Digital Commons Network^™

Artificial Neural Networks : A Comparative Study Of Implementations For Human Chromosome Classification, Nancy Akl

Theses : Honours

Artificial neural networks are a popular field of artificial intelligence and have commonly been applied to solve many prediction, classification and diagnostic tasks. One such task is the analysis of human chromosomes. This thesis investigates the use of artificial neural networks (ANNs) as automated chromosome classifiers. The investigation involves the thorough analysis of seven different implementation techniques. These include three techniques using artificial neural networks, two techniques using ANN s supported by another method and two techniques not using ANNs. These seven implementations are evaluated according to the classification accuracy achieved and according to their support of important system measures, …

End User Preference Of Customisable Features Within A Course Management System, Diana Taylor

Theses : Honours

Customisation is the means by which people alter and change various elements of their environment with the purpose of making it more suited to their immediate needs. This aids in creating a more personalised experience. People are extremely diverse in terms of age, gender, nationality, and with the dominant presence of technology people also have various levels of computer skills and experience. In the context of computer environments, customisation provides the ability to cater for a diverse user group, providing tools and options that assist with specific tasks, improve accessibility and achieve greater user satisfaction. Carter, MacLean, Lovstard & Moran …

Towards Persistent Resource Identification With The Uniform Resource Name, Luke Brown

Theses : Honours

The exponential growth of the Internet, and the subsequent reliance on the resources it connects, has exposed a clear need for an Internet identifier which remains accessible over time. Such identifiers have been dubbed persistent identifiers owing to the promise of reliability they imply. Persistent naming systems exist at present, however it is the resolution of these systems into what Kunze, (2003) calls "persistent actionable identifiers" which is the focus of this work. Actionable identifiers can be thought of as identifiers which are accessible in a simple fashion such as through a web browser or through a specific application. This …

An Investigation Of Network Security Management Methods, Shamila Anuranga Ratnayake

Theses : Honours

Network Management (NM) is concerned with reducing complexity and managing cost. The traditional NM tools and techniques are based on the Open System Interconnection (OSI) NM model. However, several drawbacks have been identified when managing a network using traditional NM tools (Sarkar & Verma, 2001). Network security is a major issue when managing a network. Even though the technology assists to reduce security risks, unless properly managed, the security measures may not do the job as expected. The State Model (SM) diagram is a new method, which may assists in managing the network. This new method may provide functionality not …

2d-Barcode For Mobile Devices, Hiroko Kato

Theses : Honours

2D-barcodes were designed to carry significantly more data than its 1D counterpart. These codes are often used in industrial information tagging applications where high data capacity, mobility, and data robustness are required. Wireless mobile devices such as camera phones and Portable Digital Assistants (PDAs) have evolved from just a mobile voice communication device to what is now a mobile multimedia computing platform. Recent integration of these two mobile technologies has sparked some interesting applications where 2D-barcodes work as visual tags and/or information source and camera phones performs image processing tasks on the device itself. One of such applications is hyperlink …

View Space Linking, Solid Node Compression And Binary Space Partitioning For Visibility Determination In 3d Walk-Throughs, Joel Anderson

Theses : Honours

Today's 3D games consumers are expecting more and more quality in their games. To enable high quality graphics at interactive rates, games programmers employ a technique known as hidden surface removal (HSR) or polygon culling. HSR is not just applicable to games; it may also be applied to any application that requires quality and interactive rates, including medical, military and building applications. One such commonly used technique for HSR is the binary space partition (BSP) tree, which is used for 3D ‘walk-throughs’, otherwise known as 3D static environments or first person shooters. Recent developments in 3D accelerated hardware technology do …

Lessons Learned With Arc, An Oai-Pmh Service Provider, Xiaoming Liu, Kurt Maly, Michael L. Nelson

Computer Science Faculty Publications

Web-based digital libraries have historically been built in isolation utilizing different technologies, protocols, and metadata. These differences hindered the development of digital library services that enable users to discover information from multiple libraries through a single unified interface. The Open Archives Initiative Protocol for Metadata Harvesting (OAI-PMH) is a major, international effort to address technical interoperability among distributed repositories. Arc debuted in 2000 as the first end-user OAI-PMH service provider. Since that time, Arc has grown to include nearly 7,000,000 metadata records. Arc has been deployed in a number of environments and has served as the basis for many other …

Weak Factorizations, Fractions And Homotopies, Alexander Kurz, Jiří Rosický

Engineering Faculty Articles and Research

We show that the homotopy category can be assigned to any category equipped with a weak factorization system. A classical example of this construction is the stable category of modules. We discuss a connection with the open map approach to bisimulations proposed by Joyal, Nielsen and Winskel.

Performance Analysis And Validation Of A Recoverable Flight Control System In A Simulated Neutron Environment, Hong Zhang, W. Steven Gray, Oscar R. Gonzalez

Electrical & Computer Engineering Faculty Publications

This paper introduces a class of stochastic hybrid models for the analysis of closed-loop control systems implemented with NASA's Recoverable Computer System. Such Recoverable Computer Systems have been proposed to insure reliable control performance in harsh environments. The stochastic hybrid models consist of either a stochastic finite-state automaton or a finite-state machine driven by a Markov input, which in turn drives a switched linear discrete-time dynamical system. Their stability and output tracking performance are analyzed using an extension of the existing theory for Markov jump-linear systems. For illustration, a stochastic hybrid model is used to calculate the tracking error performance …

A Heuristic With Bounded Guarantee To Compute Diverse Paths Under Shared Protection In Wdm Mesh Networks, Ajay Todimala, Byrav Ramamurthy

School of Computing: Conference and Workshop Papers

Establishing a fault-tolerant connection in a network involves computation of diverse working and protection paths. The Shared Risk Link Group (SRLG) [1] concept is used to model several types of failure conditions such as link, node, fiber conduit, etc. In this work we focus on the problem of computing optimal SRLG/link diverse paths under shared protection. Shared protection technique improves network resource utilization by allowing protection paths of multiple connections to share resources. In this work we propose an iterative heuristic for computing SRLG/link diverse paths. We present a method to calculate a quantitative measure that provides a bounded guarantee …

Technical Reports (1999 - 2005)

School of Computing: Technical Reports

Authors of Technical Reports (1999-2005):
Choueiry, Berthe
Elbaum, Sebastian
Goddard, Steve
Henninger, Scott
Jiang, Hong
Rothermel, Gregg
Scott, Stephen
Seth, Sharad
Soh, Leen-Kiat
Variyam, Vinodchandran

An Empirical Study Of Fault Localization For End-User Programmers, Joseph R. Ruthruff, Margaret Burnett, Gregg Rothermel

School of Computing: Conference and Workshop Papers

End users develop more software than any other group of programmers, using software authoring devices such as e-mail filtering editors, by-demonstration macro builders, and spreadsheet environments. Despite this, there has been little research on finding ways to help these programmers with the dependability of their software. We have been addressing this problem in several ways, one of which includes supporting end-user debugging activities through fault localization techniques. This paper presents the results of an empirical study conducted in an end-user programming environment to examine the impact of two separate factors in fault localization techniques that affect technique effectiveness. Our results …

A Controlled Experiment Assessing Test Case Prioritization Techniques Via Mutation Faults, Hyunsook Do, Gregg Rothermel

School of Computing: Conference and Workshop Papers

Regression testing is an important part of software maintenance, but it can also be very expensive. To reduce this expense, software testers may prioritize their test cases so that those that are more important are run earlier in the regression testing process. Previous work has shown that prioritization can improve a test suite’s rate of fault detection, but the assessment of prioritization techniques has been limited to hand-seeded faults, primarily due to the belief that such faults are more realistic than automatically generated (mutation) faults. A recent empirical study, however, suggests that mutation faults can be representative of real faults. …

Helping End-Users “Engineer” Dependable Web Applications, Sebastian Elbaum, Kalyan-Ram Chilakamarri, Bhuvana Gopal, Gregg Rothermel

School of Computing: Conference and Workshop Papers

End-user programmers are increasingly relying on web authoring environments to create web applications. Although often consisting primarily of web pages, such applications are increasingly going further, harnessing the content available on the web through “programs” that query other web applications for information to drive other tasks. Unfortunately, errors can be pervasive in web applications, impacting their dependability. This paper reports the results of an exploratory study of end-user web application developers, performed with the aim of exposing prevalent classes of errors. The results suggest that end-users struggle the most with the identification and manipulation of variables when structuring requests to …

Balancing Exploration And Exploitation: A New Algorithm For Active Machine Learning, Thomas Osugi, Deng Kun, Stephen Scott

School of Computing: Conference and Workshop Papers

Active machine learning algorithms are used when large numbers of unlabeled examples are available and getting labels for them is costly (e.g. requiring consulting a human expert). Many conventional active learning algorithms focus on refining the decision boundary, at the expense of exploring new regions that the current hypothesis misclassifies. We propose a new active learning algorithm that balances such exploration with refining of the decision boundary by dynamically adjusting the probability to explore at each step. Our experimental results demonstrate improved performance on data sets that require extensive exploration while remaining competitive on data sets that do not. Our …

Neighborhood Interchangeability And Dynamic Bundling For Non-Binary Csps, Anagh Lal, Berthe Y. Choueiry, Eugene C. Freuder

School of Computing: Conference and Workshop Papers

1. Interchangeability: An algorithm for computing interchangeability in non-binary CSPs.

2. Dynamic bundling: Integration of the above with backtrack search for solving non-binary CSPs.

3. Experiments demonstrating the benefits of dynamic bundling

·Finding multiple, robust solutions.

·Decreasing computational cost of search.

On Finding Consecutive Test Vectors In A Random Sequence For Energy-Aware Bist Design, Sheng Zhang, Sharad C. Seth, Bhargab B. Bhattacharya

School of Computing: Conference and Workshop Papers

During pseudorandom testing, a significant amount of energy and test application time is wasted for generating and for applying “useless” test vectors that do not contribute to fault dropping. For low-power testing, modification logic/ROM may be used to skip the LFSR states that generate useless test patterns. The overhead of extra logic increases rapidly with the number of such jumps. Since identification of useless patterns strongly depends on the order in which incremental fault simulation is performed, an elegant solution to this problem would be to find a minimum set of segments in the LFSR sequence, where each segment corresponds …

Profiling Deployed Software: Strategic Probe Placement, Madeline Diep, Sebastian Elbaum, Myra B. Cohen

School of Computing: Technical Reports

Profiling deployed software provides valuable insights for quality improvement activities. The probes required for profiling, however, can cause an unacceptable performance overhead for users. In previous work we have shown that significant overhead reduction can be achieved, with limited information loss, through the distribution of probes across deployed instances. However, existing strategies for probe distribution do not account for several relevant factors: acceptable overheads may vary, the distributions to be deployed may be limited, profiled events may have different likelihoods, and the user pool composition may be unknown. This paper evaluates strategies for probe distribution while manipulating these factors through …

Disseminating Usability Design Knowledge Through Ontology-Based Pattern Languages, Scott Henninger, Padmapriya Ashokkumar

School of Computing: Technical Reports

Usability patterns represent knowledge about known ways to design graphical user interfaces that are usable and meet the needs and expectations of users. There is currently a plethora of usability patterns published in books, private repositories and the World-Wide Web. The dominance of pattern discovery efforts has neglected the emerging need to organize the patterns so they can become a proactive resource for developing interfaces. This paper presents an approach using Semantic Web concepts that turns informal patterns into formal representations capable of supporting systematic design methods. Through this method, loosely coupled pattern collections can be turned into strongly coupled …

Real-Time Divisible Load Scheduling For Cluster Computing, Xuan Lin, Ying Lu, Jitender S. Deogun, Steve Goddard

School of Computing: Technical Reports

Cluster Computing has emerged as a new paradigm for solving large-scale problems. To enhance QoS and provide performance guarantees in cluster computing environments, various workload models and real-time scheduling algorithms have been investigated. The divisible load model, propagated by divisible load theory, models computations that can be arbitrarily divided into independent pieces and provides a good approximation of many real-world applications. However, researchers have not yet investigated the problem of providing performance guarantees to divisible load applications. Two contributions are made in this paper: (1) divisible load theory is extended to compute the minimum number of processors required to meet …

Dgkd: Distributed Group Key Distribution With Authentication Capability, Pratima Adusumilli, Xukai Zou, Byrav Ramamurthy

School of Computing: Conference and Workshop Papers

Group key management (GKM} is the most important issue in secure group communication (SCC). The existing GKM protocols fall into three typical classes: centralized group key distribution (CGKD), decentralized group key management (DGKM), and distributed/contributory group key agreement (CGKA). Serious problems remains in these protocols, as they require existence of central trusted entities (such as group controller or subgroup controllers), relaying of messages (by subgroup controllers), or strict member synchronization (JOT multiple round stepwise key agreement), thus suffering from the single point of failure and attack, performance bottleneck, or mis-operations in the situation of transmission delay or network failure. In …

Crtdh: An Efficient Key Agreement Scheme For Secure Group Communications In Wireless Ad Hoc Networks, Ravi K Balachandran, Byrav Ramamurthy, Xukai Zou, N.V. Vinodchandran

School of Computing: Conference and Workshop Papers

As a result of the growing popularity of wireless networks, in particular ad hoc networks, security over such networks has become very important. In this paper, we study the problem of secure group communications (SGC) and key management over ad hoc networks. We identify the key features of any SGC protocol for such networks. We also propose an efficient key agreement scheme for SGC. The scheme solves two important problems that exist in most current SGC schemes: requirement of member serialization and existence of a central entity. Besides this, the protocol also has many highly desirable properties such as contributory …

Same-Destination-Intermediate Grouping Vs. End-To-End Grouping For Waveband Switching In Wdm Mesh Networks, Mengke Li, Wang Yao, Byrav Ramamurthy

School of Computing: Conference and Workshop Papers

We investigate waveband switching (WBS) with different grouping strategies in wavelength-division multiplexing (WDM) mesh networks. End-to-end waveband switching (ETEWBS) and same-destination-intermediate waveband switching (SD-IT-WBS) are analyzed and compared in terms of blocking probability and cost savings. First, an analytical model for ETEWBS is proposed to determine the network blocking probability in a mesh network. For SD-IT-WBS, a simple waveband switching algorithm is presented. An analytical model to determine the network blocking probability is proposed for SD-IT-WBS based on the algorithm. The analytical results are validated by comparing with simulation results. Both results match well and show that ETE-WBS slightly outperforms …

A Novel Cost-Efficient On-Line Intermediate Waveband-Switching Scheme In Wdm Mesh Networks, Mengke Li, Wang Yao, Byrav Ramamurthy

School of Computing: Conference and Workshop Papers

Waveband switching (WBS) is an important technique to save switching and transmission cost in wavelength -division multiplexed (WDM) optical networks. A cost-efficient WBS scheme would enable network carriers to increase the network throughput (revenue) while achieving significant cost savings. We identify the critical factors that determine the WBS network throughput and switching cost and propose a novel intermediate waveband switching (IT-WBS) algorithm, called the minimizing-weighted-cost (MWC) algorithm. The MWC algorithm defines a cost for each candidate route of a call. By selecting the route with the smallest weighted cost, MWC balances between minimizing the call blocking probability and minimizing the …

Pre-Computation Approach To Nonlinear Simulations Of Deformable Objects, Jernej Barbic

Link Foundation Modeling, Simulation and Training Fellowship Reports

My research area involves building real-time interactive simulations of solid deformable objects. For example, liver tissue should defo1m realistically during a vi1tual surgery simulation. A virtual assembly simulation can catch airplane construction design flaws early on during the manufacturing process. Deformable objects are also useful in computer games, for example to model the motion of trees in the wind. Researchers in computational physics, mechanics, and applied mathematics have been developing algorithms for simulations of deformable objects for the past 40 years. However, the partial differential equations of solid continuum mechanics that govern deformations of physical objects are very involved and …

Applying Scenario-Based Design And Claim Analysis To The Design Of A Digital Library Of Geography Examination Resources, Yin-Leng Theng, Dion Hoe-Lian Goh, Ee Peng Lim, Zehua Liu, Ming Yin, Natalie Lee-San Pang, Patricia Bao-Bao Wong

Research Collection School Of Computing and Information Systems

This paper describes the application of Carroll’s scenario-based design and claims analysis as a means of refinement to the initial design of a digital library of geographical resources (GeogDL) to prepare Singapore students to take a national examination in geography. GeogDL is built on top of G-Portal, a digital library providing services over geospatial and georeferenced Web content. Beyond improving the initial design of GeogDL, a main contribution of the paper is making explicit the use of Carroll’s strong theory-based but undercapitalized scenario-based design and claims analysis that inspired recommendations for the refinement of GeogDL. The paper concludes with an …

A Multi-Agent Approach For Solving Optimization Problems Involving Expensive Resources, Hoong Chuin Lau, H. Wang

Research Collection School Of Computing and Information Systems

In this paper, we propose a multi-agent approach for solving a class of optimization problems involving expensive resources, where monolithic local search schemes perform miserably. More specifically, we study the class of bin-packing problems. Under our proposed Fine-Grained Agent System scheme, rational agents work both collaboratively and selfishly based on local search and mimic physics-motivated systems. We apply our approach to a generalization of bin-packing - the Inventory Routing Problem with Time Windows - which is an important logistics problem, and demonstrate the efficiency and effectiveness of our approach.

Neural Network-Based Control Of Nonlinear Discrete-Time Systems In Non-Strict Form, Jagannathan Sarangapani, Zheng Chen, Pingan He

Electrical and Computer Engineering Faculty Research & Creative Works

A novel reinforcement learning-based adaptive neural network (NN) controller, also referred as the adaptive-critic NN controller, is developed to deliver a desired tracking performance for a class of non-strict feedback nonlinear discrete-time systems in the presence of bounded and unknown disturbances. The adaptive critic NN controller architecture includes a critic NN and two action NNs. The critic NN approximates certain strategic utility function whereas the action neural networks are used to minimize both the strategic utility function and the unknown dynamics estimation errors. The NN weights are tuned online so as to minimize certain performance index. By using gradient descent-based …

A Multi-Hypothesis Approach For Salient Object Tracking In Visual Surveillance, Filiz Bunyak, Ilker Ersoy, S. R. Subramanya

Electrical and Computer Engineering Faculty Research & Creative Works

In this paper, a multiple object tracking method for visual surveillance applications is presented. Moving objects are detected by adaptive background subtraction and tracked by using a multi-hypothesis testing approach. Object matching between frames is done based on proximity and appearance similarity. a new confidence measure is assigned to each possible match. This information is arranged into a graph structure where vertices represent blobs in consecutive frames and edges represent match confidence values. This graph is later used to prune and refine trajectories to obtain the salient object trajectories. Occlusions are handled through position prediction using Kalman filter and robust …

The Inns President's Welcome, Donald C. Wunsch

Electrical and Computer Engineering Faculty Research & Creative Works

No abstract provided.