are found in all phyla. These enzymes are expressed widely in prokaryotes and eukaryotes. Nitrilases convert nitriles to corresponding acids and ammonia. They are used in industry as biocatalysts because of their specificity and enantioselectivity. These enzymes belong to the nitrilase superfamily in which members share a common &alpha / &beta / &beta / &alpha / structural fold and a unique cys, glu,lys catalytic triad with divergent N- and C-terminals.<br /> <br /> There are four atomic structures of distant homologues in the superfamily, namely 1ems, 1erz, 1f89 and 1j31. All structures have two-fold symmetry which conserves the &alpha / &beta / &beta / &alpha / -&alpha / &beta / &beta / &alpha / fold across the dimer interface known as the A surface. The construction of a 3D model based on the solved structures revealed the enzyme has two significant insertions in its sequence relative to the solved structures, which possibly correspond to the C surface. In addition there are intermolecular interactions in a region of a conserved helix, called the D surface. These surfaces contribute additional interactions responsible for spiral formation and are absent in the atomic resolution homologues.<br /> <br /> The recombinant enzyme from R.rhodochrous J1 was expressed in E. coli BL21 cells and eluted by gel filtration chromatography as an active 480 kDa oligomer and an inactive 80 kDa dimer in the absence of benzonitrile. This contradicts previous observations, which reported the native enzyme exists as an inactive dimer and elutes as a decamer in the presence benzonitrile. Reducing SDS-PAGE showed a subunit atomic mass of ~40 kDa. EM and image analysis revealed single particles of various shapes and sizes, including c-shaped particles, which could not form spirals due to steric hindrances in its C terminal.</p> <p>Chromatographic re-elution of an active fraction of 1-month old J1 nitrilase enabled us to identify an active form with a mass greater than 1.5 MDa. Reducing SDS-PAGE, N-terminal sequencing and mass spectroscopy showed the molecular weight was ~36.5 kDa as result of specific proteolysis in its C terminal. EM revealed the enzyme forms regular long fibres. Micrographs (109) were recorded on film using a JEOL 1200EXII operating at 120 kV at 50K magnification. Two independent 3D reconstructions were generated using the IHRSR algorithm executed in SPIDER. These converged to the same structure and the resolution using the FSC 0.5 criterion was 1.7 nm.<br /> <br /> The helix structure has a diameter of 13nm with ~5 dimers per turn in a pitch of 77.23 &Aring / . Homology modeling and subsequent fitting into the EM map has revealed the helix is built primarily from dimers, which interact via the C and D surfaces. The residues, which potentially interact across the D surface, have been identified and these confer stability to the helix. The conservation of the insertions and the possibility of salt bridge formation on the D surface suggest that spiral formation is common among microbial nitrilases. Furthermore, the presence of the C terminal domain in J1 nitrilase creates a steric hindrance that prevents spiral formation. When this is lost &ndash / either by specific proteolysis or autolysis - an active helix is formed.</p>

We describe the application of conditional random fields (CRF) to physiological data modeling for the application of activity recognition. We use the data provided by the Physiological Data Modeling Contest (PDMC), a Workshop at ICML 2004. Data used in PDMC are sequential in nature: they consist of physiological sessions, and each session consists of minute-by-minute sensor readings. We show that linear chain CRF can effectively make use of the sequential information in the data, and, with Expectation Maximization, can be trained on partially unlabeled sessions to improve performance. We also formulate a mixture CRF to make use of the identities of the human subjects to further improve performance. We propose that mixture CRF can be used for transfer learning, where models can be trained on data from different domains. During testing, if the domain of the test data is known, it can be used to instantiate the mixture node, and when it is unknown (or when it is a completely new domain), the marginal probabilities of the labels over all training domains can still be used effectively for prediction. / Singapore-MIT Alliance (SMA)

Current accreditation preparation methods at many universities are time consuming and tedious. This is because cost prohibits many programs from purchasing commercial assessment applications. Creating an accreditation system that utilizes readily available and internally produced software reduces the time burden and cost of an accreditation process. This work presents software applications and operating procedures for a system that integrates a high granularity database to record assessment test results and student work exhibits. Database records are linked to course and program educational outcomes. The database includes fields that link student work exhibits to outcomes using a standardized filename code. This work combines a testing application, assessment database, electronic student portfolio, and standardized File Name Generator into a cohesive assessment system. This work resulted in three applications; an updated tester application, a File Name Generator, and a hyperlinked database. The applications were used in a pilot program to collect data and produce reports showing percentage scores for each outcome.

Robust estimators for multivariate location and dispersion should be ãn consistent and highly outlier resistant, but estimators that have been shown to have these properties are impractical to compute. The RMVN estimator is an easily computed outlier resistant robust ãn consistent estimator of multivariate location and dispersion, and the estimator is obtained by scaling the classical estimator applied to the gRMVN subseth that contains at least half of the cases. Several robust estimators will be presented, discussed and compared in detail. The applications for the RMVN estimator are numerous, and a simple method for performing robust principal component analysis (PCA), canonical correlation analysis (CCA) and factor analysis is to apply the classical method to the gRMVN subset.h Two approaches for robust PCA and CCA will be introduced and compared by simulation studies.