Mathematics & Computing

# Department of Mathematics and Computing Seminars

Subscribe to future seminars

## Seminars

Filter by:

### Past Seminars

Speaker: Some Mathematical Connections with Sport Neville de Mestre  (Bond University) D109, Thursday 9 May, 2013, 11:00am to 12:00noon Mathematics Show Abstract The role of mathematics in sport is helping our sports’ scientists, coaches and competitors to improve in many ways their potential to achieve success. The talk will focus on many sports and their mathematical application at the primary, secondary and tertiary level of education, as well as the research being carried out in various areas. There will be ideas put forward that are accessible to all persons interested in sport, no matter their particular level of mathematical ability. Sports covered will include tennis, darts, football, cricket, baseball, athletics, rowing, golf, horse-racing and triathlon. Hide Abstract
Speaker: Research Issues and Challenges on Brain Informatics Ning Zhong  (Maebashi Institute of Technology, Japan, ) D109, Thursday 18 April, 2013, 11:00am to 12:00noon Other http://kis-lab.com/zhong/ () Show Abstract Brain Informatics (BI) is a new interdisciplinary and multidisciplinary field that focuses on studying the mechanisms underlying the human information processing system. It brings together researchers and practitioners from diverse fields to explore the main research problems that lie in the interplay between the studies of human brain and the research of informatics, by using powerful equipments, including functional magnetic resonance imaging (fMRI), electroencephalogram (EEG), positron emission tomography (PET), and eye-tracking. The systematic BI methodology has resulted in the big BI data, including various raw brain data, data-related information, extracted data features, found domain knowledge related to human intelligence, and so forth. In this talk, I demonstrate a systematic approach to an integrated understanding of macroscopic and microscopic level working principles of the brain by means of experimental, computational, and cognitive neuroscience studies, as well as utilizing advanced Web intelligence centric information technologies. I discuss research issues and challenges from three aspects of Brain Informatics studies that deserve closer attention: systematic investigations for complex brain science problems, new information technologies for supporting systematic brain science studies, and Brain Informatics studies based on Web intelligence research needs. These three aspects offer different ways to study traditional cognitive science, neuroscience, mental health, and artificial intelligence.Hide Abstract
Speaker: Evaluating the use of online synchronous communication to enhance learning in statistics Christine McDonald  (Mathematics and Computing, USQ) D109, Thursday 28 March, 2013, 11:00am to 12:00noon Mathematics Show Abstract This seminar will report on the findings of a study which investigated the impact of synchronous technology on distance teaching and learning through a case study of an introductory statistics course offered at the University of Southern Queensland over the Summer of 2010-11. The researcher, as participant-observer, conducted weekly online tutorials with twelve volunteers who interacted using audio, text chat and a virtual whiteboard. Analysis of interactions using the Community of Inquiry model revealed teaching, social and cognitive presences. Surveys and interviews ascertained student perceptions of the value of the tutorials. Recommendations for practice and new insights into the experience of synchronous online teaching and learning will be presented.Hide Abstract
Speaker: Multi-Layered Design of Mobile Transport Networks David Fatseas  (, ) D109, Thursday 14 March, 2013, 11:00am to 12:00noon Other Show Abstract In order to rapidly and cost effectively respond to ever increasing data traffic demands, Mobile Network Operators must consider new and innovative ways to dimension, expand and evolve their Transport Networks. One approach is via the use of layers in a Transport Network to split and route the various types of traffic flows and traffic loads (based upon cost efficiency). It is envisaged that if such a proposed Multi-Layered network routing approach is adopted and deployed into Mobile Transport Networks, it has the potential to provide significant cost benefits to a Mobile Network Operator by reducing or removing the need for capital expenditure (for example: deployment of additional transmission hardware can be postponed). This proposal is also supported by a proposed Multi-Layered Network routing approach that attempts to traffic flows in the most cost efficiently manner via the use of shortest path, lowest cost based routing algorithms (such as those used in the NETML software).Hide Abstract
Speaker: Intelligent Data Analysis and Its Application in Water Resource Ying Li  () D109, Thursday 29 November, 2012, 11:00am to 12:00noon Other Show Abstract This talk will give the brief introduction of modelling methodologies used in Water Resource Group of MDBA, as well as discuss the intelligent data analysis methods with the possibilities of implementation in water modeling at Murray-Darling Basin. A case study of NN-based prediction modeling for river water quality of the East River of China will be provided.Hide Abstract
Speaker: Smart Forms Radiya Abolkhairat  () D109, Thursday 25 October, 2012, 11:00am to 12:00noon Other Show Abstract SmartForms are efficient and powerful electronic forms that could be used for the interactions between end users and web applications systems. This research has two aims. The first is representing a comprehensive study of the existing electronic forms technologies. The second is building a declarative system that automates translating of XML schema/document to an electronic form and generating data storage. The proposed system would be built using XML technologies and one of the existing electronic forms technologies such as XForms, HTML5forms and XSLTForms. It is expected from this research to produce deep understanding of the existing electronic forms features and a system that automatically generates electronic form and data storage from an input XML schema/document.Hide Abstract
Speakers: Numerical study of nonlinear wave processes by means of discrete chain models M. Obregon, N. Raj, Yury Stepanyants, Expand M. Obregon  () N. Raj  () Yury Stepanyants  (Australian Nuclear Science and Technology Organisation, Sydney) Simplify D109, Thursday 27 September, 2012, 11:00am to 12:00noon Other Show Abstract We show that many nonlinear wave processes in dispersive and dissipative media can be numerically studied by means of chain models describable by sets of ODEs. This allows us to obtain higher accuracy results for relatively cheap price using standard ODE-solvers. The advantages of this approach in comparison with the direct numerical modeling of PDEs are: the chain model has a vivid physical meaning and can be created in laboratory conditions in various embodiments playing a role of an analogous computer; there is no need to develop a complex numerical scheme and study its stability and convergence. We demonstrate the idea in application to the modeling of solitary wave propagation in a rotating ocean described by Gardner–Ostrovsky PDE using a modified Toda chain model. We show that our results are in a good agreement with early published data obtained from the direct numerical solution of the Gardner–Ostrovsky equation.Hide Abstract
Speaker: Steady and Unsteady Free Surface Flow Past a Two-Dimensional Stern Osama Nasser Ogilat  (Department of Mathematics and Computing, Faculty of Sciences, USQ) D109, Thursday 13 September, 2012, 12:00noon to 1:00pm Mathematics Show Abstract This research is concerned with the steady and unsteady two-dimensional free surface flows past semi-infinite surface-piercing bodies in a fluid of finite depth. The problems studied can be used to model the flow of water behind the bow or stern of a wide, blunt ship. The aim of this research is to find a shape of the free surface profile created behind bodies and find conditions which allow one to minimise or completely eliminate the downstream waves. Different families of stern shapes are considered, and it is shown that the amplitude of the waves can be minimised indeed. It is found that for sterns of some specific shapes the downstream waves can be entirely eliminated. The results are generalised for two-layer semi-infinite fluid of different densities when the depth of the lower layer is finite, whereas the depth of the upper layer is infinite. The steady two-dimensional free surface flow past bodies in a fluid of finite depth is investigated analytically using the WienerHopf technique and numerically using the boundary integral technique. Furthermore, the weakly nonlinear solution is investigated using the forced Kortewegde Vries equation to describe the flow in the shallow water approximation. The unsteady two-dimensional free surface in a fluid of finite depth is investigated analytically using Laplace transform and the WienerHopf technique for Froude numbers less than one. It is demonstrated that the unsteady solution asymptotically approaches the steady state solution at large time. The results obtained can be used in the engineering practice of ship building to optimise the shape of ship hull which provides a minimal wave drag.Hide Abstract
Speakers: Test of Hypotheses for Linear Models with Non-Sample Prior Information Budi Pratikno, Shahjahan Khan, Expand Budi Pratikno  (Student, Dept. Mathematics and Computing, ) Shahjahan Khan  (Mathematics and Computing, USQ) Simplify K535, Wednesday 30 May, 2012, 1:00pm to 2:00pm Statistics Show Abstract Inferences about population parameters could be improved using non-sample prior information (NSPI). The NSPI can be classified as: (i) unspecified, (ii) specified, and (iii) uncertain (or unsure). For the three different scenarios, three different statistical tests: (i) unrestricted test (UT), (ii) restricted test (RT) and (iii) preliminary test test (PTT) defined. This study tests: (1) the intercept of the simple regression model (SRM) when there is NSPI on the slope, (2) the intercept vector of the multivariate simple regression model (MSRM) when there is NSPI on the slope vector, (3) a subset of regression parameters of the multiple regression model (MRM) when NSPI on another subset is available, and (4) the equality of the two intercepts of the parallel regression model (PRM) when the NSPI on the equality of the two slopes is available. The test statistics, their sampling distributions, and power functions are derived. Comparison of power functions are used to recommend a best test. For the computation of the power of the PTT the probability integral of the bivariate normal, Student’s t and noncentral F distributions are used. The R package is used for all computations and graphical analyses. The study shows that the PTT attains a reasonable dominance over the UT and RT for all four modelsHide Abstract
Speakers: A Mathematical Model of Financial Interventions in An Economy Hirdyesh Bhatia, Ron Addie, Expand Hirdyesh Bhatia  () Ron Addie  (Mathematics and Computing, USQ) Simplify K535, Wednesday 23 May, 2012, 12:00noon to 1:00pm Other Show Abstract We explore ways to clarify, mathematically, the perennial debate in economics over whether government deficit funded spending will cause inflation or growth -- while accepting the proposition that the economy of a nation as large as Australia is too complex to be accurately modelled mathematically.Hide Abstract
Speaker: An experience of teaching and learning in the cloud Ron Addie  (Mathematics and Computing, USQ) D109, Thursday 10 May, 2012, 11:00am to 12:00noon Other Show Abstract In a physical class-room it is sometimes handy to walk from student to student, looking at the written work of students, assisting each student in a different way. Now we can do that with our online students, if their work is "in the cloud". Recent software development tools, like Google Web Toolkit and SVG enable us to develop software which runs in a browser, producing and managing documents in the cloud, which has virtually all the functionality we expect of a program written for and running on a standalone computer. The documents being on an Internet server can then be used to advantage. When the student's documents are in the cloud, the teacher, groups of students can share work, compare documents, and learn from each other, just like being in physical contact. This seminar will describe some experience with software for network analysis and design, the netml system, which can and has be used in this way.Hide Abstract
Speaker: Steady and Unsteady Free Surface Flow Past a Two-Dimensional Stern Osama Nasser Ogilat  (Department of Mathematics and Computing, Faculty of Sciences, USQ) D109, Thursday 19 April, 2012, 12:00noon to 1:00pm Other Show Abstract The This research is concerned with the steady and unsteady two-dimensional free surface flows past semi-infinite surface-piercing bodies in a fluid of finite depth. It is assumed that the fluid is incompressible and the flow is irrotational, and the effect of viscosity and surface tension are negligible. The problems considered are physically important, since they can be used to model the flow of water near the bow or stern of a wide, blunt ship. The aim of this research is to find the shape of the free surface created behind bodies and to use this knowledge to design a two-dimensional ship stern that minimises, or eliminates, the downstream waves. Different families of plate shapes are considered, and it is shown that the amplitude of the waves can be minimised. For plates that increase in height as a function of the direction of flow, reach a local maximum, and then point slightly downwards at the point at which the free surface detaches, it appears the downstream wavetrain can be eliminated entirely. The steady two-dimensional free surface flow past bodies in a fluid of finite depth is investigated analytically using the Wiener-Hopf technique and numerically using the boundary integral technique. Furthermore, the weakly nonlinear solution is investigated using the forced Korteweg-de Vries equation (KDV) to describe the flow of the free surface. The problem of the free surface flow past a semi-infinite curved plate for the subcritical case when the Froude number F < 1 is solved. Linear problem is formulated under the assumption that the elevation of the plate is close to the undisturbed free surface level. The unsteady two-dimensional free surface in a fluid of finite depth is investigated analytically using Laplace transform and the Wiener-Hopf technique. Apart from understanding of the physics, the analytical solution has shed some light on. solving a small and long time problem in the engineering practice of ship building, the optimisation of hull shape. The linear problem is formulated by assuming that the free surface is slightly perturbed a distance d by the stern, and is then solved for the case of the flat plate for the subcritical case F < 1. It is found that the unsteady solution approaches the steady state solution as t -> ∞.Hide Abstract
Speaker: Adiabatic decay of internal solitons in the oceans due to Earth’ rotation Yury Stepanyants  (Mathematics and Computing, USQ) D109, Thursday 12 April, 2012, 11:00am to 12:00noon Other Show Abstract The adiabatic decay of different types of internal wave solitons caused by the Earth’ rotation is studied within the framework of the Gardner–Ostrovsky equation. The governing equation describing such processes includes quadratic and cubic nonlinear terms, as well as Boussinesq and Coriolis dispersions: (u_t + c u_x + α u u_x + α_{1}u^{2}u_{x} + β u_{xxx})_{x} = γ u. It is shown that at the early stage of evolution solitons gradually decay under the influence of weak Earth’ rotation proportional to the parameter γ. The characteristic decay time is derived for different types of solitons for positive and negative parameter α1 (both signs of that parameter may occur in the oceans). It is found that the adiabatic theory describes well the decay of solitons of moderate amplitudes having bell-shaped profiles. In contrast to that, large amplitude table-top solitons, which can exist when α1 is negative, are structurally unstable. Under the influence of Earth’ rotation they transfer first to the bell-shaped solitons, which decay then adiabatically and completely vanish in finite time. Estimates of the extinction time of internal solitons are presented for the real oceanographic conditions.Hide Abstract
Speaker: Introducing the eBay UCM for the Joomla! Platform Sam Moffatt  (eBay, USA) D109, Thursday 22 March, 2012, 11:00am to 12:00noon Computing Show Abstract Learn about the eBay Unified Content Model (UCM) and how it interacts with the Joomla! Platform. Hear how eBay is engaging the open source community to help build and strengthen the Joomla! Platform. Also included is a step by step walk through using the eBay UCM with the Joomla! Platform to build your own application and expose the power of the eBay UCM. We'll go through some of the design decisions to help develop understanding of one of the most powerful additions to the Joomla! Platform to revolutionise the way we think about building Joomla! extensions and applications.Hide Abstract
Speaker: Adapting Existing Code to a Parallel Environment (e.g. HPC): A Case Study Harry Butler  (Mathematics and Computing, USQ) D109, Thursday 24 November, 2011, 11:00am to 12:00noon Other Show Abstract The ability to run multiple versions of the same code with different input parameters is becoming increasingly important in many disciplines (e.g. parameter searching and sensitivity testing etc.). USQ's HPC computation environment is ideal for this purpose. This seminar will use two case studies from the environmental sciences to illustrate the tools available on HPC which allow this. The first case study illustrates how the same Fortran/C code can be run on multiple cores with different input parameters using UNIX scripts and the Portable Batch Submission (PBS) system. The second case study will illustrate how same Matlab function can be run on several cores with different input parameters.Hide Abstract
Speaker: Thinking about sampling variability and replication: Confidence intervals beat p values (AGR Seminar) Emeritus Professor Geoff Cumming   (Statistical Cognition Laboratory, School of Psychological Science, La Trobe University) D109, Friday 21 October, 2011, 11:00am to 12:00noon Statistics Show Abstract Understanding sampling variability is an important goal. A useful approach is to consider replication. What information do inferential techniques give about replication? Confidence intervals (CIs) do well: A 95% CI is, approximately, an 83% prediction interval for the mean of a replication experiment. A p value, however, gives almost no information about a replication result. CIs beat p values, although for very small N a CI may be misleading. I will present novel graphics and simulations, and emphasise cognition - how people read graphics and draw conclusions. There is more in my book: Understanding The New Statistics: Effect Sizes, Confidence Intervals, and Meta-Analysis, Routledge, 2012. www.thenewstatistics.comHide Abstract
Speaker: Windows, MatLab and the HPC Leigh Brookshaw  (Mathematics and Computing, USQ) D109, Thursday 13 October, 2011, 11:00am to 12:00noon Computing Show Abstract This seminar will introduce the basic Windows software UNIX and Matlab commands required for accessing USQ's HPC environment (e.g. WinSCP, Putty). The basic philosophy and structures used in parallel computation will be discussed in relationship to Matlab code design and implementation. Practical examples will be provided on the usage of Matlab's Parallel Toolbox and the Torque submission system within the HPC environment. Hide Abstract
Speaker: To Be Advised ... Linda Galligan  (Mathematics and Computing, USQ) D109, Thursday 11 August, 2011, 12:00noon to 1:00pm Mathematics
Speaker: Phase dynamics in active systems with nonlinear self-excitation Mayada Mohammed  () D109, Thursday 28 July, 2011, 12:00noon to 1:00pm Mathematics Show Abstract We focus on dynamics of phase of oscillations in a variety of chemical, physical and biological systems. An important mechanism of such dynamics is due to nonlinear self-excitation. We plan to numerically solve the phase equation (which is a partial differential equation) in 1D and 2D; investigate the dependence of solutions on the controlling parameters, derive different variants of the nonlinearly excited phase equation and explore regular and chaotic regimes of the phase dynamics.Hide Abstract
Speakers: Remote Access to Networking Laboratory for USQ IT and Related Students Yan Li, David Lai, Vijayakrishnan Venkatakrishnan, Stella Pang , Kevin-John Beasley , Expand Yan Li  (Mathematics and Computing, USQ) David Lai  (Mathematics and Computing, USQ) Vijayakrishnan Venkatakrishnan  () Stella Pang   () Kevin-John Beasley   () Simplify D109, Thursday 21 July, 2011, 11:00am to 12:30pm Computing Show Abstract Following the success of establishing USQ Cisco Networking Academy and embedding the Cisco Certified Networking Associate (CCNA) certificate training courses into USQ Networking and Security Major of the BIT Program in 2007, we have been working hard to set up a remote access facility for external students to access, configure and manipulate real networking equipment for students' hands-on practicals. The facility will also help on-campus students who are not able to come to Toowoomba campus for the lab work for various reasons. The remote access will provide all students maximal flexibility for their learning and laboratory needs. This networking remote access project is a practical application of USQ Open and Fleximode policy. It is the "pioneering" work in Australia in networking education for completely remote access and control of network equipment, and place USQ at the forefront of flexible education. In this seminar, we will introduce the project, present the networking remote access system through the demonstrations of several real-time lab sessions.Hide Abstract
Speaker: Coherent structures in environmental flows observations and modeling Prof. Philippe Fraunie (France)  () D109, Friday 8 July, 2011, 12:00noon to 1:00pm Mathematics Show Abstract Coherent structures are responsible for non homogeneous transport in atmosphere and ocean driving vertical mixing and pollutants concentration. High resolution numerical models and associated measurement tools have been recently developed including HF and VHF radar technique including interferometry. Front capturing techniques and Lagrangian trajectories are commonly performed to better localize vortices and barriers. Appropriate scales have to be identified from phycical processes in order to better underdstand and predict unexpected complex structures like filaments and reverse flows acting as retention zones for biogeochemical production. Examples are given both in atmospheric boundary layer and coastal flows Hide Abstract
Speaker: Non-cdn: an always authenticated CDN project Sam Moffatt  (Student, Department of Mathematics and Computing, USQ, ) D109, Thursday 30 June, 2011, 11:00am to 12:00noon Other Show Abstract Non-cdn is a CDN designed around the principle that every request should be authenticated. This is different to other CDN's where the default rule is to allow access to the system by default and in some situations provide limited access control. A perfect example of this is the Facebook CDN where if you have the direct URL to an item you can always retrieve it regardless of if you would be permitted to access the file via the Facebook web page. Non-CDN aims to provide the ability to authenticate and authorise every request, log access and provide the ability to detect potentially abusive behaviour from clients. The system's design is inspired by potential needs for securely distributing course content at USQ in ways that can safely provide the benefits of a CDN.Hide Abstract
Speaker: Statistical Methodology for Ordinal Data in Meta-Analysis Md Belal Hossain  (, ) D109, Thursday 23 June, 2011, 11:00am to 12:00noon Other Show Abstract Meta-analysis combines results from several independent studies. However, meta-analysis with more than two outcomes has not been carried out as frequently as with binary outcomes. Data from studies with several outcome categories are analysed by various methods including proportional odds model which requires a proportionality assumption'. Other methods estimate the effect size by collapsing the 2xL contingency table of randomised controlled trials (RCTs) into 2x2 table. Such merging causes loss of information and unrealistic reduction of spread by artificially creating two homogeneous categories. To overcome these drawbacks, we propose generalised odds ratio (GOR) as an effect measure for ordinal categorical outcomes. A meta-analysis method is developed using GOR for ordinal categorical outcomes under independent multinomial distribution. The proposed method is illustrated using data from five RCTs of anti-cholinesterase drug tacrine in patients with Alzheimer's disease and the results compared with that of other existing methods. Also a quasi-empirical Bayes method (QEBM) is developed for heterogeneous ordinal outcomes. This method is useful in identifying the extreme studies and improving the meta-analysis in the presence of heterogeneity. Three different examples of various degree of heterogeneity are also presented.Hide Abstract
Speaker: Modelling dispersion in turbulent boundary layers using centre manifold technique Fadhel Mohammed  () D109, Thursday 9 June, 2011, 11:00am to 12:00noon Mathematics Show Abstract A dynamical evolution equation governing concentration of contaminants, averaged across the flow, in channels and boundary layers can serve as an effective tool for prediction of the spreading of the contaminants in environmental and industrial situations. An accurate principle for constructing such an equation, based on centre manifold theory, has been proposed by Roberts and co-authors. However, the dynamical structure of turbulence has not been fully taken into account in the models. The current project aims to: (1) Formulate an analytical framework of the averaged transport of contaminants in turbulent boundary layers near smooth and rough substrates, (2) Derive advection-, diffusion-, dispersion- and higher-order coefficients for the contaminant transport in channels, (3) Develop and analyse averaged transport models for the flows through canopies, (4) Numerically justify the derived models by comparing solutions with solutions of the original (non-averaged) transport equations.Hide Abstract
Speaker: Effective Scheduling Algorithm for On-Demand XML Data Broadcasts in Wireless Environments Yongrui Qin  (PhD Candidate (Computing), ) D109, Thursday 26 May, 2011, 12:00noon to 1:00pm Computing Show Abstract The organization of data on wireless channels, which aims to reduce the access time of mobile clients, is a key problem in data broadcasts. Many scheduling algorithms have been designed to organize flat data on air. However, how to effectively schedule semi-structured information such as XML data on wireless channels is still a challenge. In this paper, we firstly propose a novel method to greatly reduce the tuning time by splitting query results into XML snippets and to achieve better access efficiency by combining similar ones. Then we analyse the data broadcast scheduling problem of on-demand XML data broadcasts and define the efficiency of a data item. Based on the definition, a Least Efficient Last (LEL) scheduling algorithm is also devised to effectively organize XML data on wireless channels. Finally, we study the performance of our algorithms through extensive experiments. The results show that our scheduling algorithms can reduce both access time and tuning time significantly when compared with existing workHide Abstract
Speaker: XML + GWT + SVG = Netml Ron Addie  (Mathematics and Computing, USQ) D109, Thursday 21 April, 2011, 1:00pm to 2:00pm Computing Show Abstract Netml is a system for analysing and designing networks. It enables networks to be graphically constructed and then analysed. It has been used for teaching at USQ and research for approximately 10 years, recently in collaboration with the City University of Hong Kong. In the last year, a graphical user interface has been added using the Google Web Toolkit (GWT) and Scalar Vector Graphics. Scalar Vector Graphics (SVG) provides vector graphics in an XML compatible form, and has been implemented in most browsers. The essential idea of the Netml system is that networks are an ideal application of XML technology, and that if this is effectively used, it becomes possible to develop a powerful collection of algorithms and make them available to a wide audience. The netml system has been the subject of numerous presentations at universities and conferences over the course of its development. Hide Abstract
Speaker: Successfully Publishing Research Articles in International Journals Shahjahan Khan  (Mathematics and Computing, USQ) D109, Thursday 7 April, 2011, 12:00noon to 1:00pm Other https://mediarep-prd-vip.usq.edu.au/b-repository/2011/1/STA/STA8304/Journal_publishing/Journal_publishing_-_Flash_(Large)_-_20110407_01.02.47PM.html (The slides and Camtasia record) Show Abstract Every research active academic and students under research training programs require publishing original research outputs in high impact, good quality international journals. Often publishing in peer reviewed journals require much more (art) than valuable new results or findings. Good writing and presentation skills and approaching to appropriate journal are often keys to success. This talk covers various issues related to conducting research, preparing articles and tips to publish in top rated international journals. Please note that the presentation is particularly useful for young researchers and research students who want to know about the steps and requirements of publishing in international journals. Research students from any discipline may benefit from the presentation.Hide Abstract
Speaker: A History of Access Control Systems (Master Dissertation Presentation) Sam Moffatt  (Student, Department of Mathematics and Computing, USQ, ) D109, Friday 1 April, 2011, 12:00noon to 1:30pm Computing thesisseminar-history.pdf 1294KB (A History of Access Control Systems (Slides)) Show Abstract What MULTICS feature has influenced both operating system design and process architecture? Which security model has become the standard for defining discretionary access control? What is the most pervasive attributed based access control system available today? Which open source application's security model distinguishes normal discretionary access control from the ability to view an item? All of these questions will be answered and more in a history of computer security examining different features of security models from MULTICS to Mac OS X. A great introduction to security systems in general. Hide Abstract
Speaker: A Personalized Ontology Model for Web Information Gathering Xiaohui (Daniel) Tao  (Mathematics and Computing, USQ) D109, Thursday 24 March, 2011, 12:00noon to 1:00pm Computing Personalized Ontology Learning and Mining for Web Information Gathering.pdf 4770KB ((Slides)) Show Abstract As a model for knowledge description and formalization, ontologies are widely used to represent user profiles in personalized Web information gathering. However, when representing user profiles, many models have utilized only knowledge from either a global knowledge base or a user local information. In this paper, a personalized ontology model is proposed for knowledge representation and reasoning over user profiles. This model learns ontological user profiles from both a world knowledge base and user local instance repositories. The ontology model is evaluated by comparing it against benchmark models in Web information gathering. The results show that this ontology model is successful. The details of this work may refer to the article published on IEEE Transactions on Knowledge and Data Engineering, 23(4):496--511, 2011.Hide Abstract
Speaker: Scalar description of three-dimensional flows of incompressible fluid Yury Stepanyants  (Mathematics and Computing, USQ) D109, Thursday 17 March, 2011, 12:00noon to 1:00pm Mathematics Stepanyants.pdf 1010KB (Scalar description of three-dimensional flows of incompressible viscous fluid (Presentation Slides)) Show Abstract As it is well-known, an essential progress in the investigation of flows of incompressible fluid may be achieved with the help of stream-function. Indeed, flow description by means of one scalar stream-function is much simpler than the description based on the three-dimensional vector field. Many interesting and physically important problems were solved by this way. However, the traditional usage of a stream-function is restricted by the assumption of certain symmetry of the flow: the method is applicable only to two-component flows, i.e. when the corresponding velocity field is effectively two-dimensional, e.g., plane flow. This restriction essentially limits the range of applicability of such approach. In this work we propose another approach, also based on the introduction of only one scalar function. However, we show that with this scalar function a wide class of non-stationary three-dimensional flows can be described. This class of flows includes both potential and vortex flows. In the latter case, the corresponding vorticity field may be two-component, in general. Characteristic features of such flows are described in details. Particular examples of flows are presented in the explicit form.Hide Abstract
Speaker: My Students Don’t Learn the Way I Do Michael de Raadt  (Mathematics and Computing, USQ) D109, Thursday 10 March, 2011, 12:00noon to 1:00pm Computing Show Abstract An understanding of how students learn can facilitate the design and creation of better teaching materials. Instruments that attempt to measure learning styles are many and varied, and their use can be controversial. This paper reports on a study of the learning styles of students in an introductory programming course, and a comparison with the learning style of their instructor. We conclude that it is not necessary to cover all possible learning styles, so long as students’ learning preferences include the modalities presented in materials.Hide Abstract
Speakers: Estimation of regression parameters when measurement error in explanatory variable Anwar Saqr, Shahjahan Khan, Expand Anwar Saqr  (Student, Department of Mathematics and Computing, USQ, ) Shahjahan Khan  (Mathematics and Computing, USQ) Simplify D109, Thursday 25 November, 2010, 12:00noon to 1:00pm Statistics Show Abstract This paper proposes a new estimation method for the parameters of a simple linear regression model where the explanatory variable is subject to measurement error. The new estimator of the regression parameters is based on the reflection of the available values of the explanatory variable. The method is straightforward, easy to implement, and performs much better than existing estimators. Moreover, it does not depend on any unrealistic assumptions about the value of any scale parameter or the reliability ratio or independence of measurement error and model error. The theoretical superiority of the proposed estimator is established by analytical results. An illustrative example for numerical comparisons is also included.Hide Abstract
Speaker: A history of access control systems Sam Moffatt  (Student, Department of Mathematics and Computing, USQ, ) D109, Thursday 18 November, 2010, 12:00noon to 1:00pm Computing Show Abstract What MULTICS feature has influenced both operating system design and process architecture? Which security model has become the standard for defining discretionary access control? What is the most pervasive attributed based access control system available today? Which open source application's security model distinguishes normal discretionary access control from the ability to view an item? All of these questions will be answered and more in a history of computer security examining different features of security models from MULTICS to Mac OS X. A great introduction to security systems in general.Hide Abstract
Speaker: How do gifted students approach difficult mathematics problems? Tim Passmore  (Mathematics and Computing, USQ) D109, Wednesday 3 November, 2010, 12:00noon to 1:00pm Mathematics Show Abstract Problem solving, in various forms and contexts, takes significant time in classroom activities throughout primary and secondary school. This area attracts the interest of both mathematics educators, seeking powerful activities in teaching and learning, and researchers, considering problem solving as a research domain. While much has been published on the ways in which students struggle with solving ordinary mathematics problems, there is still little detailed analysis of how gifted students tackle difficult problems. The aim of this project is to contribute to this aspect of problem-solving research and to suggest ways in which the pedagogy of problem solving could be improved. Experimental procedures for working with gifted students will be described and a theoretical framework for analysing their performance proposed. Some preliminary results will be presented.Hide Abstract
Speaker: Statistics for Everyone in Assessment, Analysis and Decision Shahjahan Khan  (Mathematics and Computing, USQ) D109, Wednesday 20 October, 2010, 1:00pm to 2:00pm Statistics Show Abstract The first World Statistics Day (WSD) is being globally on 20 October 2010. This is the first time the world is celebrating the many achievements of Statistics worldwide. Initiated by the International Statistical Institute (ISI) the event is being officially observed by the United Nations Statistical Commission (UNSC) under the theme "Celebrating the Many Achievements of Statistics: Service, Professionalism and Integrity" and the UN Statistics Division (UNSD) has call on governments to celebrate this day in co-operation with the national statistical services. This presentation focuses on various aspects of statistical sciences, particularly its increasingly diverse and wide range of applications in the modern society. It explores the role of statistics in the evaluation of the state of affairs as well as assessing the available resources to formulate and implement collective strategies to enhance national and international development. The use of statistical methods in the decision making as well as forecasting and predicting national priorities are also covered. Furthermore, some important contributions of statistics in health, environment, industrial and manufacturing sector are highlighted. Statistical measures essential to compare regional and global socio-economic development/status are discussed. Some of the remarkable impacts that statistics has made to the modern world are touched. Hide Abstract
Speaker: An IT Forensic Case Study: Artifacts from Google Wave Andrea Schwager  (Student, Dept. Mathematics and Computing, ) D109, Thursday 26 August, 2010, 12:00noon to 1:00pm Computing Show Abstract IT forensics is a technique of crime investigation of growing importance. It is already regularly used in investigations of both IT related crimes and crimes in general. Also, the range of techniques used by IT forensic professionals is undergoing rapid development due to the diversity and rapid development of devices, such as computers, readers, and smart phones, where evidence may be left by the parties directly or indirectly involved in an investigation. In this talk, the development of a tool specifically targeting the forensic investigation of artifacts due to the use of Google Wave will be described and the lessons learned from this study will be presented. This work was completed in fulfillment of the requirements of a Master of Computing at USQ. Andrea is employed in the Forensic Analysis Section of QLD Police.Hide Abstract
Speaker: Test of Hypotheses for Linear Models with Non-Sample Prior Information Budi Pratikno  (Student, Dept. Mathematics and Computing, ) D109, Thursday 12 August, 2010, 11:00am to 12:00noon Statistics Show Abstract As a common practice, classical inferences about a population parameter are usually drawn from the sample data alone. This applies to methods used in parameter estimation and hypothesis testing. Inferences about a population parameter could be improved using non-sample information from trusted sources. Such information, which is usually provided by previous studies or expert knowledge or experience of the researchers, and is not related to the sample data, is referred to as the non-sample prior information (NSPI). However, any NSPI value is likely to be uncertain (or unsure). The outcome of the preliminary testing on the uncertain NSPI is used in the hypothesis testing to improve the performance of the statistical test. The NSPI can be classified as: (i) unknown (unspecified), (ii) known (certain or specified), and (iii) uncertain if the suspected value is unsure. For the three different scenarios, three different statistical tests: (i) unrestricted test (UT), (ii) restricted test (RT) and (iii) preliminary test test (PTT) are defined. The current research is to test the intercept when NSPI is available on the slope for the (i) simple regression, (ii) multivariate simple regression, (iii) multiple regression and (iv) parallel regression models. The test statistics, their sampling distributions, and power functions are derived. Comparison of power functions are used to recommended a best test.Hide Abstract
Speakers: Valuing the Optionality in Australian fixed rate mortgages Nicholas Endres, Ron Addie, Expand Nicholas Endres  (Heritage Building Society, ) Ron Addie  (Mathematics and Computing, USQ) Simplify D109, Thursday 22 July, 2010, 12:00noon to 1:00pm Mathematics Show Abstract Existing techniques for modeling fixed rate mortgage products focus on the American market where the products are fundamentally different to those found in Australia. The limited optionality embedded in Australian fixed rate mortgages leads to uncertain cash flows which can adversely affect lenders. This presentation is a preliminary discussion about the ways to modify existing techniques to be relevant to the Australian market. In particular the partial prepayment condition will be modeled assuming rational action on behalf of the customer with the intent of minimising their liability. Cash flows will be simulated under a variable interest rate environment to determine appropriate pricing as a proxy for the option premium.Hide Abstract
Speaker: Construction and three-way ordination of the Wheat Phenome Atlas Kaye E Basford  (Head of School, Land, Crop and Food Sciences at UQ, ) D303, Thursday 29 April, 2010, 12:00noon to 1:00pm Statistics Show Abstract Long-term plant breeding programs generate large quantities of genealogical, genotypic, phenotypic and environment characterization data. Marker-trait association studies are commonly used to integrate and represent such data, but the results are context dependent, as trait associated markers depend on the germplasm investigated, the environments in which they are studied, and the interaction of genotype and environment. The concept of a trait-associated marker block, defined as markers in a linkage disequilibrium block that show significant association with a trait, can be used to reduce the number of trait-associated markers, control family-wise error rate and address non-independency of markers in association analysis. A Phenome Atlas can be constructed as a collection of diagrammatic representations of chromosome regions that affect trait inheritance (phenome maps) to document the patterns of trait inheritance across the genome. These maps demonstrate the complexity of the genotype-to-phenotype relationships and the context dependency of marker trait association patterns. This methodology is illustrated using the Wheat Phenome Atlas which represents the results of a genome wide association study of 20 economically important traits from the first 25 years of an international wheat breeding program. Three-way principal component analysis can be used to obtain information about which genotypes carry favourable trait-associated marker block combinations, which marker blocks discriminate among genotypes and which marker block combinations are available for any given combination of genotypes and traits. Different patterns of marker trait association profiles are observed when analysing the same genotypes for different marker blocks and trait combinations and for data obtained from different combinations of environments. These results emphasise the context dependency of association studies. Hide Abstract
Speaker: The Tablet PC University - where are we at? Birgit Loch  (Mathematics and Computing, USQ) W534, Tuesday 20 April, 2010, 12:00noon to 1:30pm Computing Show Abstract USQ funded a Senior Learning and Teaching Fellowship last year, the "tablet PC project", with the ambitious aim to establish USQ as a tablet PC university. Apart from the $50,000 fellowship funding, this project attracted a further$120,000 in in-kind support from Deans of all faculties, the Division of ICT Services, the Vice-Chancellor, and Toshiba. In this session, we will first give an overview of what has happened so far in the project, about outcomes, use and adoption of tablet PCs. A number of current tablet PC users from across the university will give short presentations on how their teaching has changed, and what impact the tablet PC has had on their students' learning. We will finish with a discussion of "where to from here", the future of tablet PCs at USQ. Hide Abstract
Speaker: Effects and experiences of automatically assessed exercises with visualization - the resource efficient way to teaching and learning Mikko-Jussi LAAKSO  (Department of Information Technology, University of Turku Lemminkisenkatu FIN-20520 , ) D109, Thursday 1 April, 2010, 12:00noon to 1:00pm Other Show Abstract Learning to program is one of the key requirements in computer science education. However, novice programmers often find it a difficult task to master. There isnt probably any single solution to solve this problem, but it is still important to develop and research new means and techniques to further improve the learning experience and outcome. In this seminar, the engagement taxonomy, the capabilities of automatic assessment and a visualisation tools (ViLLE and TRAKLA2) will be presented and discussed in terms of the benefits to the teacher and also for student learning. ViLLE, a dynamic program visualization tool developed at University of Turku, Finland. ViLLE is a language-independent visualization tool aiming at providing a more abstract view of programming, and it can be utilized both in lectures and for independent learning. It has a built-in syntax editor, with which the user can add new languages to the tool or modify the syntax of the built-in languages, including e.g. Java, C++ etc. The visualizations can be viewed in any of the (user or pre-) defined languages. To emphasize the language independency, ViLLE has a parallel view in which the execution of a program and the program code itself can be viewed simultaneously in two languages. While the execution progresses, the user can observe program outputs and changes in variable values. In addition, we can engage the learner with pop-up questions with automated assessment capabilities. TRAKLA2 is an environment for learning data structures and algorithms. The system provides algorithm simulation exercises that can be automatically graded. The grading is based on comparison between the learner made simulation sequence and a sequence produced by an actual algorithm. The system includes about 50 ready made exercises and is it developed and maintained by Software Visualization group at Aalto University, Helsinki, Finland. Hide Abstract
Speaker: Introductory Programming in a Web Context Michael de Raadt  (Mathematics and Computing, USQ) D109, Thursday 25 March, 2010, 12:00noon to 1:00pm Computing Intro Programming on Web Context.ppt 2121KB (Speaker Slides) http://www.sci.usq.edu.au/research/seminars/files/seminar220/Untitled.html (Speaker Video and Audio) Show Abstract A number of studies have recognised the benefits of using a context or theme consistently throughout an introductory programming course. Examples of contexts in which programming is related and taught include micro-worlds, robotics, games and media computation. Such contexts bring relevance to the content of programming courses. In this paper, a Web context is proposed and described. This context has been successfully used in an introductory programming course and received a positive student response.Hide Abstract
Speaker: Particle and bubble dynamics in a viscous fluid at small Reynolds numbers Yury Stepanyants  (Mathematics and Computing, USQ) D109, Thursday 22 October, 2009, 12:00noon to 1:00pm Mathematics Bubbles vs Particles.pdf 3134KB (Speaker Slides) Yury.mp3 8029KB (Speaker Audio) Show Abstract Dynamics of solid particles and gaseous bubbles in viscous density stratified fluid is studied analytically and numerically within the framework of creeping flow approximation. Equations of motion for particles and bubbles include the consideration of the buoyancy force, Stokes drag force and memory integral drag force. Expressions for the drag forces exerted on solid particles and gaseous bubbles are different. The problem studied is applicable to many practical situations where particle and bubble motion may be experienced in viscous fluids, which include aerosol and dust transport in the atmosphere, drop dynamics in pharmaceuticals and perfume, sediment and bubble motions in oceans and cavitation processes in technology. Exact analytical solutions for both the homogeneous surrounding fluid and density stratified fluid are obtained; they are analysed against numerical solutions. Results obtained are particularly relevant to the physical processes that may occur in the cooling system of the Open Pool Australian Light-water research reactor OPAL at Lucas Heights. Yury Stepanyants is a member of the Computational Engineering and Science Research Centre (CESRC).Hide Abstract
Speaker: Detecting Projected Outliers in High-dimensional Data Streams Ji Zhang  (Mathematics and Computing, USQ) D109, Thursday 15 October, 2009, 12:00noon to 1:00pm Computing ji.mp3 9578KB (Speaker Audio) SPOT_USQ seminar_Zhang.ppt 1037KB (Speaker Slides) Show Abstract In this talk, I will discuss the problem of projected outlier detection in high dimensional data streams. I will first discuss the formulation of this problem and the changes we will face when addressing this problem. I will also dwell on a new technique, called Stream Projected Ouliter deTector (SPOT), to identify outliers embedded in subspaces. SPOT devises Sparse Subspace Template (SST), a set of subspaces obtained by unsupervised and/or supervised learning processes, to detect projected outliers effectively. Multi-Objective Genetic Algorithm (MOGA) is employed as an effective search method for finding outlying subspaces from training data to construct SST. The experimental results demonstrate the efficiency and effectiveness of SPOT in detecting outliers in high-dimensional data streams.Hide Abstract
Speaker: Modelling Anomalous diffusion with application to Spiny Dendrites and Linear Reaction Dynamics Trevor Langlands  (Mathematics and Computing, USQ) D109, Thursday 8 October, 2009, 12:00noon to 1:00pm Mathematics trevor.mp3 8850KB (Speaker Audio) USQSeminar_Langlands.pdf 1186KB (Speaker Slides) Show Abstract In this talk I will discuss modelling anomalous subdiffusion with application to modelling potential propagation in spiny dendrites. Most of the brain is composed of dendrites and dendritic trees that relay postsynaptic potentials to a soma for processing and possible nerve cell firing. The mathematical description of the electrotonic properties of neuronal dendrites and dendritic trees is based on the cable equation, which models a dendrite as a homogeneous membrane cylinder with membrane capacitance, membrane resistance, and axial resistance as model parameters. In recent years advances in imaging technologies and experimental techniques have enabled detailed studies of the local structure and the distribution of dendritic spines - small protrusions that decorate typical dendrites of excitary nerve cells in the neocortex. The dendritic spines act as local traps that result in anomalous diffusion of molecules along the axial direction of dendrites. In an effort to model the electrotonic properties of neuronal dendrites with spines we have introduced a new fractional cable equation that incorporates anomalous electro-diffusion of ions. The model equations predict that larger spine densities can serve to enhance both the arrival time and the persistence of postsynaptic potentials that propagate along spiny dendrites to the soma. We describe a fractional cable equation to model anomalously slow electrodiffusion of ions in nerve cells. Fundamental solutions of this model are presented and results for firing rates and voltage attenuation are obtained in terms of the anomalous diffusion parameters. A particular application to model the passive propagation of a postsynaptic potential along a spiny dendrite is described. An interesting special case of this model is equivalent to a particular example of a fractional reaction-diffusion equation formulated by Henry and Wearne (2000, p.448-455). The solution in this example predicts negative solutions contrary to the physical reality for chemical reactions. This result has sparked new research into modelling reactions in the presence of anomalous subdiffusion and will form the second half of the this talk where we shall discuss potential alternative models for single and multi-species reactions in the case of linear reaction dynamics. These model equations are derived from a mesoscopic continuous time random walk formulation of anomalously diffusing species with linear mean field reaction kinetics. The effect of reactions is manifest in reaction modified spatio-temporal diffusion operators as well as in additive mean field reaction terms. One consequence of the non-separability of reaction and subdiffusion terms is that the governing evolution equation for the concentration of one particular species may include both reactive and diffusive contributions from other species. The general solution is derived for the multi-species system and some particular special cases involving both irreversible and reversible reaction dynamics are analyzed in detail. We have carried out Monte Carlo simulations corresponding to these special cases and we find excellent agreement with theory. The impact of this new research on the modelling of spiny dendrites will be discussed.Hide Abstract
Speaker: Giant internal waves in oceans and their possible impact on underwater navigation Yury Stepanyants  (Mathematics and Computing, USQ) D109, Thursday 27 August, 2009, 12:00noon to 1:00pm Mathematics Giant internal waves in oceans.pdf 4629KB (Speaker Slides) Yury.mp3 10602KB (Speaker Audio) Show Abstract This seminar will be targeted to an audience without specific mathematical knowledge. Possible effects of gigantic internal waves on underwater navigation are discussed. It has been shown that under certain conditions, intense internal waves may dangerously effect on coastal engineering constructions and submarine navigation. The possible role of large amplitude internal solitary waves in the wrecks of American submarines in 1960s will also be discussed. Typical models describing internal waves in density stratified oceans are outlined. Among them, the Korteweg−de Vries, Benjamin−Ono, Gardner, Ostrovsky, Kadomtsev−Petvialshvili and other equations. Solutions of these equations in the form of stationary propagating periodic or solitary waves are discussed. Characteristic parameters of waves are presented for shallow and deep basins. Experimental observations of giant solitary waves in different places of the World Ocean are summarized. The main cause of gigantic wave appearance is the barotropic tide which transforms into baroclininc perturbations and gives birth to internal waves. An example of such transformation of nonlinear internal tide with the consequent development of internal solitary waves on the Australian North West Shelf is discussed in detail. Influence of various factors such as nonlinearity, dispersion, the Coriolis effect due to Earth rotation, depth variation and horizontal variability of density stratification, on wave propagation are demonstrated through the numerical simulation. Simulations show that a wide variety of nonlinear wave shapes can be explained by a synergetic action of nonlinearity and variability of hydrology along the wave path. Hide Abstract
Speakers: The Progress Bar: A Simple Time-Management Tool for You and Your Students Michael de Raadt, Stijn Dekeyser, Expand Michael de Raadt  (Mathematics and Computing, USQ) Stijn Dekeyser  (Mathematics and Computing, USQ) Simplify D109, Thursday 23 July, 2009, 12:00noon to 1:00pm Other http://www.sci.usq.edu.au/research/seminars/files/ProgressBar/ProgressBar.html (Seminar Video Recording) http://www.sci.usq.edu.au/staff/deraadt/progressBar.html (Information about Progress Bar) Show Abstract The Progress Bar is a new tool you can use in Moodle (StudyDesk) within your courses. It's easy to set up and helpful to your students. This seminar will discuss: why such a tool is needed,results of evaluation and student feedback,impact on student progression, andhow you can set up the Progress Bar in your course.Hide Abstract
Speakers: Increasing power of robust test through pre-testing in parallelism model Shahjahan Khan, Rossita Mohamad Yunus, Expand Shahjahan Khan  (Mathematics and Computing, USQ) Rossita Mohamad Yunus  (Student, Mathematics and Computing, USQ, ) Simplify D109, Thursday 16 July, 2009, 12:00noon to 1:00pm Statistics Show Abstract For the p (> 1) simple linear models, the problem of testing the intercepts following a preliminary test on the suspected equality of slopes is considered. The unrestricted test (UT), restricted test (RT) and pre-test test (PTT) are proposed from the M-tests using the M-estimation methodology. The asymptotic distributions of the test statistics are obtained and used to derive the asymptotic power functions of the tests. Analytical and computational comparisons of the three tests are provided. The PTT achieves a reasonable dominance over the others asymptotically.Hide Abstract
Speakers: Statistical Methodology for Meta-Analysis Md. Belal Hossain, Shahjahan Khan, Expand Md. Belal Hossain  (Student, Mathematics and Computing, USQ, ) Shahjahan Khan  (Mathematics and Computing, USQ) Simplify D109, Thursday 25 June, 2009, 12:00noon to 1:00pm Statistics Show Abstract The odds ratio (OR) is one of the most popular and frequently used indices for measuring the extent of association between exposure and its outcomes in randomised controlled trials (RCTs). Meta-analysis combines data from various independent trials in estimating OR for binary outcomes to make the sample size larger so that the inference based on meta-analysis is reliable. However, there are situations in which the outcomes are on an ordinal scale with more than two cate- gories. The OR can not directly be used without arbitrarily grouping multiple levels of response into two categories. Moreover, the col- lapsing of data may cause a loss of eﬃciency. The generalised odds ratio (GOR) (Agresti, 1980) is used for summarising the diﬀerence between two stochastically ordered distributions of an ordinal cate- gorical variable without the need to assume any speciﬁc parametric models (Hosmer and Lemeshow, 1989). In this study, GOR is used for combining the treatment eﬀect in meta-analysis for ordered categori- cal outcomes. A quasi-empirical Bayes method is developed for RCTs using GOR under independent multinomial sampling procedure. This method will help identifying the extreme trials, hence improve the meta-analysis with heterogeneous trials. Both ﬁxed eﬀects and ran- dom eﬀects model are employed and compared.Hide Abstract
Speaker: Exchange Options under Jump-Diffusion Dynamics Gerald Cheang  (Nanyang Technological University, Singapore, ) D109, Thursday 21 May, 2009, 12:00noon to 1:00pm Mathematics USQ_sem.pdf 201KB (Speaker Slides) Show Abstract Margrabe provides a pricing formula for an exchange option where the distributions of both stock prices are log-normal with correlated Wiener components. Merton has provided a formula for the price of a European call option on a single stock where the stock price process contains a compound Poisson jump component, in addition to a continuous log-normally distributed component. We use Merton's analysis to extend Margrabe's results to the case of exchange options where both stock price processes also contain compound Poisson jump components. A Radon-Nikodym derivative process that induces the change of measure from the market measure to an equivalent martingale measure is introduced. The choice of parameters in the Radon-Nikodym derivative allows us to price the option under different financial-economic scenarios. We also consider American style exchange options and provide a probabilistic intepretation of the early exercise premium.Hide Abstract
Speaker: Research into academic numeracy Linda Galligan  (Mathematics and Computing, USQ) D109, Thursday 30 April, 2009, 12:00noon to 1:00pm Mathematics Show Abstract Academic numeracy is a critical awareness that allows students to become confident and competent in using mathematics and to be able to situate, interpret, critique, use, communicate and even create mathematics within their discipline’s setting. Academic numeracy is usually based on both the mathematics learnt at school and the mathematics needed in disciplines such as nursing or economics. Many students at both school and university are opting to study easier types of mathematics; universities are removing mathematical pre-requisites from award programs and not recognising the embedded mathematics within many of their courses. These trends ensure that initiatives which allow students to bridge the mathematical gaps to university are still necessary. Yet while research into mathematics at university does exist, it is still in its infancy. There are many questions not fully answered and as universities move into the culture of quality and performance matched funding, it is essential that we continue to address the following questions:What academic numeracy is needed by students at university?What are students' academic numeracy skills on entering university?What are the most effective ways to develop the academic numeracy skills of university students?This presentation will highlight some of the research undertaken and the curricula and resources developed by myself and others at USQ.Hide Abstract
Speaker: The Challenge of Assuring Data Trustworthiness Elisa Bertino  (CS Department Research Director, CERIAS, Purdue University, ) D109, Friday 24 April, 2009, 11:00am to 12:00noon Computing http://homes.cerias.purdue.edu/~bertino/ (Speaker Biography) Show Abstract With the increased need of data sharing among multiple organizations, such as government organizations, financial corporations, medical hospitals and academic institutions, it is critical to ensure that data be trustworthy so that effective decisions can be made based on these data. In this talk, we first discuss motivations and requirement for data trustworthiness. We then present an architectural framework for a comprehensive system for trustworthiness assurance based on data validation and metadata management. We then discuss an important issue in assuring data trustworthiness, that is, the evaluation of data provenance and propose a trust model for estimating the confidence level of the data and the trust level data providers. By taking into account confidence about data provenance, we introduce an approach for policy observing query evaluation. We highlight open research issues and research directions throughout the presentation. Biography: Elisa Bertino is professor of at the Computer at the Department of Computer Sciences, Purdue University and Research Director of CERIAS. Her main research interests cover many areas in the fields of information security and database systems. Her research combines both theoretical and practical aspects, addressing as well applications in a number of domains, such as medicine and humanities. She is author of several articles which appeared in International Journals and Conference Proceedings, and is co-author of the books: Object-Oriented Database Systems (Addison-Wesley, 1993) Indexing Techniques for Advanced Database Systems (Kluwer, 1997) Intelligent Database Systems (Addison-Wesley, 2001). Hide Abstract
Speaker: Finding Optimal Privacy-aware Path in Hippocratic Databases Hua Wang  (Mathematics and Computing, USQ) D109, Thursday 16 April, 2009, 12:00noon to 1:00pm Computing Show Abstract In this paper, we extend these mechanisms in order to support interorganizational business processes in Hippocratic databases. A comprehensive approach for negotiation of personal information between customers and enterprises based on user preferences is developed when enterprises offer their clients a number of ways to fulfill a service. We organize purposes into purpose directed graphs through AND/OR decomposition, which supports task delegations and distributed authorizations. Specially, customers have controls of deciding how to get a service fulfilled on the basis of their personal feeling of trust for any service customization. Quantitative analysis is performed to characterize privacy penalties dealing with privacy cost and customer's trust. Finally, efficient algorithms are given to guarantee the minimal privacy cost and maximal customer's trust involved in a business process.Hide Abstract
Speakers: Research in Echo Cancellation for Tele-health Ron Addie, Stephen Braithwaite, Expand Ron Addie  (Mathematics and Computing, USQ) Stephen Braithwaite  (PhD Student, Mathematics and Computing, USQ, ) Simplify D109, Thursday 2 April, 2009, 12:00noon to 1:00pm Computing Show Abstract Echos are one of the major problems in video conferencing and it is essential to control echo in order to achieve satisfactory quality of audio communication. Research on how to do this has been active for many years, but still faces important challenges. In February this year Ron Addie went on an "ADL Light" which took in three locations: Hong Kong, Graz (Austria), and Rome (Italy), and in two of these locations he gave seminars about echo cancellation and tele-health. In this seminar the highlights of these seminars and an outline of the research collaboration outcomes which emerged will be presented.Hide Abstract
Speaker: Exploring Network Issues in Real Time Avatar Dynamic Gesturing and Posturing in Multi-user Games Chew Wei Sheng  (Student, Mathematics and Computing, USQ, ) D109, Friday 27 February, 2009, 10:00am to 11:00am Computing Chew.mp3 13116KB (Speaker Audio) FinalSeminarFinal.ppt 7654KB (Speaker Slides) Wei Sheng, CHEW a Dessertation & a Seminar.zip 190461KB (Speaker Slides with Video) Show Abstract The aim of this research is two-fold. The first goal is to increase the expressive range of an existing game engine by mapping actor input via the keyboard onto a virtual performer. The original paradigm is that of a hand puppeteer, but instead of strings the actor manipulates the character with the keys of the keyboard. The second goal is to investigate the network performance of dynamic gesturing and posturing animations with a dead-reckoning calculation of 100, 500 and 1000 milliseconds latency per message. The results of the investigation are able to prove that the dead-reckoning technique offers a significant improvement of the network environment's quality when expressing dynamic gesturing and posturing animations in multi-user online virtual environment.Hide Abstract
Speaker: Cancer Classification Using Support Vector Machine Techniques Thatchanamoorthy Purnshatman  (Student, Mathematics and Computing, USQ, ) D109, Monday 12 January, 2009, 12:00noon to 1:00pm Bioinformatics Moorthy.mp3 5119KB (Speaker Audio) msc 8002 Seminar.pdf 1178KB (Speaker Slides) Show Abstract Machine learning offers a principled approach for developing sophisticated, automatic, and objective algorithms for analysis of high-dimensional and multimodal cancer data. This review focuses on several advances in the state of the art that have shown promise in improving detection, diagnosis, and therapeutic monitoring of disease. Key in the advancement has been the development of a more in-depth understanding and theoretical analysis of critical issues related to algorithmic construction and learning theory. These include trade-offs for maximizing generalization performance, use of physically realistic constraints, and incorporation of prior knowledge and uncertainty. The developments in machine learning, focusing on supervised and unsupervised linear methods and Support vector machine inference, which have made significant impacts in the detection and diagnosis of cancer. The approach of this master thesis consist support vector machine (SVM) techniques for cancer classification. As the SVM has many unique properties, examine the interpretation of support vector models with respect to cancer data. Support Vector Machines are especially appropriate for this classification task because of their capacity to learn non-linear classification functions from weakly and inconsistently labeled and noisy training data, as it is often seen in medical imaging. The SVM is able to learn the typical patterns for each class. When applying the SVM on a new volume, the SVM returns a confidence value or margin value for each voxel. Using a threshold filter, each voxel can be classified as tumor or non tumor. The separate analysis of each voxel might help to support the radiologist to find small tumors which might otherwise go unnoticed in the high information containing 3D images and to support him in the analysis of the tumors. Such a tool for a fast and high sensitive automatic detection is regarded as substantial help to support the radiologist. The main objective of this thesis is to extract high quality rules from trained SVMs. The rule's quality is measured in terms of accuracy, fidelity, comprehensibility and consistency. The rules will provide an explanation for the decision made by the SVM model (classifier) and can also be used for classification.Hide Abstract
Speaker: Conditional Purpose Based Access Control Model for Privacy Protection Md. Enamul Kabir  (Student, Mathematics and Computing, USQ, ) D109, Thursday 8 January, 2009, 12:00noon to 1:00pm Computing kabir.mp3 5849KB (Speaker Audio) presentation.ppt 234KB (Speaker Slides) Show Abstract This paper presents a model for privacy preserving access control which is based on variety of purposes. Conditional purpose is applied along with allowed purpose and prohibited purpose in the model. It allows users using some data for certain purpose with conditions. The structure of conditional purpose based access control model is defined and investigated through a practical paradigm with access purpose and intended purpose. An algorithm is developed to achieve the compliance computation between access purposes and intended purposes. According to this model, more information from data providers can be extracted while at the same time assuring privacy that maximizes the usability of consumers' data. This model extends traditional access control models to a further coverage of privacy preserving in data mining atmosphere. Its interior is a new structure for managing collected data in an effective and trustworthy way. This structure helps enterprises to circulate clear privacy promise, to collect and manage user preferences and consent. The implementation of the idea in the paper shows the flexibility of the model, and finally we provide comparisons of our work to other related work.Hide Abstract
Speaker: Privacy-aware Access Control with Generalization Boundaries Min Li  (PhD Student, Department of Mathematics and Computing, USQ, ) D109, Thursday 18 December, 2008, 1:00pm to 2:00pm Computing Show Abstract Privacy is today an important concern for both data providers and data users. Data generalization can provide significant protection of an individual's privacy, which means the data value can be replaced by a less specific but semantically consistent value and the personal information can be collected in a generalized form. However, over-generalized data may render data of little value. A key question is whether or not a certain generalization strategy provides a sufficient level of privacy and usability? In this paper, we introduce a new approach, called privacy-aware generalization boundaries, which can satisfy the requirements of both data providers and data users. We propose a privacy-aware access control model related to a retention period. Formal definitions of authorization actions and rules are presented. Further, we discuss how to manage a valid access process and analysis the access control policy.Finally, we extend our model to support highly complex privacy-related policies by taking into account features of obligations and conditions. Hide Abstract
Speaker: Microdata Protection Through Approximate Microaggregation Xiaoxun Sun  (PhD Student, Department of Mathematics and Computing, USQ, ) D109, Thursday 18 December, 2008, 12:00noon to 1:00pm Computing Show Abstract Microaggregation Microdata protection is a hot topic in the field of Statistical Disclosure Control, which has gained special interest after the disclosure of 658000 queries by the America Online (AOL) search engine in August 2006. Many algorithms, methods and properties have been proposed to deal with microdata disclosure. One of the emerging concepts in microdata protection is k-anonymity, introduced by Samarati and Sweeney. k-anonymity provides a simple and efficient approach to protect private individual information and is gaining increasing popularity. k-anonymity requires that every record in the microdata table released be indistinguishably related to no fewer than k respondents. In this paper, we apply the concept of entropy to propose a distance metric to evaluate the amount of mutual information among records in microdata, and propose a method of constructing dependency tree to find the key attributes, which we then use to process approximate microaggregation. Further, we adopt this new microaggregation technique to study k-anonymity problem, and an efficient algorithm is developed. Experimental results show that the proposed microaggregation technique is efficient and effective in the terms of running time and information loss.Hide Abstract
Speakers: Effects of White Matter on EEG of Multi-layered Spherical Head Models Md. Rezaul Bashar , Yan Li, Peng Wen, Expand Md. Rezaul Bashar   (Student, Department of Mathematics and Computing, USQ, ) Yan Li  (Mathematics and Computing, USQ) Peng Wen  (Faculty of Engineering, USQ, ) Simplify D109, Thursday 4 December, 2008, 12:00noon to 1:00pm Bioinformatics Show Abstract Biological tissues are multi-compartmental inhomogeneous media composed of different cellular and subcellular domains. Human head, a multi-compartmental inhomogeneous medium, is composed of scalp, skull, cerebrospinal fluid (CSF), gray matter (GM), white matter (WM), and other subcellular domains. Among these domains, skull and WM show the complicated anisotropy tissue property because of their physiological structure. However, many researchers model human head excluding WM. This research investigates the necessity of WM using four- and five- layered spherical head models. Four-layered head model excludes WM while five-layered model includes it. The piecewise homogeneous forward model using finite element method is implemented to measure electroencephalogram (EEG) on head surface for both head models. Analyzing these EEGs, this research finds the necessity of using WM to make an accurate human head model.Hide Abstract
Speaker: Priority Driven K-Anonymisation for Privacy Protection Xiaoxun Sun  (PhD Student, Department of Mathematics and Computing, USQ, ) D109, Thursday 20 November, 2008, 12:00noon to 1:00pm Computing AusDM08_Presentation.ppt 478KB (Speaker Slides) Xiaoxun.mp3 7791KB (Speaker Audio) Show Abstract Given the threat of re-identification in our growing digital society, guaranteeing privacy while providing worthwhile data for knowledge discovery has become a difficult problem. k-anonymity is a major technique used to ensure privacy by generalizing and suppressing attributes and has been the focus of intense research in the last few years. However, data modification techniques like generalization may produce anonymous data unusable for medical studies because some attributes become too coarse-grained. In this paper, we propose a priority driven k-anonymisation that allows to specify the degree of acceptable distortion for each attribute separately. We also define some appropriate metrics to measure the distance and information loss, which are suitable for both numerical and categorical attributes. Further, we formulate the priority driven k-anonymisation as the k-nearest neighbor (KNN) clustering problem by adding a constraint that each cluster contains at least k tuples. We develop an efficient algorithm for priority driven k-anonymisation. Experimental results show that the proposed technique causes significantly less distortions.Hide Abstract
Speaker: The procedural generation of indoor environments for computer games Arnaud Couturier  (Student, Mathematics and Computing, USQ, ) D109, Thursday 13 November, 2008, 12:00noon to 1:00pm Computing w0062507_MSC8002_s2_2008_oral_presentation.pdf 3258KB (Speaker Slides) Show Abstract As computer and console games become more complex, the manual level creation process also becomes more complex, time-consuming and expensive. In this research thesis, a method for automatically creating indoor levels for action games is presented. The output levels should include everything that is necessary for them to be ready to play, with minimum or no human intervention. This technique could reduce game development costs, reduce level storage requirements, and could offer new opportunities in terms of gameplay.Hide Abstract
Speakers: Workshop on the use of Sharepoint for Managing Documents in Educational Settings Ron Addie, Robert Wood, Andrea Schwager, Cynthia Wong, Stijn Dekeyser, Graeme Russell, Expand Ron Addie  (Mathematics and Computing, USQ) Robert Wood  (Strategic Information Systems, USQ, ) Andrea Schwager  (Student, Dept. Mathematics and Computing, ) Cynthia Wong  (Student, Mathematics and Computing, USQ, ) Stijn Dekeyser  (Mathematics and Computing, USQ) Graeme Russell  (Faculty of Sciences Technical Section, USQ, ) Simplify D109, Thursday 30 October, 2008, 12:00noon to 1:00pm Computing CynthiaSharePoint.ppt 166KB (Speaker Slides) MicrosoftSharepoint2.ppt 752KB (Speaker Slides) http://www.sci.usq.edu.au/research/seminars/files/seminar198/Sharepoint Seminar/Sharepoint Seminar.htm (Seminar Video Recording) Show Abstract Sharepoint is being increasingly used in education settings as a mechanism for housing documents and securely managing the changes to documents, and the workflow which these documents support. Sharepoint is already effectively serving an important need of educational institutions and industry, even though the precise nature of the services that it provides would probably be described quite differently depending on the individual situation. This seminar/worshop aims to provide a venue for IT and other staff who are using sharepoint in the educational setting to share ideas. The participants and speakers come from a diverse range of backgrounds: academic, IT support, and administrative support staff will speak about the use of sharepoint from their own experience.Hide Abstract
Speaker: Tablet PCs and screen recording Birgit Loch  (Mathematics and Computing, USQ) D109, Thursday 28 August, 2008, 1:00pm to 2:00pm Other Show Abstract In this informal presentation I will demonstrate what a tablet PC is and how it can be used for on campus and distance teaching and electronic marking of assignments. I will also explain what is needed to record screen movement (this does not require a tablet PC). Anyone interested is welcome - this presentation will be at a very basic level for people who know nothing about this technology. There will be plenty of time for questions.Hide Abstract
Speakers: Moodle as a Context for Programming Assignments Yajuvendrasinh V Mahida, Michael de Raadt, Expand Yajuvendrasinh V Mahida  (Student, Department of Mathematics and Computing, ) Michael de Raadt  (Mathematics and Computing, USQ) Simplify D109, Thursday 28 August, 2008, 12:00noon to 1:00pm Computing Moodle Block Presentation.ppt 1222KB (Speaker Slides) http://www.sci.usq.edu.au/research/seminars/files/seminar195/Presentation/Presentation.html (Seminar Video Recording) Show Abstract Moodle is the world's most widely used LMS and USQ now has Australia's largest Moodle install. Being modular and open source, Moodle allows members of the education community to contribute to the system and its development. Moodle is also a great context for students to conduct real development work and learn skills that will be valuable in the education sector. In this seminar we present a student experience of a Moodle development project and report on the potential for more such projects and the benefits that can come from this.Hide Abstract
Speaker: New case of chaotic behaviour in reaction-diffusion systems Dmitry Strunin  (Mathematics and Computing, USQ) D109, Thursday 31 July, 2008, 12:00noon to 1:00pm Mathematics Show Abstract Cellular slime moulds, catalytically oxidating CO on metal surfaces and other systems with active components can be described by diffusion equations with reactions. We uncover and discuss a new route to chaotic dynamics in such systems based on their representation as a chain of coupled nonlinear oscillators. Hide Abstract
Speaker: Teaching electronically and via the Access Grid: Annotations and Digital Ink Bill Blyth  (Australian Mathematical Sciences Institute, ) D109, Friday 27 June, 2008, 3:30pm to 4:30pm Mathematics Show Abstract In recent years it has become common for mathematics lecturers to use computer projection in their lectures and seminar presentations. Often some handwriting on a whiteboard is used for asides: for clarification, for worked examples and sketches. If the handwriting is done electronically, as is necessary over the Access Grid, it is referred to as Digital Ink. We will give a general overview of e-teaching approaches. This will include using the beamer class in LaTeX to produce pdf slides (with stepped uncovering of a slide) and annotation of any pdf file using PDF Annotator. We will demonstrate using a TabletPC to produce highlighting (note that a laser pointer is not effective in an AGR) and annotations of pdf “notes” or slides. When using a TabletPC, WindowsXP provides a very good (and fully integrated) Digital Ink with Word, PowerPoint and Excel. Digital Ink within Maple and also with an interactive whiteboard will be demonstrated. We discuss an example of marking remote student work (as a pdf file and using PDF Annotator). Hide Abstract
Speaker: Averaging and deviation for slow-fast stochastic partial differential equations Wei Wang  (Student, Department of Mathematics and Computing, USQ, ) D109, Thursday 26 June, 2008, 12:00noon to 1:00pm Mathematics Stochastic differential equation.pdf 987KB (Speaker Slides) Wei.mp3 10752KB (Speaker Audio) Show Abstract Averaging principle is an important method to extract an effective macroscopic dynamics from complex system with slow modes and fast modes. Here an averaged equation for a class of stochastic partial differential equations are derived without a Lipschitz assumption on slow modes and the rate of convergence in probability is obtained as a byproduct. Moreover the deviation between the original equations and the averaged equations is studied. It is proved by a martingale approach that the deviation is described by an Gaussian process. This gives a much better approximation.Hide Abstract
Speaker: Energy analysis of thermo-magneto-convection in a vertical layer of ferro-magnetic fluid Sergey Suslov  (Mathematics and Computing, USQ) D109, Thursday 12 June, 2008, 12:00noon to 1:00pm Mathematics Sergey.mp3 11293KB (Speaker Audio) slide.pdf 1906KB (Speaker Slides) Show Abstract We study convection in a vertical layer of ferro-magnetic fluid heated from a side and subject to the transverse magnetic field. It is found that the fluid motion is caused by interacting thermo-gravitational and thermo-magnetic mechanisms. Our experiments and computations show that the excitation of magneto-convection leads to the formation of vertically aligned stationary rolls, while gravitational convection results in horizontal rolls corresponding to a pair of counter-propagating thermal waves. The interaction of these instability modes leads to a wide spectrum of experimentally observed flow patterns including stationary rolls and standing waves of various spatial orientations. A comprehensive stability map is computed and compared with experimental flow visualisations. Disturbance energy is analysed to achieve a deeper insight into physical mechanisms driving the fluid motion.Hide Abstract
Speaker: Increasing power of robust test through pre-testing Rossita Mohamad Yunus  (Student, Mathematics and Computing, USQ, ) D109, Thursday 29 May, 2008, 12:00noon to 1:00pm Statistics rossita.mp3 6751KB (Speaker Audio) seminar4.pdf 249KB (Speaker Slides) Show Abstract Robust procedures are developed for testing the intercept of a simple regression model when it is apriori suspected that the slope has a specified value. Defining unrestricted test (UT), restricted test (RT) and pretest test (PTT) corresponding to the unrestricted (UE), restricted (RE), and preliminary test estimators (PTE) in the estimation case, the M-tests formulated in the M-estimation methodology is used to derive their asymptotic power functions. Analytical and computational comparisons of the three tests are obtained by studying the relative efficiency of their power functions. It is shown that PTT achieves a reasonable dominance over the other tests asymptotically.Hide Abstract
Speaker: Protecting Information Sharing in Distributed Collaborative Environment Min Li  (PhD Student, Department of Mathematics and Computing, USQ, ) D109, Thursday 17 April, 2008, 12:00noon to 1:00pm Computing MinLi.mp3 2494KB (Speaker Audio) Protecting Information Sharing in Distributed Collaborative Environment.ppt 263KB (Speaker Slides) Show Abstract Information sharing on distributed collaboration usually occurs in broad, highly dynamic network-based environments, and formally accessing the resources in a secure manner poses a difficult and vital challenge. Our research is to develop a systematic methodology for information sharing in distributed collaborative environments. It will ensure sensitive information and information assurance requirements, and incorporate new security constrains and policies raised by emerging technologies. We will create a new rule-based framework to identify and address issues of sharing in collaborative environments; and to specify and enforce security rules to support identified issues while minimizing the risks of information sharing through the framework.Hide Abstract
Speaker: Bayesian Statistics for avoiding Over-fitting System Models Ron Addie  (Mathematics and Computing, USQ) D109, Thursday 3 April, 2008, 1:00pm to 2:00pm Computing Show Abstract Many methods for estimating linear models and nonlinear models of Systems and of time series are available, e.g. Box-Jenkins models, methods from signal processing, radial basis functions, and neural networks. It is essential to avoid over-fitting such models but methods for avoiding over-fitting may restrict model complexity or hamper efficient use of data, or both. In this talk the Bayesian approach to model estimation, as outlined in the book on Information Theory by David MacKay, for example, will be used to develop explicit algorithms for model estimation which are fast, do not overfit, and do not restrict model complexity. Algorithms, implemented in R, for estimation of nonlinear multivariate time series will be demonstrated.Hide Abstract
Speaker: On the complexity of restricted k-anonymity problem Xiaoxun Sun  (PhD Student, Department of Mathematics and Computing, USQ, ) D109, Thursday 27 March, 2008, 12:00noon to 1:00pm Statistics APWeb08.pdf 350KB (Speaker Slides) Xiaoxun.mp3 6276KB (Speaker Audio) Show Abstract One of the emerging concepts in microdata protection is k-anonymity, introduced by Samarati and Sweeney. k-anonymity provides a simple and efficient approach to protect private individual information and is gaining increasing popularity. k-anonymity requires that every tuple(record) in the microdata table released be indistinguishably related to no fewer than k respondents. In this paper, we introduce two new variants of the k-anonymity problem, namely, the Restricted k-anonymity problem and Restricted k-anonymity problem on attribute (where suppressing the entire attribute is allowed). We prove that both problems are NP-hard for k>3. The results imply the main results obtained by Meyerson and Williams. On the positive side, we develop a polynomial time algorithm for the Restricted 2-anonymity problem by giving a graphical representation of the microdata table. Hide Abstract
Speaker: Formalising CVS/SVN: A New Concurrency Control Protocol for Documents Stijn Dekeyser  (Mathematics and Computing, USQ) D109, Thursday 20 March, 2008, 12:00noon to 1:00pm Computing FormalisingCVS.ppt 102KB (Speaker Slides) stijn.mp3 14255KB (Speaker Audio) Show Abstract Although large software companies have added collaborative features to their office suits, CVS and SVN remain the prime technologies for enabling version management and concurrent authoring of documents. Compared to concurrency control in relational databases, none of the current solutions provide absolute document correctness. In previous work we have presented a locking protocol for general XML documents which guaranteed correctness but was too inefficient for normal text documents. In this seminar we preview a new protocol tailored specifically to documents. We show that our protocol in effect improves and formalizes the workings of CVS and SVN.Hide Abstract
Speaker: pi Day Jonathan Borwein   (Dalhousie University) D109, Friday 14 March, 2008, 3:00pm to 4:00pm Mathematics pi-day-08 sem.doc 59KB (Invitation from AMSI) Show Abstract The desire, and originally the need, to calculate ever more accurate values of Pi, the ratio of the circumference of a circle to its diameter, has challenged mathematicians for millennia and, especially recently, Pi has provided fascinating examples of computational mathematics. Pi, is also uniquely part of the popular imagination. I shall describe Pi's history from its beginnings to the present time. Bio: Professor Borwein, FRSC, is a distinguished Canadian mathematician with many accomplishments and awards, who is also a leading world authority on Pi. In 2008, he is on sabbatical leave at the University of Newcastle from his Canada Research Chair, Dalhousie University. Also note that Jonathan Borwein is a leader in Access Grid technology and its use and is well known as an outstandingly good presenter.Hide Abstract
Speakers: A Model, Schema, and Interface for Metadata File Systems Richard Watson, Stijn Dekeyser, Lasse Motrone, Expand Richard Watson  (Mathematics and Computing, USQ) Stijn Dekeyser  (Mathematics and Computing, USQ) Lasse Motrone  (Masters Student, Dept. Mathematics and Computing, USQ) Simplify D109, Thursday 13 March, 2008, 12:00noon to 1:00pm Computing richard.mp3 8155KB (Speaker Audio) Show Abstract Modern computer systems are based on the traditional hierarchical file system model, but typically contain large numbers of files with complex interrelationships. This traditional model is not capable of meeting the needs of current computer system users, who need to be able to store and retrieve files based on flexible criteria. A metadata file system can associate an extensive and rich set of data with a file, thus enabling more effective file organisation and retrieval than traditional file systems. In this seminar we consider current approaches to handling file metadata, and highlight deficiencies. We introduce a data model for a database-oriented pure MDFS complete with operations and semantics. The model supports user-initiated instance and schema updates and file searches based on structured queries. We also explore the design space of a set of user interface operations intended to implement the pure model and facilitate the capturing of rich metadata. We argue that without such a simple method for users to create rich metadata, progress in this field will remain limited.Hide Abstract
Speaker: Influence of Inhomogeneous and Anisotropic Tissue Conductivity on Human Head Modelling for EEG Forward Problem Md. Rezaul Bashar   (Student, Department of Mathematics and Computing, USQ, ) D109, Thursday 21 February, 2008, 11:00am to 12:00noon Computing Bashar.mp3 7212KB (Speaker Audio) dissertation proposal.ppt 425KB (Speaker Slides) Show Abstract Fundamental problem in electrophysiology is studied by computationally modelling the macroscopic and microscopic bioelectric fields. At a macroscopic level, all tissues are homogeneous and isotropic. However, all tissues are inhomogeneous and anisotropic in microscopic level. This study investigates the influence of inhomogeneous anisotropic tissue conductivity for forward problem on human head modelling. The forward problem in electroencephalography (EEG) aims to determine the current distribution caused by the sources in the brain using potential differences measured from the scalp. The solution requires a given source inside the brain and a volume conduction model of the head with different conductivities for each layer. This study presents a comprehensive literature review on head modelling, assigning tissue conductivities according to our proposed techniques, the methods for forward computation, and the error measurements to analyse EEG influenced by tissue inhomogeneous anisotropy.Hide Abstract
Speaker: Electronic Waste – What is it & where do I put it? Malcolm Wolski  (Associate Director, Research Computing Services, Information and Communication Technology Services, Griffith University , ) Q501/2, Wednesday 20 February, 2008, 5:30pm to 6:30pm Computing Show Abstract Question: What contains lead, cadmium, chromium, mercury, nickel, zinc, tantalum, indium, vanadium, terbium, beryllium, gold , europium, titanium, ruthenium, cobalt, palladium, manganese, silver, antimony, bismuth, selenium, niobium, yttrium, rhodium, platinum, mercury, arsenic, silica and fire retardant plastics? Answer: Your desktop PC & monitor Come and find out why governments and environmentalists are getting excited about electronic waste. Malcolm Wolski is Associate Director, Information and Computing Technology Services at Griffith University where he manages the University's Research Computing Services unit. Prior to taking up his current position, Malcolm was responsible for managing the University desktop computer fleet. Malcolm worked with Dell Computers Australia to hold the first e-waste Collection Day in Queensland which was held on campus at Griffith and also undertook a research project on behalf of all Australian University Directors of ICT to look for e-waste solutions for the University sector. In 2006, Malcolm worked with the Natural Edge Project (TNEP) and Dell Computers Australia, to publish learning materials which introduce and discuss the challenges of e-waste in our society. Malcolm has an active ongoing role in Griffith University's e-Waste Research Centre (www.griffith.edu.au/ewaste), is a member of the Information and Communication Technology Services Sustainability working party and is also on the University-wide working party on sustainability.Hide Abstract
Speaker: On dimensional quantities, problem solving style and notations in mathematical courses Dmitry Strunin  (Mathematics and Computing, USQ) D109, Thursday 14 February, 2008, 12:00noon to 1:00pm Mathematics dmitry.mp3 12417KB (Speaker Audio) Show Abstract The topics listed in the title are essential part of our math courses and important for engineering mathematics and other disciplines. From my viewpoint, the ay the topics are dealt with in the Calculus books are often not satisfactory. I wrote an appendix intended for the study book of MAT2100 (Algebra and Calculus-II) on the topics. This talk can be of interest for the math group and for students (CESRC research students are welcome).Hide Abstract
Speaker: Snapshot Simulation of Internet Traffic: Fast and Accurate for Heavy-tailed Flows Ron Addie  (Mathematics and Computing, USQ) D109, Thursday 7 February, 2008, 12:00noon to 1:00pm Computing ron.mp3 132KB (Speaker Audio) snapshot.pdf 382KB (Speaker Slides) Show Abstract Internet traffic includes a small number of very long slow-moving flows and a very large number of short fast-moving flows. In this presentation, a method for simulating both types of flows, and everything in-between, will be demonstrated. The simulation technique takes many “snapshots” of typical configurations of flows. Each snapshot has a very interesting structure which can readily be seen in the form of a picture, and the conclusions from the simulations are readily understood simply by viewing a sequence of these pictures.Hide Abstract
Speaker: Model dynamics on a multigrid across multiple length and time scales Tony Roberts  (Mathematics and Computing, USQ) D109, Thursday 31 January, 2008, 12:00noon to 1:00pm Mathematics talk.pdf 132KB (Speaker Slides) tony.mp3 12557KB (Speaker Audio) Show Abstract Methods for modelling dynamics posit just two time scales: a fast and a slow scale. But many applications, including many in continuum mechanics, possess a wide variety of space-time scales; often they possess a continuum of space-time scales. I discuss an approach to modelling the dynamics of advection and diffusion with rigorous support for changing the resolved space-time scale by just a factor of two. The mapping of dynamics from a finer grid to a coarser grid may be iterated to generate a hierarchy of models across a wide range of space-time scales, all with rigorous support for the connection from one model to another. This approach will empower us to have great flexibility in modelling complex dynamics over multiple scales.Hide Abstract
Speakers: An efficient hash-based algorithm for minimal k-anonymity Xiaoxun Sun, Hua Wang, Expand Xiaoxun Sun  (PhD Student, Department of Mathematics and Computing, USQ, ) Hua Wang  (Mathematics and Computing, USQ) Simplify D109, Tuesday 8 January, 2008, 12:00noon to 1:00pm Computing Show Abstract A number of organizations publish microdata for purposes such as public health and demographic research. Although attributes of micro-data that clearly identify individuals, such as name and medical care card number, are generally removed, these databases can sometimes be joined with other public databases on attributes such as Zip code, Gender, and Age to re-identify individuals who were supposed to remain anonymous. "Linking" attacks are made easier by the availability of other, complementary, databases over the Internet. k-anonymity is a technique that prevents "linking" attacks by generalizing and/or suppressing portions of the released microdata so that no individual can be uniquely distinguished from a group of size k. In this paper, we investigate a practical model of k-anonymity, called full-domain generalization. We examine the issue of computing minimal k-anonymous table based on the definition of minimality described by Samarati. We introduce the hash-based technique previously used in mining associate rules and present an efficient hash-based algorithm to and the minimal k-anonymous table, which improves the previous binary search algorithm first proposed by Samarati.Hide Abstract
Speaker: Analysis of buffers for long-range dependent traffic: A case where asymptotic approximations fail Ron Addie  (Mathematics and Computing, USQ) D109, Thursday 22 November, 2007, 12:00noon to 1:00pm Computing Show Abstract Asymptotic approximations have been used to estimate losses in routers as buffers become larger or traffic becomes heavier. In this talk, I will describe and make use of an approximation which is accurate for a wide range of system parameters, demonstrate the consistency of this approximation with all the existing asymptotic results, and show that the asymptotic results apply only to very remote regions of the parameter space. I will also demonstrate that some asymptotic approximations are unstable, varying dramatically as the system parameters change by small amounts. It is important to understand and take into account the inaccuracy of these asymptotic results in order to maintain the important contribution of mathematical modeling to understanding and optimal control of systems.Hide Abstract
Speakers: Power of the test after pretest Md. Enamul Kabir, Shahjahan Khan, Expand Md. Enamul Kabir  (Student, Mathematics and Computing, USQ, ) Shahjahan Khan  (Mathematics and Computing, USQ) Simplify D109, Thursday 15 November, 2007, 12:00noon to 1:00pm Statistics Show Abstract The problem of testing the intercept parameter in a simple regression model following a preliminary test (PT) on the slope parameter is considered in this paper. Unrestricted test (UT), restricted test (RT) and pre-test test (PTT) for the intercept parameter are defined when the slope parameter is unknown, known and following a PT on it. Power functions of these tests are derived and compared. The impact of PT on the size and power of the PTT is also studied. The study reveals that the PTT attains a reasonable dominance over UT and RT.Hide Abstract
Speaker: A romp through retention rates Paul Fahey  (Mathematics and Computing, USQ) D109, Thursday 1 November, 2007, 12:00noon to 1:00pm Statistics PaulFahey.mp3 14287KB (Speaker Audio) seminar.pdf 1497KB (Speaker Slides) Show Abstract Recent lack of success on the Commonwealth Government’s Learning and Teaching Performance Funding (LTPF) creates the impression that USQ staff are unsuccessful at teaching and/or USQ students are unsuccessful at learning. Two of the seven criteria used in the LTPF are student progression and student retention. USQ typically rates in the lowest half dozen universities on these two measures. The challenge is to work out why. To start to investigate student progression and retention I have taken the cohort of students enrolling at USQ in Semester 1 1997 and followed their progress using 10 years of administrative data. Analyses focuses on the 20 programs with highest intake in Semester 1 1997 consisting of 13 bachelor programs (BART, BBUS, BCOM, BEDS, BEEC, BENG, BEPR, BETC, BINT, BITE, BNPO, BNUR, BSCI), 3 masters programs (MBA, MED1, MP12), and 4 non-degree programs (ACEQ, PREP, CRUG, PDEV). Outcomes of interest are a) proportion who did not return after their first semester b) proportion who did not return after their first year and c) proportion who actually graduated. A range of student characteristics were reviewed including gender, age, residency status (Australian, international onshore, international offshore, etc), country of residence, basis of admission, number of subjects attempted in their first semester, etc. The analysis aims to quantify (and check the statistical significance of) variation in outcomes between student groups. All results are presented graphically. The discussion focuses on what diagnostic information could be obtained from administrative data to assist the University understand its retention and progression statistics, how these results could be presented, who these results could be presented to and what impact, if any, this might have on USQ performance and/or USQ performance statistics. This presentation is one part of a project sponsored by the Dean of Students, funded through USQ Strategic Development Fund 2006 and led by Paul Fahey, Janet Taylor, Shane Klease, Ashley Plank and Jiuyong Li. Hide Abstract
Speakers: Increasing power of the test through pre-test - a robust method Rossita Mohamad Yunus, Shahjahan Khan, Expand Rossita Mohamad Yunus  (Student, Mathematics and Computing, USQ, ) Shahjahan Khan  (Mathematics and Computing, USQ) Simplify D109, Thursday 25 October, 2007, 12:00noon to 1:00pm Statistics rossita.mp3 10103KB (Speaker Audio) seminar1.pdf 191KB (Speaker Slides) Show Abstract This paper develops robust test procedures for testing the intercept of a simple regression model when it is apriori suspected that the slope has a specified value. Defining unrestricted test (UT), restricted test (RT) and pre-test test (PTT) corresponding to the unrestricted (UE), restricted (RE), and preliminary test estimators (PTE) in the estimation case, the M-estimation methodology is used to formulate the M-tests and derive their asymptotic power functions. Analytical and graphical comparisons of the three tests are obtained by studying the power functions with respect to size and power of the tests. It is shown that PTT achieves a reasonable dominance over the others asymptotically.Hide Abstract
Speaker: Capacity Analysis of MIMO-OFDM Broadband Channels In Populated Indoor Environments (Mathematics and Computing, USQ) D109, Thursday 11 October, 2007, 12:00noon to 1:00pm Computing ISCIT_ver03_WON.ppt 812KB (Speaker Slides) jishu.mp3 5507KB (Speaker Audio) Show Abstract The results of dynamic channel capacity measurements for a multiple-input multiple-output orthogonal frequency division multiplexing (MIMO-OFDM) system in two populated indoor environments with and without line-of-sight (LoS) will be reported. The experiment used 4 sending and 4 receiving antennas and 114 sub-carriers at 5 GHz as per draft IEEE 802.11n. MIMO-OFDM channel capacity is analysed both with fixed receiver signal-to-noise ratio (SNR) and fixed transmitter (Tx) power criteria. It is found that fixed SNR capacity increased while fixed Tx power capacity decreased in both environments by the presence of pedestrian. It is also revealed that the spread of the capacity cumulative distribution function (CDF) increased due to the pedestrians in both environments with both criteria.Hide Abstract
Speaker: How to Give a Proposal Seminar Michael de Raadt  (Mathematics and Computing, USQ) D109, Wednesday 3 October, 2007, 10:00am to 11:00am Computing howToGiveAProposalSeminar.zip 6079KB (Seminar Slides and Audio Zipped) http://www.sci.usq.edu.au/research/seminars/files/seminar173/breeze/index.html (Seminar Slides and Audio Streaming) Show Abstract This is an informative lecture describing how to give a research proposal seminar and good presentation practices.Hide Abstract
Speaker: Everything you ever wanted to know about Identity Theft but didn't dare ask Nick Tate  (Chair, ACS Queensland, ) Q501/2, Wednesday 26 September, 2007, 6:00pm to 7:00pm Computing Nick Tate - BIO for ACS - 2007.pdf 39KB (Nick Tate Bio) Show Abstract How secure is the Internet? This talk will attempt to answer this question by looking at some of the security issues which confront internet users and reviewing some of the types of attack that occur, before offering some suggestions for the future. Hide Abstract
Speakers: Comparing Writely and Moodle Online Assignment Submission and Assessment Khaleel Petrus, Michael Sankey, Expand Khaleel Petrus  (Mathematics and Computing, USQ) Michael Sankey  (Learning and Teaching Support Unit, USQ, ) Simplify D109, Thursday 20 September, 2007, 12:00noon to 1:00pm Computing khaleel.mp3 9175KB (Speaker Audio) seminar_writely_Moodle.ppt 543KB (Speaker Slides) Show Abstract This paper discusses the diverse and changing teaching environments in the higher education sector. With the extensive use of Internet and the availability of numerous online interaction tools, students are increasingly participating in, and requesting access to, features such as electronic course materials, online forums, automated response facilities and mechanisms for electronic assignment submission. As USQ moves to provide more of these flexible online environments, particularly to assist in the assessment of student learning outcomes, it is equally important for lecturing staff to select the most appropriate mechanisms and assignment submission tools to assist their students in this process. This paper reports on a pilot study performed to investigate student perceptions of using two online collaborative software systems for submitting assignment work and receiving feedback; Writely and Moodle. An online survey was presented to the students to quantify their preferences in using these two online facilities with respect to intuitiveness, convenience of use and responsiveness. Comments were also sought from students to: Explain why they made their particular choices and report on any problems they encountered. The survey also offered them the opportunity to provide further feedback. The quantitative data provides a clear indication that the students prefer using the Moodle system over Writely while the qualitative data give a clear indication as to why this was the case.Hide Abstract
Speaker: Delegating revocations and authorizations Hua Wang  (Mathematics and Computing, USQ) D109, Thursday 20 September, 2007, 11:00am to 12:00noon Computing hua.mp3 10399KB (Speaker Audio) presentationCBP07new.pdf 153KB (Speaker Slides) Show Abstract Delegation models based on role-based access control management have been known as flexible and efficient access management for data sharing on distributed environment. Delegation revocations are a significant functionality for the models in distributed environment when the delegated roles or permissions are required to get back. However, problems may arise in the revocation process when one user delegates user $U$ a role and another user delegates a negative authorization of the role. This paper aims to analyse various role-based delegation revocation features through examples. Revocations are categorized in four dimensions: Dependency, Resilience, Propagation and Dominance. According the dimensions, sixteen types of revocations exist for specific requests in access management. We present revocation delegating models, and then discuss user delegation authorization and the impact of revocation operations. Finally, comparisons with other related work are indicated.Hide Abstract
Speakers: Estimation of univariate normal mean using p-value Md. Enamul Kabir, Shahjahan Khan, Expand Md. Enamul Kabir  (Student, Mathematics and Computing, USQ, ) Shahjahan Khan  (Mathematics and Computing, USQ) Simplify D109, Thursday 13 September, 2007, 12:00noon to 1:00pm Statistics Enamul.mp3 9730KB (Speaker Audio) mean_p, seminar.pdf 162KB (Speaker Slides) Show Abstract This paper estimates the mean of a normal distribution in presence of non-sample prior information regarding the value of the mean. We define the unrestricted estimator (UE), restricted estimator (RE), p-value based restricted estimator (PRE) and p-value based preliminary test estimator (PPTE) based on the sample, prior information, p-value of an appropriate test and combination of all them respectively. The relative performance of these estimators is studied on the basis of bias and mean square error (MSE) criteria. Both analytical and graphical comparisons are investigated. If the p-value is reasonably small, the PRE almost uniformly over performs the UE and PPTE, regardless of the level of significance.Hide Abstract
Speaker: Capturing screen video - with a free tool Birgit Loch  (Mathematics and Computing, USQ) D109, Thursday 6 September, 2007, 12:00noon to 1:00pm Other Show Abstract This informal presentation will show how to get access to (and get started with) an easy to use, free tool, for recording any movement on your screen. The differences between this tool, and licensed software such as Camtasia and Captivate will be pointed out. Screen capture can be used for instance to prepare training videos for software packages (for staff and students), or to record writing on the computer with a stylus (e.g. development of mathematical formulae). Recordings can be done ad-hoc, to answer a distance student enquiry, or put together after preparation.Hide Abstract
Speakers: Second ElluminateLive Session Shirley Reushle, Birgit Loch, Peter Evans, Expand Shirley Reushle  (Education, USQ, ) Birgit Loch  (Mathematics and Computing, USQ) Peter Evans  (USQ, ) Simplify D303, Wednesday 8 August, 2007, 12:00noon to 2:00pm Computing Elluminate2nd session.pdf 49KB (Flyer) Show Abstract ElluminateLive is a virtual classroom and web conferencing tool built specifically for live (real time), multimedia collaboration. Sessions can be recorded and made available for play back. EL can be used for teaching (DISTANCE ON CAMPUS, DISTANCE), research, presentations, meetings…Hide Abstract
Speaker: Ethics in Computing Michael Thompson  (Division of Information and Communication Technology Services, USQ, ) D303, Thursday 2 August, 2007, 11:00am to 12:00noon Computing Ethics in Computing.doc 126KB (Speaker Notes) mikeThompson.mp3 11245KB (Speaker Audio) Show Abstract Ethics can cover a wide range of issues associated with: Commerce, Computer Abuse, Social Justice Issues, Speech Issues, Intellectual Property, Privacy and Risks. The seminar will provide examples of how organisations such as USQ and the Australian Computer Society (ACS) provide advice and information to our constituents about issues associated with ethics through Codes of Practice and Professional Conduct. Case study examples from the ACS will illustrate the various scenarios that computing graduates will be faced with when they join force and will illustrate how legislation and codes of practice can assist.Hide Abstract
Speaker: A world tour of dynamical systems, stability,and chaos Rowena Ball  (Department of Theoretical Physics and the Mathematical Sciences Institute, ANU, ) T120, Tuesday 31 July, 2007, 11:00am to 12:00noon Mathematics SelfRegulator.avi 32596KB (Video shown during presentation) usq-seminar.pdf 3193KB (Speaker Slides) Show Abstract This journey begins at a large nineteenth century house in northern England. The route then takes us to a museum in Regensburg, thence to an office tower in Hong Kong. We end up at an oil refinery, its towers defining the skyline of a seamy port city in complicated chiaroscuro. During the tour I review aspects of the mathematics of dynamical systems, stability, and chaos within a historical framework that draws together the two major threads of its early development: celestial mechanics and control theory, and focusing on qualitative theory. From this perspective I show how concepts of stability enable us to classify dynamical equations and their solutions and connect the key issues of nonlinearity, bifurcation, control, and uncertainty that are common to time-dependent problems in natural and engineered systems. We also learn about one of the key players in the story: the 1958 Chrysler Imperial. Some work-in-progress on stability of complex networks may be discussed.Hide Abstract
Speaker: Allocation of individual marks in student teams: a test of teamwork? Paul Fahey  (Mathematics and Computing, USQ) T119, Thursday 26 July, 2007, 12:00noon to 1:00pm Statistics Dep07.ppt 538KB (Speaker Slides) paul.mp3 12431KB (Speaker Audio) Show Abstract In 2002 the USQ Faculty of Engineering and Surveying introduced four mandatory, problem-based subjects into its curriculum. In Problem Solving 2 students work in teams of seven to produce reports on three problems (one report every four weeks). Each team submits a single report for marking and identification of any individual contributions to the report is actively discouraged. With the 3 reports worth 90% of the total assessment, students need their team to work well in order to succeed. Team marks are produced by normal academic marking of the team reports. (The content is largely statistics and physics.) The transition from team mark to individual mark is informed by team consensus of the relative contribution of each team member. In this paper, I will quickly overview the design and management of Problem Solving 2 and then move on to an evaluation of the assessment system. This evaluation will review as best as possible: Validity (mainly comparing individual student’s marks against their GPA from non team-based units). Reliability (correspondence between team consensus of contribution scores and confidential ratings from each individual team member). Acceptability (take-up rates of the assessment options and qualitative feedback) These evaluations have their limitations but all results are consistent with a fair and acceptable assessment scheme for team-based learning. Speaker Bio Paul Fahey was a founding member of the teaching team for ENG2102 and taught it for four years. Through this involvement in Paul was a member of the team which received the USQ “Design and Delivery of Teaching Materials” Award 2005 and member of the team which received the Australasian Association for Engineering Education’s “Excellence in Curriculum Team Project in Engineering Education” Award 2005.Hide Abstract
Speakers: Innovative Teaching - Socrates' New Tablet Sarah Bauer, Mason Wong, Ray Lai Wai Khuen, Expand Sarah Bauer  (NetOptions, ) Mason Wong  (Toshiba) Ray Lai Wai Khuen  (Heulabs Software) Simplify D303, Wednesday 18 July, 2007, 12:30pm to 2:00pm Computing Show Abstract Find out about exciting new software to make your teaching more dynamic and engaging. Come along to the demonstration of educational software using a Tablet PC.Hide Abstract
Speakers: Improving the Algebra Knowledge and Skills of Prospective Secondary Teachers Nerida Ellerton, Ken Clements, Expand Nerida Ellerton  (Department of Mathematics, Illinois State University, ) Ken Clements  (Department of Mathematics, Illinois State University, ) Simplify D303, Monday 16 July, 2007, 12:00noon to 1:00pm Mathematics Show Abstract In this paper we analyse recent data from the United States showing that 50 “top” students who had completed full algebra programs in secondary school tended to know very little of the algebra they had been taught, in the key algebra content areas of linear and quadratic equations, and linear inequalities. Data from a successful planned intervention, aimed at upgrading the algebra subject knowledge and associated pedagogical content knowledge of prospective primary and middle-secondary teachers, will be presented and briefly analysed in the paper. The paper will close with discussion of the question of what needs to be done in schools and mathematics teachers education programs in the United States of America, and in similar programs elsewhere (e.g., Toowoomba, Australia), in order that students will develop relational understanding of important algebra concepts. Hide Abstract
Speakers: Elluminate Live v8 Trial Shirley Reushle, Birgit Loch, Expand Shirley Reushle  (Education, USQ, ) Birgit Loch  (Mathematics and Computing, USQ) Simplify D303, Monday 16 July, 2007, 11:00am to 12:00noon Computing ELflyer.pdf 104KB (Eluminate Flyer) Show Abstract USQ has a history of award winning for innovation in distance education. To regain and retain a leading edge in education, it is vital that we spend time exploring and using emerging technologies to support pedagogical innovation and the Vice Chancellor’s flexibility agenda. To achieve this, USQ academics need access to these technologies. Elluminate Live is a virtual classroom, web conferencing tool built specifically for live (real time), multimedia collaboration. EL can be used for teaching, research, presentations, meetings and includes features such as session recording, typed chat, audio (multiple speakers) and video, whiteboard, application and desktop sharing, graphing calculator, moderator/student mode, quizzes, cross platform, and broadband or dial-up access. The trial is being conducted July – December 2007. This purpose of the infomation session is to discuss and demonstrate some of the features of Elluminate Live, and outline the trial process including the expectations of those involved in the trial, set up information, details of the Moodle site attached to the trial, and the evaluation process. To see recorded demonstration, please contact Birgit.Hide Abstract
Speaker: Procedural Generation Terrains for Games using Fractals and L-Systems Kae Thong Chew (Max)  (Student, Mathematics and Computing, USQ, ) W534, Thursday 12 July, 2007, 12:00noon to 1:00pm Computing max.mp3 6832KB (Speaker Audio) Research Dissertation Slide ver2.ppt 1948KB (Speaker Slides) Show Abstract As the worlds within computer games becoming more complex with each generation of hardware and software, it is becoming increasingly burdensome for artists and developers to create these multiple detailed worlds with visual and structural realism. With each new generation of hardware, developers can create even more fantastic universes in which players can immerse themselves. This project investigates the development of procedural generation terrains for games by using fractals and L-Systems. The aim is to design and investigate the use of L-System fractals to generate a game terrain map suitable for implementation in a computer game. The proposed project will try to produce new methods and experimental results on how L-Systems can be used further to generate game map terrains for new research areas. One of the new research areas includes introducing L-Systems to generate interior game maps or interior building architecture models. The created models solutions prove to be a reference material and provide a basis for future researches and developments in generating game map terrains procedurally.Hide Abstract
Speaker: Reducing the Effects of Multiplicity in Microarrays Anita Pelecanos  (Student, Mathematics & Computing, USQ, ) D109, Friday 6 July, 2007, 1:00pm to 2:00pm Bioinformatics anita.mp3 12947KB (Speaker Audio) Combatting the Effects of Multiplicity in Microarrays.ppt 802KB (Speaker Slides) Show Abstract cDNA microarrays measure thousands of gene expression levels (RNA abundance in cells) at the same time to identify genes that are differentially expressed. Searching for differentially expressed genes in microarrays requires testing thousands of hypothesis tests simultaneously. The validity of the statistical outcomes is compromised due to multiple testing. In multiple testing, it is difficult to control the overall number of false positives (Type I errors) while still maintaining power. Issues arise from determining an appropriate Type I error rate to select, along with a method to control this error rate for many joint hypothesis tests. Methods to achieve Type I error control have progressed from controlling the familywise error rate to controlling the false discovery rate. Some approaches to control the familywise error rate and the false discovery rate for microarray studies are discussed. Application of such procedures to a simulated data set and microarray data set are demonstrated to quantify the number of rejections made and how many of them are correctly rejected. An indication of the false positive rate in microarrays is illustrated in a case study. Hide Abstract
Speakers: Copying in Mathematics and Computing , Michael de Raadt, Expand   (Mathematics and Computing, USQ) Michael de Raadt  (Mathematics and Computing, USQ) Simplify D109, Thursday 28 June, 2007, 12:00noon to 1:00pm Other Copying.ppt 354KB (Speaker Slides) Show Abstract This Departmental discussion will focus on: Pressures on students Why students copy Informing students Better assignments Detection methods Hide Abstract Show References Sheard, J., Carbone, A., and Dick, M. 2003. Determination of factors which impact on IT students' propensity to cheat. In Proceedings of the Fifth Australasian Conference on Computing Education - Volume 20 (Adelaide, Australia). T. Greening and R. Lister, Eds. Conferences in Research and Practice in Information Technology Series, vol. 140. Australian Computer Society, Darlinghurst, Australia, 119-126. Sheard, J., Carbone, A., and Dick, M. 2003. Determination of factors which impact on IT students' propensity to cheat. In Proceedings of the Fifth Australasian Conference on Computing Education - Volume 20 (Adelaide, Australia). T. Greening and R. Lister, Eds. Conferences in Research and Practice in Information Technology Series, vol. 140. Australian Computer Society, Darlinghurst, Australia, 119-126.Hide References
Speaker: XML and Web Service Security Survey Li-li Sun  (Student, Mathematics and Computing, USQ, ) D109, Thursday 14 June, 2007, 12:00noon to 1:00pm Computing lili.mp3 6936KB (Speaker Audio) XML security survey(final).ppt 151KB (Speaker Slides) Show Abstract Meeting security requirements for privacy, confidentiality and integrity is essential in order to move business online. Extensible Markup Language (XML) promoted by the World Wide Web Consortium (W3C) is rapidly emerging as the new standard language for semi-structured data representation and exchange on the Internet. XML documents may contain private information that cannot be shared by all user communities. Securing XML data, therefore, is becoming increasingly important. In 2002, several specifications progressed toward providing a comprehensive standards framework for securing XML-based applications have been presented. These applications can effectively to protect information contained in a website. In this talk, we will present XML and web service security main standards and most specifications for those standards. We will discuss each standard which connects with protecting XML based documents. We will also briefly describe the connections of those standards based on existing technologies.Hide Abstract
Speakers: Mathematical skills of students entering Engineering and Science studies at USQ Nic Jourdan, , Expand Nic Jourdan  (Mathematics and Computing, USQ, )   (Mathematics and Computing, USQ) Simplify D109, Thursday 14 June, 2007, 11:00am to 12:00noon Mathematics Nic.mp3 12078KB (Speaker 1 Audio - Nic Jourdan) pat.mp3 4533KB (Speaker 2 Audio - Pat Cretchley) Seminar June 07.ppt 124KB (Speaker Slides) Show Abstract We report on the mathematics competencies of 206 Engineering and Science students commencing an algebra and calculus course at USQ in the first semester of 2006. To inform course design and student support in the face of growing student diversity, skills were assessed in six fundamental areas, and compared with those measured in 2001. The findings revealed reasonable skills with arithmetic, fractions, and index laws but serious ongoing weaknesses in areas of algebra, functions, and trigonometry. These findings have important implications for program development at USQ. This work forms part of a USQ Teaching & Learning Research Grant project that includes the development of online diagnostic procedures, and the counseling of students into suitable studies.Hide Abstract
Speakers: SQLify Stijn Dekeyser, Michael de Raadt, Expand Stijn Dekeyser  (Mathematics and Computing, USQ) Michael de Raadt  (Mathematics and Computing, USQ) Simplify Outside USQ, Friday 1 June, 2007, 12:00noon to 1:00pm Computing Conventicle-SQLify.ppt 274KB (Speaker Slides) Stijn.mp3 7315KB (Speaker Audio) http://www.it.uu.se/research/group/upcerg/Publications/proceedingsKoliCalling2006/system3.pdf (Paper in proceedings of Koli Calling 2006) Show Abstract (Presented at Queensland CSEd Conventicle) In recent years a small number of web-based tools have been proposed to help students learn to write SQL query statements and also to assess students' SQL writing skills. SQLify is a new SQL teaching and assessment tool that extends the current state-of-the-art by in-corporating peer review and enhanced automatic assessment based on database theory to produce more comprehensive feedback to students. SQLify (pronounced as squalify) is intended to yield a richer learning experience for students and reduce marking load for instructors. In this paper SQLify is compared with existing tools and important new features are demonstrated.Hide Abstract
Speakers: Community in Large Classes Michael Bulmer, , Expand Michael Bulmer  (Department of Mathematics, University of Queensland, )   (Mathematics and Computing, USQ) Simplify D109, Thursday 24 May, 2007, 12:00noon to 1:00pm Mathematics MathsEd_May_Affect_Bulmer_07.ppt 165KB (Patricia's Introduction) USQ2007a.ppt 908KB (Michael's Speaker Slides) Show Abstract The notion of "community" in mathematics classes can have a range of meanings, such as a community of learners or a community of practice. This talk will give an overview of these notions of community in large classes and some approaches and experiences in fostering their growth . We will also discuss current research for improving our understanding of affective aspects of student learning in mathematics and statistics. Hide Abstract
Speakers: Protecting Small Flows from Large Ones for Quality of Service in the Internet Ron Addie, Oleksiy Yevdokimov, Stephen Braithwaite, David Millsom, Expand Ron Addie  (Mathematics and Computing, USQ) Oleksiy Yevdokimov  (Mathematics and Computing, USQ) Stephen Braithwaite  (PhD Student, Mathematics and Computing, USQ, ) David Millsom  (Cisco Systems, California, ) Simplify D109, Thursday 17 May, 2007, 12:00noon to 1:00pm Computing ron.mp3 11347KB (Speaker Audio) Show Abstract This paper is a significant step contributing to scalable management of Quality of Service for Internet flows by discriminating against large flows. The idea of discriminating against flows on the basis of size has certain optimal properties which mean that it may be much better than discriminating for flows on the basis of user-assigned classifiers. Notably, it is not necessary to classify flows at the network edge, all flows are advantaged, policing is unnecessary, and global changes to the Internet architecture are not necessary, since each router can independently identify large flows and put into effect the proposed discrimination. In this paper, as well as discussing the whole idea, we define a method for measuring flow size which is well adapted to this application, and we estimate key performance parameters of this approach, such as: how many large flows need to be managed, how quickly they will arrive and depart, what performance will be experienced by small flows, and what performance will be experienced by large flows. This paper will be presented at the International Conference on Digital Telecommunications, in Silicon Valley, California, in July 2007.Hide Abstract
Speakers: Workshop/Seminars on MIMO Technology Karla Ziri Castro, John Leis, Ron Addie, Hajime Suzuki, Expand Karla Ziri Castro  (Electrical, Electronic & Computer Engineering, USQ, ) John Leis  (Electrical, Electronic & Computer Engineering, USQ, ) Ron Addie  (Mathematics and Computing, USQ) Hajime Suzuki  (Wireless Technologies Laboratory, CSIRO ICT Centre, Sydney, ) Simplify T120, Friday 11 May, 2007, 10:00am to 4:00pm Computing Workshop Program.pdf 24KB (Program of Workshop Events and Breaks) Show Abstract In recent years, there has been an increased interest in the study of multiple-input multiple-output (MIMO) systems in multipath environments as an approach that can offer significant bandwidth efficiency in broadband wireless applications. The MIMO approach can yield significant gains for both link and network capacities, with no additional energy or bandwidth consumption when compared to conventional single-array diversity methods. To obtain an in depth idea about the MIMO technology we are organising a seminar/workshop. In the workshop we have distinguish speakers who will deliver their knowledge and expertise on the field of MIMO technology. This workshop is sponsored by the Computational Engineering and Science Research Centre (CESRC)Hide Abstract
Speaker: Investigating Artificial Intelligence Performance in Turn Based Strategy Game using Genetic Algorithms Jiwan Kandangwa  (Student, Mathematics & Computing, USQ) D109, Thursday 10 May, 2007, 12:00noon to 1:00pm Computing jiwang.mp3 9550KB (Speaker Audio) Seminar.ppt 441KB (Speaker Slides) Show Abstract Conventional turn based strategy games have artificial intelligence (AI) that is hard-coded or based on predefined scripts. As a player gains experience during game play they begin to predict the AIs strategy. This makes it easy to defeat on a regularly basis. Players can also develop new strategies over time whereas the AI strategies are static. This makes the game less enjoyable and the human player loses interest. This dissertation will present an investigation into the ways to improve the game-play by dynamically adjusting the AI’s strategy using Genetic Algorithms (GA) for counteracting the player’s strategy allowing the AI to observe the player and learn during game play. For experimental clarity a simple turn-based strategy game called Advanced Protection is redesigned with and without GA.Hide Abstract
Speaker: Coordinate transforms underpin multiscale modelling and reduction in deterministic and stochastic systems Tony Roberts  (Mathematics and Computing, USQ) K305, Thursday 26 April, 2007, 12:00noon to 1:00pm Mathematics tony.mp3 13148KB (Speaker Audio) Show Abstract A persistent feature of complex systems in engineering and science is the emergence of macroscopic, coarse grained, coherent behaviour from the interactions of microscopic agents (molecules, cells) and with their environment. In current modeling, ranging from ecology to materials science, the underlying microscopic mechanisms are known, but the closures to translate microscale knowledge to a system level macroscopic description are rarely available in closed form. Kevrekidis proposes new equation free' computational methodologies to circumvent this stumbling block in multiscale modelling. Nonlinear coordinate transforms underpin analytic techniques that support these computational methodologies. But to do so we must cross multiple space and time scales, in both deterministic and stochastic systems, and where the microstructure is either smooth or detailed. Using examples, I describe progress in using nonlinear coordinate transforms to illuminate such multiscale modelling issues.Hide Abstract
Speaker: Noise-free Stochastic Resonance Thomas Stemler   (University of Western Australia, ) K305, Tuesday 24 April, 2007, 12:00noon to 1:00pm Mathematics Thomas.mp3 11844KB (Speaker Audio) thomas.pdf 1943KB (Speaker Slides) Show Abstract A classical technique of theoretical physics and applied mathematics is to model dynamical systems with many degrees of freedom by a suitable stochastic model. For example, thermodynamics successfully models the many degrees of freedom by Gaussian white noise. It has been discovered recently that stochastic differential equation models also provide useful descriptions of certain chaotic systems with relatively few degrees of freedom. This talk will describe some recent successes of these techniques with particular references to experiments on a low-dimensional chaotic system implemented as an electronic circuit. The stochastic models give new insights into non-linear resonance phenomena. One well known counter-intuitive example is that non-linear systems can resonance when forced with a periodic signal and noise, so called stochastic resonance. Recently it has been discovered that certain chaotic systems (displaying intermittency) also show this kind of resonance. This is called noise-free stochastic resonance. "Noise-free" emphasises the fact that the role of noise is taken over by the fast deterministic degrees of freedom while the close connection to the resonance in stochastic systems is pointed out by "stochastic resonance". We will show that experimental results obtained from the deterministic electronic circuit are in good agreement with the prediction of the stochastic theory.Hide Abstract
Speaker: Phenol Formaldehyde (PF) resins Paul Faigl  (USQ Fibre Composite Design and Development Group, ) D303, Thursday 1 March, 2007, 11:00am to 12:00noon Mathematics Show Abstract Phenol Formaldehyde (PF) resins are known for their excellent heat, fire and chemical resistance and low cost. Their main disadvantage is a lack of flexibility and brittleness. In our Fibre Composite Design and Development (FCDD) group at the USQ we work extensively with PF resins and have achieved flexibility already somewhere between poly- and vinyl esters. The Holy Grail of this practically oriented research is to achieve extendibility around 5-7% with flexural stress >100 MPa (100.106 N.m-2 ~1000 atm ~1.45.104 psi). This will bring us to the region of epoxy resins. P + F react in a complex condensation reaction, where xP + yF generate polymeric product + water. The reaction proceeds initially as addition but because of branching at each step, many reaction outcomes are possible. As in the evolution process, only one outcome is finally realized. A good model should be able to predict a likely outcome, also and particularly when the reactants are offered other partners apart from P and F. In the discussion, we will introduce some basic chemical concepts of the PF condensation. We invite you to help us to gain more insight into the reaction landscape by means of suitable mathematical tools. These tools remain to be developed. It is a challenging task with immediate practical applications. Hide Abstract
Speakers: Redeveloping Networking Courses in Collaboration with Cisco, the IT Industry Giant Yan Li, David Lai, Peter Thomas, Expand Yan Li  (Mathematics and Computing, USQ) David Lai  (Mathematics and Computing, USQ) Peter Thomas  (Cisco, ) Simplify D303, Wednesday 28 February, 2007, 2:00pm to 3:00pm Computing Network Cisco Seminar.ppt 534KB (Speaker Slides (Yan and David)) Speaker 1 - Ron Addie (Welcome).mp3 921KB (Speaker Audio - Speaker 1) Speaker 2 - Yan Li.mp3 3929KB (Speaker Audio - Speaker 2) Speaker 3 - David Lai.mp3 1068KB (Speaker Audio - Speaker 3) Speaker 4 - Peter Thomas.mp3 8101KB (Speaker Audio - Speaker 4) Show Abstract USQ Mathematics and Computing department has teamed up with Cisco to provide networking courses that combine theoretical knowledge with hands on experience and qualifications from IT industry. Students can put their knowledge and skill into practice in our newly equipped laboratory with cutting edge Cisco networking hardware. This will prepare our IT graduates to become competent networking professionals. In this seminar, we will talk about how we incorporate courses from Cisco Network Academy Program into the Networking major of the Bachelor of IT (BIT). We will use one of our networking courses, CSC3407 Computer Communication and Internetworking, as a showcase. Cisco System Engineer Mr Peter Thomas from Cisco head office in Brisbane will introduce Cisco to us and explain how the Cisco Certified Network Associate (CCNA) and Cisco Certified Network Professional (CCNP) programs can enhance the job opportunities of USQ BIT graduates. Hide Abstract
Speakers: Experiences with producing transmodal study materials for computing courses Ian Richards, Birgit Loch, Expand Ian Richards  (Mathematics and Computing, USQ) Birgit Loch  (Mathematics and Computing, USQ) Simplify D303, Tuesday 30 January, 2007, 2:00pm to 3:00pm Computing TransmodalSeminar-3.ppt 1382KB (Speaker Slides) Show Abstract As USQ constantly strives to improve its teaching methods, a particular focus is the enhancement of study materials to take account of different student learning modalities. While on-campus students in computing courses gain access to face-to-face interaction including voice, whiteboard explanations and supervised exercises, external students have mostly been limited to printed material with some multimedia components delivered via the web. On campus students are also seeking access to more flexible forms of delivery, due to work commitments and limited contact hours with staff. In this seminar the presenters will describe their experiences in dealing with a number of authoring systems to produce transmodal study materials for computing courses, including ICE, Breeze, Captivate, Camtasia and Audacity. Example content will be demonstrated and student feedback reviewed. Some recommendations for the ongoing development of course materials in the Department of Mathematics and Computing will be presented. An opportunity will be provided to discuss how transmodal delivery can be incorporated in other computing courses.Hide Abstract
Speakers: Flows in networks - understanding how to manage them even when their definition is vague Ron Addie, Oleksiy Yevdokimov, Expand Ron Addie  (Mathematics and Computing, USQ) Oleksiy Yevdokimov  (Mathematics and Computing, USQ) Simplify D303, Thursday 25 January, 2007, 12:00noon to 1:00pm Computing flowstates.ppt 137KB (Speaker Slides) Ron.mp3 9758KB (Speaker Audio (appologies for poor audio quality)) Show Abstract The activities in networks all take the form of flows. Intuitively, these flows are obvious – some correspond to file downloads, some to web page access, and others to Internet banking. However, it is not so clear where a flow begins and ends, or where to draw the boundary between one flow and another. This is quite frustrating because there are great opportunities for making networks perform better by discriminating between flows. Large file downloads, for example, will receive virtually perfect performance,, even if they are given lowest priority, and by doing this all the small flows will benefit a great deal. In this seminar we define a vector space of flow states. How should we define the closeness of one flow state to another?. The mathematical field of topology provides a way to do this. In this talk we will discuss some ways to define a topology (how close one flow state is to another) and how this might be useful. In particular, the issue of defining flows more precisely is bypassed by this approach because instead of attempting to standardize the definition of flow we can talk about how similar one flow state is to another, and this is all that is really necessary.Hide Abstract
Speaker: Normal form transforms separate slow stochastic modulation from stochastic oscillatory systems Tony Roberts  (Mathematics and Computing, USQ) D303, Thursday 25 January, 2007, 11:00am to 12:00noon Mathematics nftsssmsos.pdf 313KB (Speaker Slides) Tony.mp3 9647KB (Speaker Audio (appologies for poor audio quality)) Show Abstract Modelling stochastic systems has many important applications. Normal form coordinate transforms are a powerful way to untangle interesting long term dynamics from undesirably detailed microscale dynamics. We explore normal forms of stochastic differential equations when the stochastic dynamics are oscillatory. A stochastic Hopf bifurcation provides a specific example to illustrate general principles. We extract a model for the stochastic modulation of the oscillations. The results will help us accurately model multiscale stochastic systems. Hide Abstract
Speaker: Simple Guide to Prospective PhD Students Shahjahan Khan  (Mathematics and Computing, USQ) D303, Tuesday 23 January, 2007, 2:00pm to 3:00pm Other Show Abstract Steps to start a PhD DegreeEntry requirementsMajor Research Areas in the DeptFocus on individual research interests and research profileContacts of prospective supervisorsInfo on programs and applicationHide Abstract
Speaker: A Comparative Study of Classification Methods for Microarray Data Analysis Hong Hu  (Student, Mathematics and Computing, USQ, ) D303, Thursday 7 December, 2006, 11:00am to 12:00noon Computing ausdm1.ppt 216KB (Speaker Slides) Show Abstract In response to the rapid development of DNA Micro array technology, many classification methods have been used for Microarray classification. SVMs, decision trees, Bagging, Boosting and Random Forest are commonly used methods. In this paper, we conduct experimental comparison of LibSVMs, C4.5, BaggingC4.5, AdaBoostingC4.5, and Random Forest on seven Microarray cancer data sets. The experimental results show all ensemble methods outperform C4.5. The experimental results also show that all five methods benefit from data pre-processing, including gene selection and discetization, in classification accuracy. In addition to comparing the average accuracies of ten-cross validation tests on seven data sets, we use two statistical tests to validate findings. We find that Wilcoxon signed ranks test is better than sign test for such purpose.Hide Abstract
Speaker: XForms In Practice - Implementations and Examples Suleiman Odat  (Student, Mathematics & Computing, USQ) D303, Thursday 7 December, 2006, 10:00am to 11:00am Computing Show Abstract XForms is another Web form technology that enables users to gather information online. They illustrate new ways of dealing with forms and information entries in the client-side, and of capturing this data and processing them in the server-side. These methods reduce the time and work required to author online forms, and maximize the quality of gathering and capturing their data to provide a rich user interaction with the form, and to enable XML tools to validate and process the captured data. In this presentation, we show the advantages of using the W3C XForms over other Web forms technologies such as the W3C HTML forms, Adobe LiveCycle forms, and Microsoft InfoPath forms by providing real examples to author and process forms online.Hide Abstract
Speaker: Universal regimes in dynamics of a free turbulent jet: full k-epsilon model Dmitry Strunin  (Mathematics and Computing, USQ) D303, Thursday 30 November, 2006, 11:00am to 12:00noon Mathematics Show Abstract This is the final part of our analysis of universal regimes in the k-epsilon model of expanding turbulent jet driven by the turbulent diffusion. We start by recalling previous results on a drastically simplified version of the model, for which we found an attractor explicitly. For the full model the situation is much more complex, however the analysis indicates that the attractor exists here as well. This work can be viewed as a kind of "proof of existence" but unlike numerous purely mathematical existence theorems which do not tell how the attractor looks we (approximately) found it.Hide Abstract
Speakers: Using Google Docs to Collaborate on Papers and Study Books Stijn Dekeyser, Richard Watson, Peter Sefton, Expand Stijn Dekeyser  (Mathematics and Computing, USQ) Richard Watson  (Mathematics and Computing, USQ) Peter Sefton  (RUBRIC Technical Manager, DeC and Management Services, USQ, ) Simplify D303, Thursday 30 November, 2006, 10:00am to 11:00am Computing GoogleDocsSeminar.pdf 95KB (Speaker Slides) Show Abstract Researchers and Lecturers have many software options to write papers and course material on their own, but fewer options when they want to collaborate with others. USQ has developed some systems (eg GOOD and ICE) that can be used by non-technical persons, each having strengths and weaknesses. In this seminar we present our findings on using Google Docs (formerly Writely.com) and compare it to other approaches. A live demo of Google Docs is included in the seminar. In addition, we demo an extension to the system that we developed to generate high-quality output. Hide Abstract
Speakers: The State of the Art in Game AI Standardisation Billy Yue, Penny de Byl, Expand Billy Yue  (Student, Mathematics and Computing, USQ, ) Penny de Byl  (Mathematics and Computing, USQ) Simplify D109, Thursday 23 November, 2006, 12:00noon to 1:00pm Computing billy.mp3 9370KB (Speaker Audio) CGIE2006.ppt 2567KB (Speaker Slides) http://www.cgie2006.murdoch.edu.au/program.html (Program for the CGIE2006 conference) http://www.igda.org/ai/ (The AIISC homepage) Show Abstract The Artificial Intelligence Interface Standards Committee (AIISC) was formed in 2002 to develop interface standards for AI middleware in the computer game industry. In this seminar, we provide an overview of the work completed to date by the working groups making up the AIISC. We describe the issues each group is addressing and examine their relevance and importance to the domain of computer games development, and discuss their benefits and shortcomings. We also outline the current progress of each group, and the direction they will be taking in the future in order to implement a usable and widely accepted game AI API.Hide Abstract
Speaker: Securing Service Sharing over Networks for Mobile Users Using Service Network Graphs David Lai  (Mathematics and Computing, USQ) D109, Thursday 16 November, 2006, 11:30am to 12:00noon Computing Show Abstract Sharing of services among wired and wireless networks provides users with more variety of services. In this paper, we concentrate on how to achieve secure service sharing for mobile users in an aggregation of heterogeneous networks. Our approach is to extended the Service Network Graph (SNG) concept to handle mobile users rambling within a guest network which is part of an SNG. Adhering to this approach, if the guest network has delegated authentication authority to the home network, the extended Service Network Graph can also set up a transient account by sharing a secret between the user and the authentication servers on the guest network and the home network. This may serve as a temporary account and facilitates mobile users to obtain services on the fly. Lastly, we discuss how to use the extended SNG for mobile IP authentication.Hide Abstract
Speaker: Improving the Performance of Stream Control Transmission Zhongwei Zhang  (Mathematics and Computing, USQ) D109, Thursday 16 November, 2006, 11:00am to 11:30am Computing SCTP.ppt 152KB (Speaker Slides) Show Abstract Stream Control Transmission Protocol (SCTP) is a transport protocol combining the advantages of TCP and UDP. SCTP has many desirable features including multihoming, multistreaming, and partial data reliability. These features have made SCTP perform much more effective in multimedia networking applications and more flexible in enhancing security, they also operate better in wireless environment which traditional transport protocols are ineffective and cumbersome. Before the transmission, an application using SCTP needs to establish an `association' between the client and the server. The establishment of the association requires a number which will be used to create multiple streams. However, SCTP has not specified a method or suggested any ideas of determining the number. In our paper, we focus on the SCTP's performance over the wireless networks. The ideas is to extend the SCTP with a process of determining an optimal number prior to the association establishing. We examine the modified SCTP on a simulated wireless network, and the experiment results of the simulation using NS2 have shown the modified SCTP is feasible and also demonstrated the modified SCTP's superiority of performance over TCP and UDP over the wireless networks.Hide Abstract
Speaker: Computer Game Demo Presentations by CSC3418 (Computer Games Programming) Students D109, Monday 30 October, 2006, 12:00noon to 2:00pm Computing Show Abstract Everyone is invited to attend the student’s final presentations of the computer game demos they have been creating for CSC3418 in S2, 2006. The presentations will be given by 4 teams of students who will demonstrate their game idea, run a short example of the game and briefly outline how they have developed their application. This presentation forms part of their final assessment.Hide Abstract
Speakers: Applied Attitude Pathfinding Quynh Duong, Penny de Byl, Expand Quynh Duong  (Student, Dept Maths and Computing, USQ) Penny de Byl  (Mathematics and Computing, USQ) Simplify D109, Thursday 26 October, 2006, 11:30am to 12:00noon Computing pippy.mp3 5031KB (Speaker Audio) presentation.ppt 630KB (Speaker Slides) Show Abstract This work examines the possibility of applying Search Relevancy technique with Attitude Pathfinding. A simplified search algorithm is used and a relevancy percentage is assigned to the search results. Most relevant results will then be chosen as preferred waypoints for an artificial intelligent bot to navigate using a modified A Star pathfinding algorithm called Attitude Pathfinding. With this system, we want to demonstrate the interactive aspect of Attitude Pathfinding in practice as well as exploring the capabilities of combining two AI fields.Hide Abstract
Speakers: Procedural Generation Terrains for Games using Fractals and L-systems Kae Thong Chew, Penny de Byl, Expand Kae Thong Chew  (Student, Dept Maths and Computing, USQ) Penny de Byl  (Mathematics and Computing, USQ) Simplify D109, Thursday 26 October, 2006, 11:00am to 11:30am Computing chew.mp3 7310KB (Speaker Audio) Research Proposal Slide ver1.ppt 2283KB (Speaker Slides) Show Abstract As the worlds within computer games becoming more complex with each generation of hardware and software, it is increasingly more burdensome for artists and developers to create these detailed worlds as the customers have high taste of computer graphics visual realism and demand more from game developers Therefore, alternative methods have been researched to approach this problem and procedural content generation method is an approach that can offload some of the routine, repetitive rendering tasks to the software or graphics hardware, so that the human resources can be applied to the non-routine and creative aspects of game development. This project investigates the development of procedural generation terrains for games by using fractals and L-systems to generate them. The aim is to design and investigate the use of L-systems fractals in generating a game terrain map suitable to be implementing in a game environment. The result will try to yield new methods and experimental results on how a 3-D game engine will react towards the game map generating procedurally by producing new theory result or disprove the fact of using L-systems fractals to generate game map terrains. The models which will be created can prove to be a reference material and as a basis for future research and develop.Hide Abstract
Speakers: Benefits of Linking Modelling To Experimental Approaches – A Case Study For Wind Erosion John Leys, Harry Butler, Expand John Leys  (Department of Natural Resources, ) Harry Butler  (Mathematics and Computing, USQ) Simplify T122, Wednesday 25 October, 2006, 11:00am to 1:00pm Mathematics Benefits of Linking Modelling To Experimental Approaches.doc 35KB (Outline Document) CEMSYS talk Oct 06.pdf 19506KB (Speaker Slides) JohnLeys.mp3 12286KB (Speaker Audio) Show Abstract In recent years, considerable effort has been made in constructing integrated dust modelling systems which couple modules for atmospheric, land-surface and aeolian processes with land-surface parameter databases. The aim has been to capture the main processes and environmental factors for dust-storm development, transport and deposition. Integrated modelling of Australian dust storms has only been carried out a few times in Australia, notably (Shao and Leslie 1997) and (Shao et al. 2006). Both these studies reported a lack of experimental data were available to validate the model against. The most extensive observational data set available is that of the Bureau of Meteorology (BoM) which contains data on wind erosion activity across Australia for the last 45 years and at selected locations, records go back over 100 years. These formal meteorological records have been used to map wind erosion activity in Australia since the early work of (McTainsh and Pitblado 1987) and have since been quality assessed, refined and stored in the Australian Dust Event Database (DEDB) at Griffith University. As Leys et al (2006) report, the DEDB has significant gaps in the data, both temporally and spatially, and they have established DustWatch in an attempt to resolve some of these problems. We are now at the point of being able to link the modelling with experimental data thus increasing the value of both products. This approach has been acknowledged by Catchment Management Authorities (CMA) who are supporting DustWatch, the Desert Knowledge CRC who are co-funding the modelling with the CMAs and the Land and Water Audit, who is funding trails that will test the functionality of the DustWatch / modelling approach to monitoring and resource assessment . This presentation will discuss the benefits of combining modelling and experimental observations for the monitoring and understanding of wind erosion in Australia. Hide Abstract Show References Leys J, McTainsh G, Strong C, Heidenreich S, Biseaga K (2006) DustWatch: Using community networks to improve wind erosion monitoring in Australia. In ' 6th International Conference on Aeolian Research.' (Eds W Nickling and J Gillies) (University of Guelph: Guelph). McTainsh GH, Pitblado JR (1987) Dust storms and related phenomena measured from meteorological records in Australia. Earth Surface Processes and Landforms 12, 415-424. Shao Y, Leslie LM (1997) Wind erosion prediction over the Australian continent. Journal of Geophysical Research 102, 30,091-30,105. Shao Y, Leys JF, McTainsh GH (2006) Numerical simulation of a severe dust event in eastern Australia. Journal of Geophysical Research. (Accepted)Hide References
Speaker: Reversed Engineering of Gene Networks for Bovine Skeletal Muscle: Development and Applications Toni Reverter-Gomez  (CSIRO Livestock Industries, Brisbane, Australia, ) D109, Thursday 19 October, 2006, 11:00am to 1:00pm Bioinformatics Reverter_USQ_Seminar.ppt 8306KB (Speaker Slides) Show Abstract This work examines how gene networks for bovine skeletal muscle can be inferred from expression data that spans a comprehensive number of experimental treatments in muscle and fat tissues. We use the example of myogenin (MYOG), a muscle-specific transcription factor essential for the development of skeletal muscle, and develop a network comprising 22 genes (or network nodes) that are connected by 33 significant associations. With this network, we simulate an evolutionary process based on digital organisms to address two fundamental questions: What global changes are required in the MYOG network to generate an extreme organism? And, what (re)configurations in the network can be expected that results in a minimal and maximal change in protein output after knocking out MYOG? Results revealed the existence of two distinct clusters of genes with coordinated expression within the network. Speaker Bio Dr Antonio (Toni) Reverter is Principal research scientist with the Bioinformatics Group of CSIRO Livestock Industries. Toni's background is in statistical genetics, more specifically in methods for large-scale genetic evaluation and parameter estimation. His work in CSIRO involves the statistical analysis of gene-expression and mapping data including whole-genome SNP genotypes for complex traits in livestock. More specifically, he develops and applies novel mathematical, computational and statistical methods to disentangle the information encapsulated in datasets generated by high-throughput genomic techniques. Toni was the recipient of the inaugural 2005 Eureka Prize for Bioinformatics.Hide Abstract
Speaker: Two-Equation Model Resolves Meam Flow Resonances in Subcritical Flow Systems Sergey Suslov  (Mathematics and Computing, USQ) D109, Thursday 12 October, 2006, 12:00noon to 1:00pm Mathematics aims06v.pdf 358KB (Speaker Slides - PDF) aims06v.ps 1432KB (Speaker Slides - PS) Sergey.mp3 11263KB (Speaker Audio) Show Abstract Amplitude equations of the Landau type, which describe the dynamics of the most amplified periodic disturbance waves in slightly supercritical flow systems, have been known to form reliable and sufficiently accurate low-dimensional models capable of predicting the asymptotic magnitude of saturated perturbations in weakly nonlinear regimes beyond the bifurcation point. However the derivation of similar models capable of predicting the threshold disturbance amplitude (below which the solution returns to its undisturbed state and above which it undergoes the transition to a different nonlinear state) in systems bifurcating subcritically faces multiple resonances which lead to the singularity of model coefficients. Here we suggest a methodology of deriving a two-equation dynamical system of coupled cubic amplitude equations with non-singular coefficients which resolve the resonances and are capable of predicting the threshold amplitude for weakly nonlinear subcritical regimes and the qualitative features of such flows. As an example, a developed procedure is applied to a system of Navier-Stokes equations describing a subcritical plane Poiseuille flow. The developed model predictions are found to be in reasonable agreement with experimentally detected threshold amplitudes reported in literature.Hide Abstract
Speaker: The Way LMSs Should Be Scott Sorley  (ICT Infrastructure & Systems, USQ, ) D109, Thursday 21 September, 2006, 12:00noon to 1:00pm Computing LMS-SJS.ppt 1285KB (Speaker Slides) ScottSorley(MissingLastFewMinutes).mp3 5896KB (Speaker Audio (missing last few minutes)) http://www.sci.usq.edu.au/USQstaff/wiki/doku.php?id=start:lms_innovation_at_usq (Discussion "Should USQ Be Innovative with its LMS" (Restricted Access)) Show Abstract One person's (hopefully informed) personal opinion on where current and past Learning Managements Systems fail, why they aren't strategic systems for Universities and how they fail tutors, students and administrators. Topic will be expanded with ideas where future direction and LMS development should go and how these pervasive systems can be transformed into strategic business drivers.Hide Abstract
Speakers: Unlearn what you have learned about the LMS Stijn Dekeyser, Penny de Byl, Michael de Raadt, Expand Stijn Dekeyser  (Mathematics and Computing, USQ) Penny de Byl  (Mathematics and Computing, USQ) Michael de Raadt  (Mathematics and Computing, USQ) Simplify D109, Thursday 21 September, 2006, 11:00am to 12:00noon Computing LMS introduction.ppt 562KB (Speaker Slides - Speaker 1 Stijn Dekeyser) Moodle as an LMS.ppt 362KB (Speaker Slides - Speaker 2 Penny de Byl) Speaker1StijnDekeyser.mp3 6738KB (Speaker Audio - Speaker 1 Stijn Dekeyser) Speaker2PennydeByl.mp3 4454KB (Speaker Audio - Speaker 2 Penny de Byl with Richard Young) Speaker3MichaeldeRaadt.mp3 3180KB (Speaker Audio - Speaker 3 Michael de Raadt) Unlearn what you have learned about the LMS.ppt 220KB (Speaker Slides - Speaker 3 Michael de Raadt) http://moodle.org/ (Moodle Website) http://www.educause.edu/ir/library/pdf/ers0302/rs/ers0302w.pdf (EduCause ECAR Study) http://www.sci.usq.edu.au/moodle (Dept M&C Moodle Site) Show Abstract What is a Learning Management System? What technologies are behind them and what differentiates them? Who's using what LMS and why? What can we do with an LMS in the future? Come and join in a topical discussion and the University decides its LMS future.Hide Abstract
Speaker: Progressive Teaching with Tablet Technology Birgit Loch  (Mathematics and Computing, USQ) D109, Thursday 14 September, 2006, 12:00noon to 1:00pm Mathematics birgit.mp3 12116KB (Speaker Audio) USQ_tablet_talk_14-9-06after.ppt 1252KB (Speaker Slides) Show Abstract Non-interactive computer technologies, such as PowerPoint slides, are standard tools used in modern lectures, yet, if slides are prepared entirely before delivery, the effectiveness of these methods for the cognition of (mathematical) concepts is questionable. Such formats limit flexibility and encourage passive learning, and do not engage students since lectures cannot be adjusted based on audience reaction. In this talk, I will describe results from a two year study into using handwriting on the computer for on-campus lecture presentations. I will begin with a motivation for the implementation of this technology, concentrating on highly visual disciplines such as - but not restricted to - those using mathematical symbols. I will then explain the various hardware and software configurations used during this study, and point out benefits and problems arising from each. Student and instructor attitude towards this technology will be discussed, and I will talk about the applicability for lecture delivery to distance students using screen capture software. I will conclude with an outline of future directions and a brief look at alternative hardware setups. Part of this study was conducted with Associate Professor Diane Donovan, University of Queensland.Hide Abstract
Speaker: A normal form coordinate transform simplifies stochastic dynamics Tony Roberts  (Mathematics and Computing, USQ) D109, Thursday 14 September, 2006, 11:00am to 12:00noon Mathematics tony.mp3 12124KB (Speaker Audio) ztalk.pdf 310KB (Speaker Slides) Show Abstract Modelling stochastic differential equations has many important applications. Normal form coordinate transforms are a powerful way to untangle interesting long term dynamics from uninteresting but necessary microscale dynamics. We will explore normal forms of stochastic differential equations when the dynamics has both slow modes and quickly decaying modes. The thrust is to derive normal forms useful for macroscopic modelling of detailed microscopic systems. Thus I endeavour to remove all fast time processes, as well as reducing the dimensionality of the dynamics. Sri Namachchivaya, Leng and Lin (1990) emphasise the importance of quadratic stochastic effects "in order to capture the stochastic contributions of the stable modes to the drift terms of the critical modes." We too will derive such important quadratic effects. The results will help us model accurately multiscale stochastic transitions and stochastic discretisations. Hide Abstract
Speaker: Ageing and aged-care enabled by innovative technology: the state-of-the-art and current research Jeffrey Soar  (Collaboration for Ageing & Aged Care Informatics Research, USQ, ) D109, Tuesday 12 September, 2006, 12:00noon to 1:00pm Statistics 06 sci coll3.ppt 6558KB (Speaker Slides) JeffSoar.mp3 14036KB (Speaker Audio) Show Abstract Most developed countries are experiencing unprecedented increases in the percentages of older people. The population 65 years and over is projected to more than double by 2050. The greatest rate of growth is in the population aged 85 years or over. This group will have the greatest need for health and disability support. It is not only the increasing numbers of people requiring care in the future but also social changes that will increase pressures on aged care and health services. The level of informal care that was provided for frail aged family members some decades ago is less viable today. Governments have responded with a range of strategies and policy initiatives such as abolition of the compulsory retirement age, promoting productive ageing and enhancing the capacity for home-based care. There is particular interest in technology to assist active ageing and aged care to enable extending active and productive lives and facilitate care to be delivered according to consumer preferences for place and time. These changes will reflect the impacts of technologies that have transformed other industries, enabled new products and services and provided delivery at the convenience of the consumer. There will be many opportunities for new and imaginative research. Research is needed to guide changes that will be required in policy, strategy, funding, work-practices, integration of technology into care, future roles for the professions and new models of care. This presentation will discuss the state-of-the-art in assistive technology for active ageing and care in the home, the gaps and research opportunities. It will also discuss the research program of the CAAIR research cluster at USQ: http://www.usq.edu.au/business/research/conferences/aging/default.htm Speaker Bio Jeffrey Soar came to academia from a long career in IT management in health and aged care. He held CIO/IT Manager positions in Veterans’ Affairs, NSW Health and the New Zealand Ministry of Health. He is the founder and Secretary of the national association, Aged Care Informatics Australia http://websites.golden-orb.com/hisa2/100132.php and founder and organiser of the annual Aged Care Informatics Conference. His research focuses on the benefits of ICT for active ageing and aged care. Hide Abstract
Speakers: Do students SQLify? Improving Learning Outcomes with Peer Review and Enhanced Computer Assisted Assessment of Querying Skills Lee Tien Yu, Stijn Dekeyser, Michael de Raadt, Expand Lee Tien Yu  (Masters Student, Mathematics and Computing, USQ, ) Stijn Dekeyser  (Mathematics and Computing, USQ) Michael de Raadt  (Mathematics and Computing, USQ) Simplify D109, Thursday 31 August, 2006, 12:00noon to 1:00pm Computing Lee Tien Yu.mp3 8815KB (Speaker Audio) preliminary[2].ppt 110KB (Speaker Slides) Show Abstract In this seminar, I am going to introduce the design and implementation of a new tool, called SQLify, which is an SQL teaching and assessment tool that extends the current state-of-the-art by incorporating peer review and enhanced automatic assessment to arrive at a richer learning experience for students. I will examine difficulties students have in learning SQL and introduce various features integrated into SQLify which can help students improve their SQL skills.Hide Abstract
Speaker: Classical and Bayesian Approaches to prediction distribution for multivariate simple regression model with multivariate normal errors Shahjahan Khan  (Mathematics and Computing, USQ) D109, Thursday 31 August, 2006, 11:00am to 12:00noon Statistics pres-pred-usq-aug06.pdf 118KB (Speaker Slides) Shahjahan Khan.mp3 9992KB (Speaker Audio) Show Abstract The classical and Bayesian approaches to the prediction distribution for a set of future responses for a multivariate simple regression model with unknown slope, intercept and scale parameters are discussed in this talk. The errors of the model, from both the performed and future experiments, are identically and independently distributed as multivariate normal variables. Conditional on a set of realized responses, another set of the future unrealized responses follow a matrix T distribution. The number of degrees of freedom of the prediction distribution depends on the size of the realized sample, and the dimension of the regression parameters in the model. The prediction distribution obtained by the classical method is the same as that derived under the Bayesian method with uniform prior. Some applications of the prediction distribution are discussed.Hide Abstract
Speaker: Optimal rule discovery and applications Jiuyong Li  (Mathematics and Computing, USQ) D109, Thursday 24 August, 2006, 12:00noon to 1:00pm Computing Optimal.ppt 536KB (Speaker Slides) Show Abstract Rules are one of the most expressive and human understandable representations of knowledge; a rule based method produces self-explanatory results. Therefore, rule discovery has been a major issue in machine learning and data mining. Heuristic and association rules are two dominant schemes for rule discovery. However, heuristic rule discovery fails to find many globally optimal rules. Association rule discovery yields too many rules and is infeasible for many applications. In this talk, I will present an efficient alternative for rule discovery, optimal rule discovery, and discuss its relationships with other rule discovery schemes. I will also discuss its applications. One application is to build classification systems that tolerate missing values. Another application is to find risk patterns in a real world medical data.Hide Abstract
Speaker: The Future of File Storage and Metadata Lasse Motroen  (Mathematics and Computing, USQ) D109, Thursday 17 August, 2006, 12:00noon to 1:00pm Computing LasseMotroen.mp3 9355KB (Speaker Audio) Seminar.ppt 292KB (Speaker Slides) WinFS_IWish_720x486_2mbs.wmv 40289KB (Video Shown) Show Abstract People are, today, faced with the challenge of organizing a growing set of files and information. The traditional hierarchical file system generally imposes the restriction that a file may only belong to a single folder or directory and as a result that file can be associated with no more than one category. Most files could be placed in several categories, based on their metadata, however organizing files in this manner may be difficult for the user. For this reason, Metadata File Systems, along with several applications designed to assist in the organization of files, have been proposed. In my presentation, I will endeavor to provide an overview of proposed solutions to this problem, discuss their shortcomings and present how this metadata problem might be solved in future. I will also discuss some concepts that should be applied when designing a User Interface for a Metadata File System.Hide Abstract
Speaker: Improved Estimation in Dynamic Linear Regression Model Zahirul Hoque  (Mathematical and Physical Sciences, University of Newcastle, ) D109, Thursday 10 August, 2006, 12:00noon to 1:00pm Statistics Hoque-Gerlach(JSPI).pdf 131KB (Paper Presented) ZahirulHoque.mp3 9244KB (Speaker Audio) ZH-Canada.pdf 77KB (Speaker Overheads) Show Abstract This paper studies the preliminary test and shrinkage estimators of linear state space regression model via Kalman filtering. The performance of the estimators, with respect to mean square error, has been investigated. It has been revealed that under certain conditions both preliminary test and shrinkage estimators outperform Kalman filter but shrinkage estimator is superior to the preliminary test estimator. Hence, our result invalidates the minimum mean square error property of Kalman filter. Hide Abstract Show References Bancroft, T.A. (1944). On biases in estimation due to the use of the preliminary test of significance, Ann. of Math. Stat., 15, 190-204. Duncan, D.B. and Horn, S.D. (1972). Linear Dynamic Recursive estimation from the Viewpoint of Regression Analysis, Jour. of Am. Stat. Asso., 67, 815-821. Gerlach R., Carter C. and Kohn R. (2000). Efficient Bayesian inference in dynamic mixture models, Jour. of Am. Stat. Asso., 95(451), 819-828. Kalman, R.E. (1960). A New Approach to Linear Filtering and Prediction Problems, Jour. of Bas. Eng., 82, 34-45. Kalman, R.E. and Bucy, R.S. (1961). New Results in Filtering and Prediction Theory, Jour. of Bas. Eng., 83, 95-108. Khan, S., Hoque, Z. and Saleh, A.K.Md.E. (2002). Estimation of the slope parameter for linear regression model with uncertain prior information, Jou. of Stat. Res., 36(1), 55-73. Khan, S., Hoque, Z. and Saleh, A.K.Md.E. (2005). Estimation of the intercept parameter for linear regression model with uncertain non-sample prior information, Stat. Pap., 46(3), 379-395. Meinhold, R.J. and Singpurwalla, N.D. (1983). Understanding the Kalman Filter, The Am. Stat., 37(2), 123-127.Hide References
Speaker: Modelling MIMO Technologies for Indoor WLANs Noel Roshan Francis  (Masters Student, Mathematics & Computing, USQ) D109, Thursday 10 August, 2006, 11:00am to 12:00noon Computing Modelling MIMO Sim 2.ppt 194KB (Speaker Slides) NoelRoshanFrancis.mp3 6157KB (Speaker Audio) Show Abstract In this presentation I will provide an overview on the modelling of Multiple-Input Multiple-Output (MIMO) wireless systems in an indoor environment. The performance of a MIMO channel is primarily affected by multi-path and shadowing effects caused by the surrounding environment. I have modelled an indoor 2x2 MIMO system using the open source Network-Simulator 2 (NS2) and measure its performance through the system bit-error rate (BER), which is a measure of the reliability of the entire wireless system. Simulation results indicate that the performance of MIMO systems in indoor environments is considerably better than conventional Single-Input Single-Output (SISO) systems used in most commercially available wireless networks.Hide Abstract
Speaker: Nature of mathematical discovery and inquiry activities Oleksiy Yevdokimov  (Mathematics and Computing, USQ) D109, Thursday 27 July, 2006, 11:00am to 12:00noon Mathematics OleksiyAudio.mp3 14724KB (Speaker Audio) RR_PME30_Yevdokimov.ppt 168KB (Speaker Slides) Show Abstract How we deal with the nature of mathematical discovery and inquiry activities has always had a strong focus among many researchers' interests. Nowadays the phenomenon of mathematical discovery, its mechanism and mental processes remain in the educational research spotlight. Indeed, the concepts of inquiry and mathematical discovery have many common features with the learning process and can be considered together. Much previous work on the process of mathematical discovery in the mathematics education literature has often concentrated on mathematicians and their research practice without clear indication of the needs and objectives of the learning process and the subjects involved in it. I will present the results of research which focused on studying students' inquiry work from a psychological point of view. Inquiry activities of students in a classroom were analysed through the evaluation of the character of these activities within the learning process with regard to a mathematician's research practice.Hide Abstract
Moodle Workshop
Speakers:
Khaleel Petrus, Penny de Byl, David Lai, Expand
Khaleel Petrus  (Mathematics and Computing, USQ)
Penny de Byl  (Mathematics and Computing, USQ)
David Lai  (Mathematics and Computing, USQ)
Venue/Date/Time:C205, Thursday 20 July, 2006, 9:00am to 12:30pm
Discipline:Computing
Abstract:
RSVP:There are limited spaces in the workshop due to limited number of PCs in the lab. Email petrus@usq.edu.au by Monday 17th July. Kindly supply USQ log-in, course code and course name to create a Moodle account. Also, please have your course materials accessible online (perhaps on your H: drive).
Details:You will be guided and helped to setup Moodle for your course in an easy and step by step way.
Schedule:
 Session 1 9:00-10:40 Course setup Penny & David Lai Break 10:40-11:00 Tea and Biscuits Session 2 11:00-12:30 Assignments Khaleel
Speaker: Implementation of AQMs on Linux made Easy Stephen Braithwaite  (PhD Student, Mathematics and Computing, USQ, ) D303, Thursday 6 July, 2006, 12:00noon to 1:00pm Computing finalPres.pdf 240KB (Speaker slides) StephenAudio.mp3 11705KB (Speaker Audio) http://www.sci.usq.edu.au/staff/braithwa/MastProj/ (Stephen's Project Resource Page) Show Abstract This seminar provides a report on a project to implement a Mice and Elephants queueing discipline, which favours short flows over long flows, on Linux. The project has three aims. The first aim of the project is to produce a prototype Mice and Elephants router for the purpose of further evaluation of the Mice and Elephants strategy and the Shortest Job First strategy. A demonstration of the router will be provided after the seminar. The second aim of the project is to make a contribution to Linux by making my implementation as code that is both fit for distribution with Linux and useful in a small business or domestic setting. The third aim is to explore and document a method of creating Linux queueing disciplines in general. Hide Abstract
Speakers: Moodle online learning environment Khaleel Petrus, Penny de Byl, Stijn Dekeyser, Expand Khaleel Petrus  (Mathematics and Computing, USQ) Penny de Byl  (Mathematics and Computing, USQ) Stijn Dekeyser  (Mathematics and Computing, USQ) Simplify D303, Thursday 6 July, 2006, 11:00am to 12:00noon Computing MoodleSeminarKhaleel.mp3 7789KB (Speaker Audio (Speaker 1 - Khaleel Petrus)) MoodleSeminarPenny.mp3 3453KB (Speaker Audio (Speaker 2 - Penny de Byl)) MoodleSeminarStijn.mp3 1921KB (Speaker Audio (Speaker 3 - Stijn Dekeyser)) Seminar_Moodle.ppt 1118KB (Snapshots from demonstration) http://cdc.humboldt.edu/lms/whatis.html (What is Moodle?) http://www.egovmonitor.com/node/3460 (Open University going Moodle) http://www.sci.usq.edu.au/research/seminars/?seminarID=21 (Previous web services + Moodle seminar) Show Abstract Moodle is an open source online environment and is praised as one of the successful distance education environments. Among the interesting and extensive evaluations of Moodle against Blackboard (proprietary) Professor Kathy D. Munoz, of Humboldt State University reported (Feb 2005) student satisfaction; (Which do you prefer, Moodle or Blackboard?) was as follows:No preference - 42.9%Moodle - 35.7%Blackboard - 21.4% We will introduce Moodle and present our experiences using it and also some maintenance, installation, Gradebook and students enrolment lists issues will be addressed briefly. The usage of the environment will be introduced in step-by-step hands-on style to explore Moodle. Among the environment capabilities explored will be: adding files, updating links, assignment submission, marking feedback, surveys, students interaction activities, discussion groups, Wikis and online assessment capabilities.Hide Abstract
Speaker: Resolve subgrid and interelement interactions to discretise stochastically forced diffusion Tony Roberts  (Mathematics and Computing, USQ) D109, Thursday 29 June, 2006, 11:00am to 12:00noon Mathematics talk.pdf 245KB (Speaker slides) Show Abstract Constructing discrete models of stochastic partial differential equations is very delicate. I apply modern dynamical systems theory to support spatial discretisations of the stochastically forced diffusion equation. The trick to the application of centre manifold theory is to divide the physical domain into finite sized elements by introducing insulating internal boundaries which are later removed to fully couple the dynamical interactions between neighbouring elements. The approach automatically parametrises the microscale, subgrid structures within each element induced by spatially varying stochastic forcing. The crucial aspect of this work is that we explore how a multitude of noise processes interact within and between neighbouring elements. Noise processes with coarse structure across a finite element are the most significant noises for the discrete model. Their influence also diffuses away to weakly correlate the noise in the spatial discretisation.Hide Abstract
Speaker: Microarray Gene Expression Data Analysis: a comparison of false discovery rate controlling procedures Anita Pelecanos  (Student, Mathematics & Computing, USQ, ) D109, Thursday 22 June, 2006, 11:00am to 12:00noon Bioinformatics Show Abstract Abstract: Microarrays allow for fast evaluation and illustration of differential expression (variation in the amount of protein present in a given cell or tissue) of thousands of genes on an entire genome scale. Current statistical inference problems with microarray data regularly involve thousands of genes that implicte thousands of hypothesis tests. A problem with having many hypothesis tests is controlling the Type I error rate (the rate at which pronouncing a gene is differentially expressed when it is not). The Type I error rate increases dramatically with the number of hypotheses. Many procedures have been proposed to deal with these issues by controlling different error rates.Hide Abstract
Speaker: Stein's Unbiased Risk Estimator (SURE) for optical flow Mingren Shi  (Mathematics and Computing, USQ) D109, Thursday 25 May, 2006, 11:00am to 12:00noon Statistics OpSureTalk.pdf 126KB (Speaker overheads) Show Abstract Optical flow estimation is a problem raising from image restoration. Optical flow is the apparent motion of the brightness/intensity patterns observed when a camera is moving relative to the objects being imaged. Optical flow equation is ill-posed; there are two variables (optical flow vector has horizontal and vertical directions), but only one equation. Hence it has infinity many solutions. To obtain a unique solution, a constraint should be imposed on optical flow vector. Local constant Lucas-Kanade method is an effcient one of regularization or smoothing methods to estimate the optical flow. The smoothing parameter is the neighbourhood size n under the assumption that the optical flow in a neighbourhood of the central pixel is a constant. The size n, however, is tradeoff. If it is too small, the numerical errors may be excessively amplified. If it is too large, the solution may be not close to the real solution. So there is an optimal value of the n size. Stein's Unbiased Risk Estimator (SURE) was originally used to estimate the mean of a multivariate normal distribution. We use it to find the optimal value of the tuning parameters n in the context of image restoration. Numerical experiments illustrate that SURE is effective for this purpose.Hide Abstract
Speakers: E-ink liberates online learning Birgit Loch, Christine McDonald, Expand Birgit Loch  (Mathematics and Computing, USQ) Christine McDonald  (Mathematics and Computing, USQ) Simplify D109, Thursday 18 May, 2006, 11:00am to 12:00noon Mathematics Show Abstract Electronic Ink (e-Ink) allows the user to write and draw on a computer; it is the electronic equivalent of ink. In this presentation we will discuss the feasibility and demonstrate the use of a freely available chat client for the teaching of distance students. This client incorporates e-Ink, which is driven by a mouse or stylus. Advantages are obvious for highly visual disciplines (such as mathematics), since it is now possible to write (eg formulae) and draw (eg diagrams, graphs) on a computer, in a chat environment, in real time. We report on outcomes of a pilot study conducted in two first year undergraduate mathematics-based courses, both offered in distance mode only, with a group of volunteering students. A Departmental Learning and Teaching Grant has allowed us to continue this project as an integral part of the problem solving component of a large first year mathematics course.Hide Abstract
Speaker: Investigating artificial intelligence performance in turn based strategy game using genetic algorithm Jiwan Kandangwa  (Student, Mathematics & Computing, USQ) D109, Thursday 11 May, 2006, 11:00am to 12:00noon Computing Show Abstract Conventional turn based strategy games have artificial intelligence (AI) that is hard-coded or based on predefined scripts. As a player gains experience during game play they begin to predict the AIs strategy. This makes it easy to defeat on a regularly basis. Players can also develop new strategies over time whereas the AI strategies are static. This makes the game less enjoyable and the human player loses interest. This seminar will present an investigation into the ways to improve the game-play by dynamically adjusting the AI's strategy for counteracting the player's strategy allowing the AI to observe the player and learn during game play.Hide Abstract
Speaker: Rank set sampling versus simple random sampling in the estimation of mean and ratio Ganes S. Ganesalingam  (Institute of Information Sciences and Technology, Massey University, New Zealand) D109, Thursday 20 April, 2006, 11:00am to 12:00noon Statistics Show Abstract It is common in practice that the experimental units can be ranked easily using measurable covariate than quantification of the main variable of interest which requires expensive measurements. In such a situation the Ranked Set Sampling (RSS) is more beneficial and cost effective when compared to the traditional Simple Random Sampling (SRS). Ranked sat sampling employs judgement ordering to obtain the actual sample and hence yield a sample of observations that is more representative of the underlying population. Therefore, either greater confidence is gained for a fixed number of observations, or for a desired level of confidence, a smaller number of observations is needed. In either way it is a big gain to the researcher in terms of time and money. In this talk I introduce the basic concepts of RSS and illustrate with a real life application.Hide Abstract
Speaker: CISCO university research program Brian Kissell  (CIO, Queensland Studies Authority) D109, Wednesday 12 April, 2006, 11:00am to 11:30am Computing
Speaker: Indoor MIMO and SISO wireless LAN systems Noel Roshan Francis  (Masters Student, Mathematics & Computing, USQ) D109, Thursday 23 March, 2006, 10:00am to 11:00am Computing Show Abstract This presentation covers the detailed description of a comparison study between the performance of indoor Multiple-Output (MIMO) and Single-Output (SISO) Wireless LAN (WLAN) systems. In this project we will show, for the case of multiple-transmit and multiple-receive antennas, how much MIMO systems outperform conventional SISO systems particularly when noise or channel interference from surrounding objects dominates. During the presentation we will describe the selected methodology to quantify these differences for specific cases studied, and discuss further areas of research.Hide Abstract
Speakers: The spoils of innovative teaching Michael de Raadt, Ron Addie, Birgit Loch, Expand Michael de Raadt  (Mathematics and Computing, USQ) Ron Addie  (Mathematics and Computing, USQ) Birgit Loch  (Mathematics and Computing, USQ) Simplify D109, Thursday 16 March, 2006, 10:00am to 11:00am Computing Speaker Notes.pdf 920KB (Speaker notes) http://www.sci.usq.edu.au/staff/deraadt/award (Excellence in Teaching Award nomination information) http://www.sci.usq.edu.au/staff/deraadt/award/documents/breeze (Breeze Presentation) Show Abstract The teaching team for Foundation Computing is successfully combining modern educational methods with innovative delivery. Foundation Computing is offered to more than a thousand students each year from diverse disciplines and backgrounds. The course is training tomorrow's computer users for a rapidly changing computing world. Come and find out how we:Handle up to 4800 paperless student assignments a semester with a small teaching team,Employ electronic peer review,Shift teaching from marking to interactionMake intensive use of tailored communication tools, andCreate better outcomes for students. The teaching team is currently preparing an application for the Faculty Teaching Design and Delivery Award. This presentation gives a preview of our application.Hide Abstract
Speakers: Electronic forms Stijn Dekeyser, Ron Addie, Suleiman Odat, Expand Stijn Dekeyser  (Mathematics and Computing, USQ) Ron Addie  (Mathematics and Computing, USQ) Suleiman Odat  (Student, Mathematics & Computing, USQ) Simplify D109, Thursday 9 March, 2006, 10:00am to 11:00am Computing Show Abstract Electronic forms have been around for a long time and have become commonplace with the advent of the World Wide Web. However, HTML forms have many shortcomings and therefore new technologies, based largely on XML, are being developed. In this seminar we briefly describe what is set to come to your web browser in the coming year or so, before presenting some research results obtained by the MC Forms Research Interest Group over the last 15 months. This includes results published in three recent papers including a world-class publication. We also briefly report on an implementation strategy for our ideas, and present a student research project proposal. Most of the seminar is targeted to a general audience.Hide Abstract
Speaker: Time files: An ADL review (Mathematics and Computing, USQ) D109, Thursday 9 March, 2006, 10:00am to 11:00am Statistics Show Abstract In S2, 2006, I was on Academic Leave, spending time at the Department of Primary Industries and Fisheries, the Walter and Eliza Hall Institute of Medical Research, and the highly esteemed University of Southern Queensland. In this non-technical, talk, I will discuss what I did and what I achieved during this ADL.Hide Abstract
Speaker: Numerical issues of exploring spatio-temporal instabilities Sergey Suslov  (Mathematics and Computing, USQ) D303, Saturday 4 February, 2006, 11:30am to 12:00noon Mathematics Show Abstract The concept of spatio-temporal instability relates to the development of initially localised small amplitude disturbance envelopes superimposed onto the base solution whose stability is investigated. Three drastically different scenarios can be distinguished: the envelope can decay in time and space signifying absolute stability of the base solution; the disturbance amplitude can grow, yet the edges of the envelope can propagate in the same direction leaving an undisturbed field behind which is characteristic for convective instability; and finally, the edges of a growing envelope can move in the opposite directions so that absolute instability is observed when the base solution is eventually perturbed everywhere. These phenomena which arise in a wide spectrum of situations ranging from plasma physics to fluid flows have been extensively studied analytically over the last four decades using asymptotic theory of Fourier integrals representing disturbance envelopes. Yet most of the modern physical applications are too complicated to result in an analytical dispersion relation that would enable "pen and paper" analysis, therefore they require fully numerical treatment. Despite the fact that the theory of spatio-temporal instabilities is nowadays well understood the development of its computationally efficient numerical analogue remains at the leading edge of research. This talk will identify major theoretical steps and numerical difficulties in determining the spatio-temporal character of instabilities and will suggest efficient ways of resolving them for a wide range of physical problems.Hide Abstract
Speaker: Model the inertial dynamics of the thin film flow of power law fluids and other generalised Newtonian fluids Tony Roberts  (Mathematics and Computing, USQ) D303, Thursday 2 February, 2006, 11:00am to 11:30am Mathematics
Speaker: ... a very discrete affair (Mathematics and Computing, USQ) D109, Thursday 12 January, 2006, 11:00am to 12:00noon Mathematics DiscreteAffair.pdf 480KB (Speaker slides)
Speaker: Introduction to ICE (Integrated Content Environment) software and how it uses version control Peter Sefton  (RUBRIC Technical Manager, DeC and Management Services, USQ, ) D109, Wednesday 7 December, 2005, 10:00am to 11:00am Computing Show Abstract ICE is a content management/single-source publishing system which is optimised for writing courseware. ICE uses the Subversion version control system to manage content allowing collaborative, distributed authoring. ICE is cross platform (Linux, Mac OS X, Windows) Open Source Software, released under the GPL. Currently it is being developed by Software Development Services and is in pilot stage of production. ICE enables authors to edit content for print, web and CD delivery using a similar method to GOOD except authors use a word processor to write content.Hide Abstract
Speakers: USQ data integration web services and its Moodle applications Stijn Dekeyser, Leigh Brookshaw, Penny de Byl, Expand Stijn Dekeyser  (Mathematics and Computing, USQ) Leigh Brookshaw  (Mathematics and Computing, USQ) Penny de Byl  (Mathematics and Computing, USQ) Simplify D109, Thursday 1 December, 2005, 11:00am to 12:00noon Computing its_web_services.pdf 40KB (Speaker slides) Show Abstract In this workshop we will demonstrate web services that are being implemented by MC and ITS staff to enable programmatic access to data managed by the Peoplesoft database. We will very briefly talk about the theory behind web services, but focus mainly on a How-To presentation. The goal is that you come away knowing how you can optimally make use of this exciting breakthrough. Finally we also present how these web services will be integrated in the Moodle modules being developed in our department.Hide Abstract
Speaker: Intelligent Gene Regulatory Networks - an integration of Microarray and Ontological knowledge Vijayarengan Ramanujachary  (Masters Student, Mathematics and Computing, USQ) D109, Monday 21 November, 2005, 11:00am to 12:00noon Bioinformatics Show Abstract A large volume of data and information about genes and gene products has been stored in various molecular biology databases. A major challenge for knowledge discovery using these databases is to identify related genes and gene products in disparate databases . With the wide availability of the gene expression data induced towards the modelling of the Gene Regulatory Network. The proposed model is one such that integrates Gene Regulatory Networks with integration of the microarray data sets and Gene Ontological (GO) knowledge. This model retrieves the microarray data set (particular environment or disease) and it maps it with the Gene Ontological knowledge. The mapping and integration of the microarray data and the ontological knowledge are performed with the powerful computational methods and outputs are expressed as Gene Regulatory Network with powerful.Hide Abstract
Speaker: Natural aerosols: their impact on regional Australia Harry Butler  (Mathematics and Computing, USQ) D109, Wednesday 26 October, 2005, 12:00noon to 1:00pm Other Show Abstract The industrial age has seen an increase in atmospheric aerosols. While most of this increase in aerosols has occurred as a result of industrial processes, corresponding changes in farming and land management has also resulted in an increase in natural aerosols, such as dust and smoke. This increase in natural aerosols in Australia is illustrated by the severe dust and smoke events of the 1980's and the 1990's. These events affected large areas of Australia and New Zealand. This seminar discusses the impact that these aerosols have on: health, the environment and the economy of regional Australia. It also discusses the role monitoring and modelling plays in managing these natural aerosols in regional environments.Hide Abstract
Speaker: The DustWatch project Craig Strong  (DustWatch Northern Coordinator, Griffith University, ) D109, Friday 21 October, 2005, 1:00pm to 2:00pm Other Show Abstract Did you know that Australian dust can contain up to 35% organic material? Incredible considering that, on average, Australian soils have an organic component of less than 1%! Loss of these fine nutrient-rich organic (and inorganic) components by wind erosion makes dust more than just a nuisance; it is a serious land degradation issue. Researchers at Griffith University and NSW Department of Natural Resources have been using Bureau of Meteorology data to understand how different areas of the continent are susceptible to wind erosion. The Dust Storm Index (DSI) map (see attached) highlights the dusty hotspots. Limiting though is the patchy distribution and poor representation of stations in the semi-arid regions. In an attempt to address this, DustWatch, an Australia-wide network of volunteer observers has been set up to improve the understanding of wind erosion. This talk will explore how DustWatch can complement existing government monitoring programs to provide improved estimates of the extent, severity and impact of dust storms, and increase community awarness about duststorms and wind erosion. The 2002 - 2003 season is used as a case study. The great thing about DustWatch is that participation is easy, the dust will come to you!Hide Abstract
Speaker: Can Quality of Service be achieved in the Internet by better queue management in routers? Ron Addie  (Mathematics and Computing, USQ) D109, Monday 12 September, 2005, 12:00noon to 1:00pm Computing Show Abstract For some time now researchers have sought after an architecture and associated protocols which can provide statistically guaranteed quality of service in the Internet. Two very complex architectures for this purpose, IntServ and DiffServ, have been proposed, but have not achieved significant penetration. One reason for this lack of practical success is that the Internet does quite well without either of these quite complex systems. Another active area of research which overlaps considerably with the DiffServ initiative is Active Queue Management (AQM). Active Queue Management is attractive because it can be implemented in each router independently. No complex architecture is required to achieve its results. However, the achievements of AQMs up to now have been quite modest. In this seminar, which reports, to some degree, on a paper recently accepted for publication at TENCON, to be held in Melbourne later this year, a new AQM is proposed which has the potential to achieve statistically guaranteed quality of service in the Internet without a network wide change to the Internet architecture.Hide Abstract
Speaker: Modelling resonant subcritical transitions to turbulence Brett Richards  (Student, Mathematics and Computing, USQ, ) D109, Monday 15 August, 2005, 12:00noon to 1:00pm Mathematics Show Abstract Mechanisms which cause a subcritical transition of smooth, regular fluid flow to turbulence are important factors in designing efficient fluid transport systems. This is particularly true for practical applications such as pipelines or flow mixing devices. In this project we use an amplitude-expansion method to address the nature of localized transitions to turbulence that occur for mostly laminar flow through a wide rectangular channel. In particular, we extend previous techniques to examine the problem of mathematical resonance caused by interacting perturbations. Additional solution terms provide adjustments to a recognisable system of linear equations which we solve to find the mean flow distortion and higher order disturbance harmonics. We hope that computational results will justify the appropriateness of our model based on existing experimental data and will not leave us high and dry.Hide Abstract
Speaker: Attractors in nonlinear diffusion (with application to turbulence) Dmitry Strunin  (Mathematics and Computing, USQ) D109, Monday 8 August, 2005, 12:00noon to 1:00pm Mathematics Show Abstract The aim of this talk is to report on the research I am doing in 2005 and to invite anyone who might be interested for collaboration. Diffusion processes are everywhere around us in nature and technology; they describe, for example, propagation of heat (diffusion of temperature). A well known feature of the diffusion with constant diffusion coefficient is the existence of an attractor - decaying Gaussian-shaped solution generated by an instantaneous heat source. The Gaussian attracts different regimes started by different initial conditions. In many situations, however, the diffusion coefficient is not constant but depends on a diffusing agent. In such cases an attractor still exists, which I illustrate by examples. More complicated are situations where there are more than one diffusing agent coupled with each other. An example, which I analyse here, is an expansion of a turbulent jet where the diffusion of the turbulent energy is coupled with the dynamics of the flow velocity. I find the respective attractor and discuss its connection with the centre manifolds theory.Hide Abstract
Speaker: Techniques to Improve Traffic Conditioning in Differentiated Services Networks Ardavan Naeini  (Mathematics and Computing, USQ) D109, Wednesday 3 August, 2005, 12:00noon to 12:30pm Computing Show Abstract Quality of Service (QoS) has been under discussion for nearly two decades. There are many factors involved in providing a reasonable level of performance to the users of the Internet, specifically the users of demanding applications such as multimedia and Voice over IP (VoIP). So far, there have been architectures such as Differentiated Services (DiffServ) and Integrated Services(IntServ) both proposed to improve the QoS on the Internet through prior reservations. In addition, due to the presence of IP in the TCP/IP suite protocol, the Internet is currently providing 'best-effort' delivery. To improve the QoS in the 'best-effort' delivery system of the Internet, several congestion control mechanisms have been proposed over the last decade. In this presentation, which is the first part of a two-series presentation, I am going to present an introduction to my Master of Computing (MComp) research project proposal, highlighting the fundamental requirements of delivering QoS. Thereafter, I will talk about some techniques based on AI which can be used to improve the traffic conditioning, and consequently, can lead to improvements in QoS on the DiffServ Network. In addition, I will introduce a bunch of Active Queue Management (AQM) mechanisms already proposed. To conclude my presentation I will introduce a few researchers who are working on the same topic in academia within Australia. Hide Abstract
Speaker: The ICE-EM school mathematics project Michael Evans  (Australian Mathematical Sciences Institute) K302, Wednesday 27 July, 2005, 1:00pm to 5:00pm Mathematics Show Abstract ICE-EM Mathematics is a new program being developed for students in Years 5 to 10 by the International Centre of Excellence for Education in Mathematics (ICE-EM). The program comprises teacher professional development, a series of textbooks, multimedia materials and continuing teacher support. It will be completed over three years. ICE-EM is funded by the Australian Government through the Department of Education, Science and Training and is managed by the Australian Mathematical Sciences Institute (a consortium of 25 university mathematics departments, CSIRO Mathematical and Information Sciences, the Australian Bureau of Statistics and the Australian Mathematics Trust). The experienced ICE-EM development team is working with teachers, mathematicians and curriculum experts who provide detailed feedback as materials are produced. Each year-level program will be piloted in all types of schools across Australia for an entire school year. Teachers involved in the pilot will provide regular feedback to the development team, and the final program will reflect this collaboration.Hide Abstract
Speaker: Performance issues for Voice Over Internet Protocol (VOIP) in Networks Jishu das Gupta  (Student, Mathematics and Computing, USQ, ) D109, Thursday 7 July, 2005, 12:00noon to 1:00pm Computing Show Abstract There is a general consensus that the Quality of Service (QoS) of Voice Over Internet Protocol (VOIP) is of growing importance for research and study. The impact of VOIP performance in access networks is particularly important in regions where Internet resources are limited and the cost of improving these resources is prohibitive. We investigate the contribution of the access network to the overall performance of VOIP services and the ways in which access networks can be designed to improve VOIP performance. Issues of interest include the choice of coding rate, dynamic variation of coding rate, packet length, methods of controlling echo, and the use of Active Queue Management (AQM) in Access Network routers. Methods for analysing the impact of the access network on VOIP performance will be surveyed and reviewed. Also, we consider some approaches for improving performance of VOIP by doing some experiments using NS2 simulation software with a view for gaining a better understanding of design of access networks.Hide Abstract
Speaker: Power-variance models for modelling monthly and annual Australian rainfall (Mathematics and Computing, USQ) D109, Thursday 30 June, 2005, 12:00noon to 1:00pm Statistics Show Abstract Modelling rainfall presents difficulties: one is that rainfall is both continuous and discrete. The discrete component corresponds to exactly zero rainfall. Some researchers circumvent this by using two models---one for determining the presence and absence of rainfall, another for the rainfall amount. Here, we use power-variance (Tweedie) generalized linear models, which can explictly model continuous data with exact zeros. We demonstrate: - there is a physical basis for using these models; - that the parameters, in some cases, lend themselves to a useful interpretation; and - show the models fit the data well using monthly rainfall data from Charleville. We then model Australian annual rainfall, and develop contour maps for the power-variance model index parameter defining which member of the power-variance family is best used. We then briefly discuss further research using the presented ideas. The bulk of this presentation has been accepted as a refereed paper at the forthcoming International Workshop on Statistical Modelling, to be held in Sydney in July, and is joint work with Dr Neil White (Department of Primary Industries and Fisheries, Toowoomba).Hide Abstract
Speaker: An Infrastructure for Service Authentication and Authorisation Revocation in a Dynamic Aggregate of Networks David Lai  (Mathematics and Computing, USQ) D109, Wednesday 29 June, 2005, 12:30pm to 1:00pm Computing Show Abstract According to the 2005 Computer Crime and Security Survey conducted by Computer Security Institute and FBI, the average loss per year due to unauthorised access to information is US$303,234.00 per respondent of the survey. It reflects the importance of computer security, especially authentication, authorisation and revocation of access rights. For distributed networks systems without a global authentication service, authentication of users from another autonomous network is a major security issue in service sharing. In this presentation, we will present our work on Network Service Sharing Infrastructure (NSSI) by which many networks are linked together for service sharing. Within NSSI, individual networks authenticate and grant authorisations independent of each other by using their own authentication information repository. NSSI enables authentication and authorisation results to be relayed to other linked networks to access a shared service while individual networks still maintain their own authentication scheme or authentication requirements. NSSI facilitates dynamic aggregation of networks for service sharing with minimum administrative overhead.Hide Abstract Speaker: Developing Sustainable Strategies to Perfect Chinese E-Commerce System Zhen Wang (Student, Mathematics and Computing, USQ) D109, Wednesday 22 June, 2005, 12:30pm to 1:00pm Computing Show Abstract This presentation outlines a study in the E-Commerce framework currently undergoing in China. Its objective is to probe sustainable and effective strategies to perfect E-Commerce Systems within the context of China. This study will closely analyse the macro environment of E-Commerce in China; it identifies E-Commerce models, techniques, security, application platform, standards, certification, target markets and consumers etc, then analyses the gap between the informationisation, barriers, bottlenecks, and opportunities of the Chinese E-Commerce system. Among the topics of facilitating development of a positive, effective and satisfied E-Commerce system, "trust" becomes an essential but concerned issue, which is not easy to be addressed technically in short term. Therefore, with a systematic analysis on Chinese E-Commerce activities, the research will then focus on the concept of trust and techniques of fostering a trust relationship between consumers and merchants. It is envisaged that the study may be extended to assess the influence of government policy and national law on the level of a whole country.Hide Abstract Speaker: Understanding and Analysing AQM Workings Using Java Graphical Simulation Shen Chong (Student, Mathematics and Computing, USQ) D109, Wednesday 22 June, 2005, 12:00noon to 12:30pm Computing Show Abstract Traditionally, the concepts and protocols of computer communications and internetworking are illustrated by lecturers using text, images and videos. But the concept of Internet routing and some advanced strategies of queue management such as AQM (Active Queen Management) are characterised by complex, dynamic and concurrent processes. Thus, teaching and learning in this area have been challenging and extremely difficult to both lecturers and students. In recent years, computer simulation has become an important tool in assisting the understanding of computer network and testing the new protocols. This research aims at providing an interactive and animated learning environment that can significantly improve the understanding and retention of computer network and data communications. In particular, the objective of this research is to assist both lecturers and students on understanding the queue management such as AQM workings. The approach is based on the advanced Java technology such as graphic library and multithreading. Expected outcomes from this research include a fully animated and interactive environment that simulates the dynamic process of buffer management of the router and a set of tools for analysing the network performance. Hide Abstract Speaker: How do you judge a catch? The mechanics and the perception (Mathematics and Computing, USQ) D109, Wednesday 15 June, 2005, 12:00noon to 1:00pm Mathematics Show Abstract When a ball is hit towards you, how do you know which way to run in order to be there when the ball arrives? The initial speed of the ball is difficult to see, the ball is greatly affected by air drag, and it is almost always spinning. Previous published studies of this problem have been rather naive. In this talk I develop some models which incorporate effects of the drag crisis and spin and their influence on the fielder's perception. I even conclude with some coaching tips!Hide Abstract Speaker: Virtual plants and leaf surface modelling Birgit Loch (Mathematics and Computing, USQ) D109, Wednesday 1 June, 2005, 12:00noon to 1:00pm Mathematics Show Abstract This talk will begin with an introduction to virtual plants. Virtual plants are developmental plant models that incorporate topological and geometrical information, which can be used to produce a visualisation. They are not just computer-generated images, but computer models of the structural dynamics of individual plants in space. Virtual plants are tools for plant scientists and teachers in biology, ecology and pest management. I will then go into more detail for one of my research interests - the generation of leaf surface models. Leaves play a vital role in the development of a plant, as they are major resource collectors. Models of leaves may be used for visualisation purposes only, or they may also incorporate leaf function such as photosynthesis. While the modelling of plant architecture has been researched extensively over the last decades, models of leaf surfaces have mostly been generated with low accuracy or level of detail, and have often been handcrafted. I will explain techniques for the creation of detailed, accurate models of leaf surfaces. Such models may be used as parts of virtual plants for applications in fields as diverse as the arts, agriculture or computer games. These techniques are mathematical methods of surface fitting based on data that has been sampled from real leaves. Data collection, surface fitting and - time permitting - subsequent boundary improvement of resulting surface models will be addressed.Hide Abstract Speaker: Modelling precipitation using generalised estimating equations Taryn Axelsen (Mathematics and Computing, USQ) D109, Thursday 12 May, 2005, 12:00noon to 1:00pm Statistics Show Abstract Precipitation has a major influence on virtually all human activity and thus the quantification of this erratic phenomena is very important. My honours project will model precipitation data using generalised estimating equations (GEE). Generalised estimating equations are used to incorporate the non-independent structure of precipitation. The Tweedie distribution has been used in the past to model both the occurrence and amount of precipitation simulataneously. It is hoped that GEEs can be used, together with the Tweedie distribution, to create a model for precipitation that models both the occurrence and amount of precipitation, whilst taking into account that the data is dependent.Hide Abstract Speaker: Compound solitons and compactons in systems with large-scale dispersion Yury Stepanyants (Australian Nuclear Science and Technology Organisation, Sydney) T119, Friday 29 April, 2005, 12:00noon to 1:00pm Mathematics Show Abstract As is known, the wave-breaking phenomenon caused by nonlinearity can be averted either by small-scale dissipation, described by Bateman-Burgers equation, or small-scale dispersions, described by Korteweg-de Vries=20 (KdV) equation. In both these cases, steady-state solutions in the form of shock waves, periodic cnoidal waves, or solitons appear as a result of interplay between nonlinearity and dissipation/dispersion. In this talk, after an extensive tutorial introduction, I will discuss whether the large-scale dispersion can prevent wave-breaking. The situations where the large-scale dispersion is important in physical systems are widely spread. The corresponding model equation, the reduced Ostrovsky equation, has the same universality as the classical KdV=20 equation. It was derived for many applications including waves in oceans, nonlinear acoustics, plasma physics, relaxing media, media with random parameters, etc. Detailed analysis of solitary and periodic stationary solutions to this equation will be presented in the talk. The relationship of the reduced Ostrovsky equation with its full version combining quadratic nonlinearity with two types of dispersion, small-scale and large-scale, will be also discussed. An interesting specific feature of these equations, the zero-mass integral, will be pointed out, and its effect on the solution structure will be studied. Different types of compound solitons, spikons (sharp-crest solitons with infinite derivatives), loop solitons, compactons (solitons on a finite support), and periodic waves will be presented and classified.Hide Abstract Speaker: XML undeniable signature Li-li Sun (Student, Mathematics and Computing, USQ, ) T119, Friday 22 April, 2005, 12:00noon to 1:00pm Computing Show Abstract XML signatures are designed for the security of XML document transactions. They are involved in the aspects of authentication, data integrity and support for non-repudiation to the data they sign. After introducing the key XML concepts, in this talk, we will present a new signature method for XML based on RSA undeniable signature technique.Hide Abstract Speaker: Electronic Peer Review: A large cohort teaching themselves? Michael de Raadt (Mathematics and Computing, USQ) T120, Wednesday 16 March, 2005, 11:00am to 12:00noon Computing ElectronicPeerReview.ppt 180KB (Speaker slides) Show Abstract Foundation computing is one of the largest courses in the University. A system of electronic submission and peer reviewing with instructor moderation is now being used in this course. Why such a system is being used?How is it implemented?What are the benefits for instructors and students?How do students rate the system?Hide Abstract Show References Anderson, G & Boud, D. (1996) Extending the role of peer learning in university courses, Research and Development in Higher Education, 19, 15-19. Bloom, B.S (Ed) (1056) Taxonomy of Educational Objectives Handbook 1: Cognitive Domain. New York: Longman, Green & Co. Brook, C & Oliver, R (2003) Online learning communities: Investigating a design framework. Australian Journal of Educational Technology, 19(2), 139-160. Davies, R & Berrow, T. (1998) An evaluation of the use of computer supported peer review for developing higher level skills. Computers Educ, 30(1/2), 111-115. Griffiths, S, Houston, K & Lazenbatt, A. (1995) Enhancing Student Learning Through Peer Tutoring in Higher Education. Coleraine: Educational Development Unit, University of Ulster. Johnson, S. & Aragon, S. (2003). An Instructional Strategy Framework for Online Learning Environments, New Directions for Adult and Continuing Education, 100, Winter 2003, Wiley Periodicals, Inc. Kurhila, J, Miettinen, M, Nokelainen, P, Floreen, P & Tirri, H. (2003) Peer-to-Peer Learning with Open-Ended Writable Web, Proceedings of the 8th Annual Conference on Innovation and Technology in Computer Science Education, 173-178, ACM Press. McLoughlin, C & Luca, J. (2001) Quality in Online Delivery: What Does it Mean for Assessment in E-Learning Environments? Proceedings of the 18th Annual Conference of the Australian Society for Computers in Learning in Tertiary Education, 417-427. Prins, F.J, Sluijsmans, D.M.A, Kirschner, P.A &Strijbos, J. (2005) Formative peer assessment in a CSCL environment: a case study, Assessment & Evaluation in Higher Education, 30(4), 417-444. Saunders, D. (1992) Peer tutoring in higher education. Studies in Higher Education, 17(2), 211-218. Venkatesh, V., Morris, M., Davis, G. & Davis F. (2003). User Acceptance of Information Technology: Toward a Unified View, MIS Quarterly, 27(3), 425-478.Hide References Speaker: Stereotyping of emotions for use in creating believable artificial intelligences Billy Yue (Student, Mathematics and Computing, USQ, ) D109, Thursday 10 March, 2005, 11:00am to 12:00noon Computing Show Abstract One field in Artificial Intelligence (AI) is the study and construction of intelligent agents. In virtual reality type domains where these agents are called upon to model human beings, a major problem has been structuring and processing their knowledge in a way that makes them more believable. To this end, stereotyping principles have been investigated and used as a way to perceive and process emotions in an intelligent agent. The agent's attitude towards its environment, and entities within its environment, are used to generate emotional states, which in turn generate the agent's behaviour. Emotions produce believable characters that appear to think, make decisions, and act on their own will, to produce a suspension of disbelief.Hide Abstract Speaker: The Akaroa2 Simulation Package Don McNickle (Management Science, Management Department, University of Canterbury, New Zealand, ) D109, Thursday 29 July, 2004, 12:00noon to 1:00pm Computing Show Abstract One approach to parallel simulation is to launch multiple copies of a simulation program on a network of cheap, available processors - "multiple replications in parallel" (MRIP). The Akaroa2 Automatic Parallel Simulation Manager (named after a town on Banks Peninsula founded by French settlers in 1840) is an example of this approach. As well as efficiency gains - often achieving linear speed-up of simulation times using processors that otherwise would be idle - an added attraction of the MRIP approach is that it lends itself to increasing automation and user-friendliness, so that users with limited understanding of the theory of simulation analysis can use the package successfully. Since a master control module is already required to launch the simulation copies it is straightforward to build in modules that carry out the analysis of simulation output, and control the simulation experiment. To get statistically credible results, especially for steady-state simulations of complex systems, the unpredictability of the required run length means that sequential simulation, where the simulation runs until a pre-specified level of precision of the results is reached, is the only practical method. Thus much of our research has concentrated on finding methods of output analysis which work well under both the MRIP and sequential simulation scenarios. Hide Abstract Speaker: A comparative study of herd restrictions due to bovine tuberculosis in an area of high badger removal with an area of low badger removal in four counties in Ireland Gabrielle Kelly (School of Mathematical Sciences, University College Dublin, Ireland, ) D109, Wednesday 21 July, 2004, 12:00noon to 1:00pm Statistics Show Abstract The existence of a Mycobacterium bovis-infected population of badgers in Ireland has been linked as a source of infection for the cattle population in Ireland. A study was commissioned by the Department of Agriculture and undertaken over a five year period 1997-2002 which involved removal of badgers from four geographical areas. We examine the effect of this on the cattle and compare with control areas. We will look at the study design and some of the survival analyses carried out on the data.Hide Abstract Speaker: Some Aspects of Correlations in Queues Don McNickle (Management Science, Management Department, University of Canterbury, New Zealand, ) D109, Thursday 15 July, 2004, 12:00noon to 1:00pm Computing Show Abstract It can be a mistake to talk about two loosely related topics in a seminar: the change from one to the other tends to lose the audience. Here I will talk about three topics, which have a theme of correlated processes in common. 1. Product form networks (AJackson@ networks in their simplest form) possess the surprising property that in steady state the numbers of customers at various nodes are independent. Part of the reason for this relates to time reversal - if you show a movie of a Jackson network backwards you see another Jackson network Yet clearly, even in these systems there are dependencies - Aupstream@ nodes influence Adownstream@ nodes at a later point in time.First we look at these time-lagged correlations, calculating them by the Randomization Method. 2. Then we look at some processes in simple finite non-reversible systems - just overflow and output processes. Since these can be modelled as finite state Markov renewal processes the cross-correlations between flows can be calculated by inverting Laplace transforms using Maple. 3. Finally, since even these simple systems are getting too hard to model we will consider some general approximations based on fitting the moments of counting processes, which often work surprisingly well. A method of direct approximation for the expected waiting time in the queue is extended to these cases.Hide Abstract Speaker: A review of XML data management techniques Rodney Topor (School of Computing and Information Technology, Griffith University, ) H102, Tuesday 29 June, 2004, 12:00noon to 1:00pm Computing Show Abstract Information on the Web is increasingly being stored and transferred in the form of XML documents. Given such large volumes of data in this form, researchers have been applying database theory and technology to problems of design, storage, and access to XML data. In this seminar, I will review recent research intodesign and normalisation of XML schemas,approaches to representing and storing XML document collections, anddesign and implementation of query languages for accessing XML data.Hide Abstract Speaker: Implementation of a Mice and Elephants Queuing discipline on Linux Stephen Braithwaite (PhD Student, Mathematics and Computing, USQ, ) D109, Thursday 24 June, 2004, 12:00noon to 1:00pm Computing Show Abstract A Mice and Elephants strategy is one that favors short traffic flows over long traffic flows. Packets arriving at a router are classified according to the sum of bytes or packets in the flow that the packet belongs to. Packets belonging to short flows (mice) are given priority over packets that belong to long flows (elephants). It has been shown in mathematical models and in simulations that a "Mice and Elephants" strategy can produce substantial benefits in situations where there is congestion over a link at the edge of the Internet. This is a proposal to create a router that uses the "Mice and Elephants" strategy so that the strategy can be tested out in real situations. I aim to create this router by obtaining a suitable source code copy of the Linux Kernel and modifying the code that implements the queuing of packets. The resulting router will be a prototype version of a router for the purpose of evaluating a "Mice and Elephants" strategy.Hide Abstract Speaker: Classroom performance system (CPS) Cameron Craig (McGraw-Hill Aust Pty Ltd) D303, Thursday 17 June, 2004, 12:00noon to 1:00pm Other Show Abstract Cameron Craig's role is responsibility for innovation and e-learning initiatives in Australia, New Zealand and Asia. McGraw-Hill Australia is pleased to partner eInstruction Pty Ltd (www.einstruction.com.au) in providing the Classroom Performance System to students and instructors. This partnership follows great success in the U.S. where McGraw-Hill Higher Education and eInstruction have been offering CPS at a relatively low cost with a McGraw-Hill text for use in a Higher Education course. Adopting CPS provides unparalleled interactivity for students and is a powerful resource for instructors. What is the Classroom Performance System: CPS utilises a wireless handset, a simple receiver unit, and proprietary software. Using a standard projector, instructors can use the CPS software to run a series of questions during the lecture where the questions are displayed on screen and students respond with their handset. Responses are received via the receiver unit(s) and the results are calculated on the instructor’s computer and a histogram displayed on screen. CPS responses can be tracked by instructors as each student uses an individual handset where a unique serial number is registered by students and each response is recorded into a grade book. It is a great way of receiving immediate feedback! It encourages student attendance and engagement in lectures, it can be used to promote discussion, give and grade quizzes and track attendance. Over 100 institutions world wide are using CPS and over 400 using similar systems. In Australia, Deakin University, University of Ballarat, La trobe University, Edith Cowan University, University of Western Australia and University of Western Sydney have confirmed usage of CPS for Semester 2 , 2004. Hide Abstract Speaker: Enhancing the IMS Q&TI Specification by adding support for parameterised questions Damien Clark (School of Computer Science, Central Queensland University, ) Y392, Tuesday 15 June, 2004, 11:00am to 12:00noon Computing Show Abstract In recent years, the use of electronic online quiz systems has become quite popular due to pressures of increasing class sizes, and the need for more efficient methods of assessing distance students. One of the main benefits of using multiple choice questions (MCQ) with online quiz systems is automated marking. However, to help address the problems of cheating and to allow students the opportunity to resit quiz tests, it is necessary to develop a large number of MCQ items from which a random selection is made. This ensures that each quiz taken is sufficiently different. To develop good quality MCQs takes considerable time; therefore, what time is saved in marking is lost in development of quality MCQs. To offset this overhead, a method can be used whereby certain elements of the question are parameterised. Randomly selected option(s) can be inserted to alter the details of the question, while still testing the same content area. Creating one parameterised question can result in a large or infinite number of different quiz questions at low cost. Independent or collaborating researchers have developed many custom built online quiz systems. Each uses a different data format to represent quiz questions, and results in poor or no interoperability. Questions developed for one system cannot be re-used in another. To stem the flow of incompatible custom built systems, the IMS Global Learning Consortium have developed an IMS specification for describing questions and tests called The Question and Test Interoperability Specification. Institutions and vendors are now implementing their systems to support this specification such that interoperability can be achieved of tests between different quiz testing products. The IMS Question & Test Interoperability Specification is quite complete and is extensible to allow for functionality that does not fit within the specification guidelines. However, this is at the cost of breaking interoperability. An enhanced version of the IMS Q&TI Specification has been developed to support the interoperable storage of parameterised quiz questions. Using this enhanced specification, ten multiple choice and fill in the blank questions from a range of subject areas have been defined which each implement parameterisation. To demonstrate these ten parameterised questions, a prototype implementation of a quiz system has been developed, which supports the enhanced specification.Hide Abstract Speaker: NetXML - a language for describing and manipulating networks Abdulla Zareer (Student, Mathematics and Computing, USQ) D109, Friday 14 May, 2004, 12:00noon to 1:00pm Computing Show Abstract Technical descriptions of networks for the purpose of simulation, graphical representation, management, design, and so on, are currently stored in a variety of special data formats. In this project, an XML language for network data has been developed and used for two sample applications. The applications are: (1) An algorithm for computing the end-to-end availability of a network (a measure of its reliability), and (2) Depiction of a network as a graphic displaying various features selectively by means of coloured graphics, with informative labels.Hide Abstract Speaker: Streamlining the USQ Network: Quantum simulation of the USQ WAN link Shanthi Joseph (Student, Mathematics and Computing, USQ) T110, Thursday 13 May, 2004, 1:00pm to 2:00pm Computing Show Abstract The Poisson Pareto Burst Process (PPBP) is widely accepted as a simple and accurate model of Internet traffic. In this project, the PPBP and Lindley's equation have been used to model the USQ WAN link. The model is used to study the performance of the WAN link under varying conditions of traffic and resource allocation. The performance metrics include buffer overflow and bandwidth utilisation. The traffic conditions such as rate of arrival of new bursts and the duration of bursts are modelled using PPBP parameters. The network resource allocation is modelled using link capacity and queue capacity parameters. Analytical techniques and real traffic data have also been used in the analysis of the model. Quantum simulation techniques have been investigated to effectively study rare events such as network congestion. In quantum simulation, multiple simulation processes are used, each incorporating independent random events. Simulation processes that do not generate favourable events are replaced with processes that create favourable events. As a result, favourable events occur more frequently and may be studied in more detail. A generic software framework to implement quantum simulation in a parallel computing environment has been developed and tested in the APAC National Facility. The project has demonstrated the potential use of simulation to be a valuable tool in network analysis, design, provisioning and maintenance. Hide Abstract Speaker: External Education at The Open Polytechnic of New Zealand John Green (The Open Polytechnic, Wellington, New Zealand, ) D109, Thursday 22 April, 2004, 12:00noon to 1:00pm Computing Show Abstract John Green presents a brief overview of pedagogical issues, and techniques used for external education, at The Open Polytechnic of New Zealand. John is a Senior Lecturer in the Open Polytechnic’s School of Information Sciences and Humanities. John is at present a visiting fellow in the Department of Mathematics and Computing with Michael de Raadt, looking at our methods of dealing with external students. John is keen to meet with anyone wishing to discuss such issues. The Open Polytechnic is the leading specialist provider of open and distance learning at tertiary level in New Zealand and the largest tertiary institution in terms of student numbers, with currently over 30,000 enrolments annually. The Polytechnic celebrates its 60th Anniversary in 2006. Students learn either via printed material with online peer and tutor support or through fully online courses supported by Nexted. During his presentation John will also touch on collaborative research with Dr Zlatko Kovacic on learning style profiles. Hide Abstract Speaker: Generic Gesture Recognition using Key Frames and Motion Change Analysis Benedict Radclyffe (Student, Mathematics and Computing, USQ) Y392, Thursday 8 April, 2004, 12:00noon to 1:00pm Computing Show Abstract Human gestures can be divided into a sequence of significant poses with transitional movement. Previous research explores the characteristics of gestures, and gesture recognition methods. Work has also been completed that attempts to model human motion using key frames and a simplified skeletal structure. A method for domain independent gesture recognition using key frames and constraints on motion will be proposed as an alternative to the traditional methods that concentrate on the use of neural networks.Hide Abstract Speaker: Efficient XPath Axis Evaluation for DOM Data Structures Jan Hidders (University of Antwerp, Belgium, ) D109, Thursday 4 March, 2004, 11:00am to 12:00noon Computing Show Abstract We propose algorithms for implementing the axes for element nodes in XPath given a DOM-like representation of the document. Each algorithm assumes an input list that is sorted in document order and duplicate-free and returns a sorted and duplicate-free list of the result of following a certain axis from the nodes in the input list. The time complexity of all presented algorithms is at most O(1 + m) where 1 is the size of the input list and m the size of the output list. This improves upon results of others: they presented algorithms with linear time complexity also, but these are linear in the size of the entire document whereas our algorithms are linear in the size of the intermediate results which are often much smaller.Hide Abstract Speaker: Satisfiability of XPath Expressions Jan Hidders (University of Antwerp, Belgium, ) D109, Thursday 26 February, 2004, 11:00am to 12:00noon Computing Show Abstract We investigate the complexity of deciding the satisfiability of XPath 2.0 expressions, i.e., whether there is an XML document for which their result is nonempty. Several fragments that allow certain types of expressions are classified as either in PTIME or NP-hard to see which type of expression make this a hard problem. Finally, we establish a link between XPath expressions and partial tree descriptions which are studied in computational linguistics.Hide Abstract Speakers: Difficulties of Simulating Internet Traffic Ron Addie, Zhi Li, Expand Ron Addie (Mathematics and Computing, USQ) Zhi Li (Student, Mathematics and Computing, USQ, ) Simplify D109, Tuesday 24 February, 2004, 12:00noon to 1:00pm Computing Show Abstract Simulations of TCP/IP protocols usually adopt an arbitrary mix of offered traffic, e.g. so many ftp sessions and so many http sessions. Then the behaviour of the network is studied. Is this because the concept of a "normal" stationary pattern of traffic is not applicable to TCP/IP networks? Maybe not! Another explanation is that there is a normal stationary pattern of TCP/IP traffic but it is too difficult to simulate. Hence we are forced to select arbitrary traffic configurations, none of which typical of the whole range of possibilities. We will describe a traffic model for TCP/IP traffic which has broad acceptance, and explain why simulating this model is very difficult. We will then consider:how we might overcome the difficulties of simulating a network carrying this traffic;whether it is important to simulate a stationary "normal" model of traffic, or are "arbitrary traffic configurations" satisfactory?Hide Abstract Speaker: Some Statistical Behaviors on Mobile Agents for Network Routing Hong Shen (Japan Advanced Institute of Science and Technology (JAIST), ) D109, Sunday 22 February, 2004, 10:00am to 11:00am Computing Show Abstract Deployment of mobile agents in network-based applications has attracted lots of attentions recently. How to control the activities of agents is crucial for effective application of mobile agents. In this talk I will focus on the application of mobile agents in network routing. First I will introduce our mathematical models for agent-based network routing. I will then show the analysis of two important activity properties of mobile agents: the probability of success in finding the destination and the distribution of mobile agents running in the network. Our results show that the number of mobile agents in the network can be controlled by adjusting the number of agents generated per request and the number of jumps each mobile agent can move. Thus, we can improve network performance by tuning relevant parameters locally. This reveals new theoretical insight to the statistical behaviors of mobile agents and provide useful tools for effectively managing mobile agents for routing in large networks.Hide Abstract Speaker: Bayesian Inference with WinBUGS and R Daniel Burrell (Genetic Solutions; PhD student at University of Newcastle) D109, Friday 20 February, 2004, 11:00am to 12:00noon Statistics Show Abstract The Bayesian interpretation of probability has long been considered too dificult an approach to be widely applicable. The reasons for this stem from the many and varied complications associated with conditional probability distributions and the difficulties that so often arise in evaluating directly, multivariate integrals of not so 'well-behave' functions. However, with the advent of faster and more powerful computers, and the continuing development of Markov-chain Monte Carlo (MCMC) methods, the Bayesian approach has become a much more viable option in many statistical contexts. The WinBUGS software encorporates the high-level BUGS language into a Windows GUI to provide a relatively easy way to develop and run parameter simulations for Bayesian statistical models, allowing Bayesian inferences to be made about parameters of interest given specific observed data. WinBUGS includes a command-line interface for those users who prefer the Unix/Linux feel of its predecessor BUGS, as well as a graphical modelling tool known as DoodleBUGS. Of itself, WinBUGS is good for the beginning Bayesian analyst, but is a bit of a 'black box' to those who wish to do more technical analysis. However, Rubin et al. (2003) have developed an R package: R2WinBUGS, that allows WinBUGS to communicate with the far more general and powerful statistical and graphical language R. I will give a brief tutorial showing how WinBUGS works and what it is capable of, followed by a demonstration of how to make WinBUGS communicate with R, leading to much greater flexibility and increased functionality. Also, I will discuss how to utilise the R language directly to perform Bayesian inference, providing the analyst with complete control in a Bayesian modelling context. Hide Abstract Speaker: Coupling Algorithms for Multi-Physics Problems Hermann G Matthies (Institute of Scientific Computing, Technical University Braunschweig, Germany, ) D109, Friday 13 February, 2004, 11:00am to 12:00noon Mathematics Show Abstract Multi-Physics Problems---here in particular the emphasis will be on fluid-structure coupling---will be approached from the point of seeing them as separate subsystems which have to be coupled in the proper way. This way, each subsystem still can be analysed separately, can be discretised or otherwise approximated separately, and, last but not least, existing software to solve the subproblems may be reused as part of a larger system. Assuming that each of the subproblems describes the evolution of the state of that subsystems (usually through a differential equation in some appropriate Banach space), the coupling conditions will usually introduce "algebraic" constraints, i.e. they only involve the states of the subsystems, but may very well be differential operators in space. The complete problem is then an abstract differential algebraic system of equations (DAE). When implicit time-stepping methods are used to solve the DAE, coupled equations involving all subsystems have to be solved, but only solvers for the subsystems are available. With the aid of the subsystem-solvers the global equation is re-formulated in fixed-point form. Frequently used approaches are non-linear block Jacobi and block Gauss-Seidel. These methods are simple back and forth iterations, but they are relatively slow and they occasionally fail. An explanation of this occasional failure is given, as well as a Newton-like method which does not suffer from such difficulties, and only uses the subsystem solvers. Hide Abstract Speaker: Tractor Guidance Systems: A case study of generalised least squares (GLS) and linear mixed-effect models (LMEs) Andrew Powierski (Student, Mathematics and Computing, USQ, ) D109, Friday 13 February, 2004, 11:00am to 12:00noon Statistics Show Abstract Tractor guidance systems are a popular tool used by farmers to keep the tractor and the implement being towed at a level of acceptable performance. Two types of tractor guidance systems are: "NovAtel (Beeline)" and "Trimble", both of which are tracked using the Global Positioning System (GPS). Three experiments have been performed in an attempt to compare these two types of receivers with the manufacturer's specifications. In this analysis, we have longitudinal data, which is correlated data. Classical regression techniques cannot be applied to this type of data due to the failure of assumptions which are required to perform regression analysis. The statistical analysis on the data sets will be performed using both GLS and LME models, both of which account for correlated data. This presentation is based upon my Honours project proposal and I will outline the analysis I plan to perform.Hide Abstract Speaker: Part 2: Expansion of turbulence in shear stratified flow: degenerate self-similarity Dmitry Strunin (Mathematics and Computing, USQ) D109, Wednesday 11 February, 2004, 11:00am to 12:00noon Mathematics Show Abstract We present numerical and analytical analysis of the evolution of a turbulent zone between two layers with different densities (light layer is on top of heavy layer) moving with different velocities. Pumped up by the shear the turbulence works against the buoyancy forces. As time elapses the profile of the turbulent energy becomes self-similar before decaying to zero in finite time. It seems that the self-similarity cannot be classified as one of two known types of self-similarity.Hide Abstract Speaker: Expansion of turbulence in shear stratified flow: interesting case of self-similarity Dmitry Strunin (Mathematics and Computing, USQ) D109, Thursday 22 January, 2004, 11:00am to 12:00noon Mathematics Show Abstract We present numerical and analytical analysis of the evolution of a turbulent zone between two layers with different densities (light layer is on top of heavy layer) moving with different velocities. Pumped up by the shear the turbulence works against the buoyancy forces. As time elapses the profile of the turbulent energy becomes self-similar before decaying to zero in finite time. It seems that the self-similarity cannot be classified as one of two known types of self-similarity.Hide Abstract Speaker: Conflict Scheduling of Transactions on XML Documents Stijn Dekeyser (Mathematics and Computing, USQ) D109, Thursday 15 January, 2004, 11:00am to 12:00noon Computing Show Abstract In the last few years an interest in native XML databases has surfaced. With other researchers we argue that such databases need their own provisions for concurrency control since traditional methods are inadequate to capture the complicated update-behavior that is possible for XML documents. Ideally, updates should not be limited to entire document trees, but should involve subtrees and even individual elements. Providing a suitable scheduling algorithm for semistructured data can significantly improve collaborative systems that store their data --- e.g. Word processing documents or vector graphics --- as XML documents. In this presentation I will give an overview of the work we have done in Antwerp during the last two years. The talk will be more or less a copy of my talk that I will give in Dunedin (Australasian Database Conference) next week.Hide Abstract Speaker: Embedding 'key skills' in undergraduate mathematical sciences programmes Ken Houston (University of Ulster, Northern Ireland, ) D109, Friday 12 December, 2003, 11:00am to 12:00noon Mathematics Show Abstract I will consider a rationale for teaching "key skills" and for embedding them in an undergraduate programme. I will illustrate possible ways of doing this in mathematical sciences programmes by describing experiences at the University of Ulster and elsewhere in the UK.Hide Abstract Speaker: Introducing An Object Oriented Network Simulation Environment in C++ - A test bed for Network Protocols David Lai (Mathematics and Computing, USQ) D109, Friday 5 December, 2003, 11:00am to 12:00noon Computing Show Abstract In this presentation, Omnet++ is introduced. A typical client server network is used as an example to illustrate how to set up a simulation. We will then set up input queue(s) for a router as an example of using modules as building blocks for an investigation.Hide Abstract Speaker: Asymptotic accuracy in queueing theory Ron Addie (Mathematics and Computing, USQ) D109, Thursday 4 December, 2003, 11:00am to 11:30am Computing Show Abstract Recently results have shown that a single server queue Poisson Pareto Burst Process input (which is a realistic model for most traffic carried by the Internet today)has a tail which is bounded by hyperbolic functions. We show that the hyperbolic upper and lower bounds for this system can be very misleading, that this hyperbolic tail result is relevant only from a certain threshold onwards, and the magnitude of this threshold may be very large. We also show that any hyperbolic upper and lower bounds for a tail of the stationary waiting time complementary distribution necessarily become further apart as the rate of the process increases. The importance of this result is that the character of the queueing behaviour of systems which carry Internet traffic is quite different from what has been up to now widely accepted. In fact, behaviour is determined by a slowly moving process of the long bursts. This insight potentially can be exploited to better control Internet traffic. This talk will be presented at the Australian Telecommunication, Networking and Applications Conference which will be held next week in Melbourne. Hide Abstract Speaker: Improving the adaptability of AQM algorithms to traffic loads using Fuzzy Logic Zhi Li (Student, Mathematics and Computing, USQ, ) D109, Thursday 27 November, 2003, 11:00am to 12:00noon Computing Show Abstract Adaptive RED (ARED) aims to keep average queue length around a predefined queue length. If so, average delay will be offered to Internet users and simultaneously high link utilisation will be achieved. However, ARED algorithm merely makes average queue length loosely converge to the target length with low responsiveness and low stability. One reason for this is ARED uses a fixed increase step-size to adjust the maximum dropping probability to the current traffic load. In this paper, fuzzy logic (FL) controller has been designed to adaptively obtain the increase step-size of the maximum dropping probability based on the current traffic conditions. The simulation results show the proposed FL scheme improves the ARED's adaptability to the current traffic loads, consequently achieving the desired target queue length with better performance than ARED, while keeping very similar drop rate and link utilization.Hide Abstract Speaker: Formal Authorisation Approaches for Role-Based Access Control Based on Relational Algebra Operations Hua Wang (Mathematics and Computing, USQ) D109, Thursday 27 November, 2003, 11:00am to 12:00noon Computing Show Abstract With people's increased consciousness the need for electronic commerce to protect their private information and security of applications, system administrators are continuing to implement access control mechanisms as well as to retain a critical and complex aspect of security administration. We develop formal authorisation allocation algorithms for role-based access control (RBAC). The formal approaches are based on relational structure, and relational algebra and operations. The process of user-role assignments is an important issue in RBAC because it may modify the authorisation level or imply high-level confidential information to be derived while users change positions and request different roles. There are two types of problems which may arise in user-role assignment. One is related to authorisation granting process. When a role is granted to a user, this role may be conflict with other roles of the user or together with this role; the user may have or derive a high level of authority. Another is related to authorisation revocation. When a role is revoked from a user, the user may still have the role from other roles. To solve the problems, we present an authorisation granting algorithm, and weak revocation and strong revocation algorithms that are based on relational algebra. The algorithms can be used to check conflicts and therefore to help allocate the roles without compromising the security in RBAC. We describe how to use the new algorithms with an anonymity scalable payment scheme. Finally, comparisons with other related work are discussed. Hide Abstract Speaker: There must be a better way to mix application output and text in study materials (Mathematics and Computing, USQ) D109, Thursday 20 November, 2003, 11:00am to 12:00noon Statistics Show Abstract In the Department of Mathematics and Computing (and no doubt other Departments), it is often necessary to create Study Materials centred around some computer application. This may be explicit (such as actually teaching an application) or implicit (as is common in statistics and mathematics, where R, MATLAB, Scilab, etc may be used to demonstrate statistical or mathematical principles). Commonly, it is necessary to create the study materials and incorporate the application language and output in the text. It may also be necessary to include tables, pictures, results, and so on that are outputs from the program. Generally, this is achieved by running the programs, capturing the output, and placing this output in the document. Pictures can be created, saved, and introduced into the text. The code that created this output and graphics can be stored and re-used if necessary. This is fine; but if something changes (such as a new version of the software is released, or a data file changes or is updated, etc.), the whole process must be repeated: figures re-created and re-saved; the computer code altered, and so on. There must be a better way. There is; come along to this seminar and discover the better way to incorporate application output into study materials, and bypass these problems. This seminar demonstrates the principles in one specific context (using LaTeX and R) as used in the preparation of one set of study materials for 2003, but we hope it challenges others to extend the idea to other contexts. Hide Abstract Speaker: Direct Interesting Rule Generation Jiuyong Li (Mathematics and Computing, USQ) C204, Wednesday 5 November, 2003, 11:00am to 12:00noon Computing Show Abstract An association rule generation algorithm usually generates too many rules including a lot of uninteresting ones. Many interestingness criteria are proposed to prune those uninteresting rules. However, they work in a post-pruning process and hence do not improve the rule generation efficiency. In this paper, we discuss properties of informative rule set and conclude that the informative rule set includes all interesting rules measured by many commonly used interestingness criteria, and that rules excluded by the informative rule set are forwardly prunable, i.e. they can be removed in the rule generation process instead of post pruning. Based on these properties, we propose a Direct Interesting rule Generation algorithm, DIG, to directly generate interesting rules defined by any of 12 interestingness criteria discussed in this paper. We further show experimentally that DIG is faster and uses less memory than Apriori.Hide Abstract Speaker: Detection of Plagiarism in Computer Programming Assignments (Mathematics and Computing, USQ) D109, Wednesday 29 October, 2003, 11:00am to 12:00noon Computing Show Abstract Plagiarism at Universities is becoming an increasingly more common problem and it is not only in written assignments, but also within programming assignments. The plagiarism within programming assignments is more difficult to detect as programs can be changed in numerous ways, e.g. changing comments, identifiers, functions, the constructs used and the structure of the program. There have been numerous efforts made to produce detection programs which are not only capable of detecting even the most subtle of changes, but also flexible enough to be used for more then one programming language. One of the major aims of this study is to investigate the current state and produce a set of benchmarking tests that can be applied to any plagiarism detection software to give an unbiased comparison to others tested.Hide Abstract Speaker: Transition to Turbulence in Pipes and Channels Nick Trefethen (Oxford University and University of Queensland) D109, Thursday 23 October, 2003, 12:00noon to 1:00pm Mathematics Show Abstract High Reynolds number fluid flows are turbulent in practice, yet classical linear stability analysis predicts they should be laminar. This paradox remained largely unresolved for much of the 20th century. This talk will describe the explanation of subcritical transition to turbulence that emerged in the 1990s, which owed much to numerical simulations and to fundamental ideas of numerical linear algebra. The "essence of the matter" can be conveyed by a simple ODE model containing just two or three variables.Hide Abstract Speaker: Web of Knowledge Demonstration Sue Dowe (Faculty Librarian, Engineering & Surveying, Maths & Computing, USQ) D109, Thursday 9 October, 2003, 12:00noon to 1:00pm Other Show Abstract The live demonstration will begin with an explanation of what ISI Web of Knowledge is and where to find it on the Library Home Page. Then it will cover:ISI Web of Science (included in Web of Knowledge):How many and which journals does it cover?How to find relevant articles on topics;How to examine a document's historical origin (the cited references) and easily follow links to its current position in the research literature (times cited);How to discover who is citing a particular author's research, and how it is being used to support current research.ISI Current Contents (also included in Web of Knowledge):How to view, browse and search the complete tables of journals and books in your area of interest and within your selected date range; How to link to full-text online journal articles How to link to evaluated Web sites in all disciplines.Hide Abstract Speaker: Independence properties for Wishart matrices Jacek Wesolowski (Student, Mathematics and Computing, USQ) D109, Thursday 18 September, 2003, 12:00noon to 1:00pm Mathematics Speaker: Graph visualization: towards practical applications Xiaodi Huang (Dept. Mathematics and Computing, USQ) D109, Thursday 11 September, 2003, 12:00noon to 1:00pm Computing Show Abstract Many classical graph Layout algorithms have been developed during past decades. However, these algorithms face difficulties in a practical context, such as the overlapping node problem, large graph layout and dynamic graph layout. This talk will report some of research results from my ongoing PhD studies. First, a framework for graph visualization is proposed. A new approach to removing overlapping nodes, called Force-Transfer, is then developed to greatly improve the Force Scan algorithm, which is a classical algorithm. Third, two approaches are presented for large graph layout: filtering and clustering. In order to filter _noise_ or less important nodes in a given graph, an approach called NodeRank is proposed to compute importance scores of nodes, and then control appearances of nodes in the layout by adjusting a threshold. Fourth, two methods are given to clustering graphs: KNN and NSKM. After identification of seed nodes as initial members of clusters, the K-nearest neighbor search, and a novel Node Similarity Matrix, are employed to seek a group of nodes with most affinities or similarities, and then group them to form abstract nodes. Finally, a MMD approach is introduced to layout the clustered graphs. A prototype called PGA was developed to demonstrate the proposed algorithms and approaches.Hide Abstract Speaker: Highlights of the 2003 Data Mining Winter School at Monash University Khaleel Petrus (Mathematics and Computing, USQ) D109, Thursday 7 August, 2003, 12:00noon to 1:00pm Computing Show Abstract Data mining practices and experiences will be explained through practical examples. Then, some theory and algorithms will be explained. These include Rule discovery, Neural Networks, Bayesian Networks and regression models. Tasks of data mining are described. Then, data warehousing will be explained. CSIRO practical experiences are presented. Finally, future research challeges are presented.Hide Abstract Speaker: Multiple and negative target rule mining Jason Jones (Student, Mathematics and Computing, USQ) D109, Thursday 31 July, 2003, 12:00noon to 1:00pm Computing Show Abstract Data Mining is a flourishing and relatively new field of Computer Science whereby we gather knowledge from information. Rule mining is a popular data mining method whereby we discover relationships and express them linguistically. Clearly, high accuracy of a set of rules under application is crucial, and is an important area of research. In this presentation, I will briefly discuss algorithms developed to date, and introduce a new approach which aims at increasing the accuracy of rule based predictions.Hide Abstract Speaker: Palm Size Online Games Sunny Koh (Student, Mathematics and Computing, USQ) D109, Wednesday 30 July, 2003, 12:00noon to 1:00pm Computing Show Abstract With a total sales of 12,066,375 units in 2002 (Gartner Dataquest January 2003), PDAs are becoming more and more commonplace. According to NPD FunWorld_ and NPD TechWorld_, the Video and Computer industry is worth$6.35 Billion US Dollars in 2001. A convergence of the 2 markets have created a demand for PDA based Games. Big names like Sony, Microsoft and Electronic Arts are pursuing this overlap market with products. Most of these shipping at the moment are either single player or have limited multiplayer capabilities. With the advance of Palmtop Technologies, the drive towards a Massive Multiplayer Online Game for the PDA is on. An MMOG is an online persistent virtual world populated by thousands of players and non-player characters. The world is persistent in that players' characters or attributes develop over time and the world exists and grows whether the players are online participating in it or not. Players generally participate in it by paying a monthly fee. With a huge potential market for PDA based MMOGs, is this technical possible with current technologies? What would be the limitations and considerations in the design and development of such products?Hide Abstract
Speaker: Use of Wavelets for Modelling and Simulation of Multi-Scale Phenomena Oleg V Vasilyev  (Department of Mechanical Engineering, University of Colorado, Boulder, ) D109, Monday 30 June, 2003, 12:00noon to 1:00pm Mathematics Show Abstract Today there are a number of problems in engineering and science, which share a single common computational challenge: the ability to solve and/or model accurately and efficiently a wide range of spatial and temporal scales. Different scales are often not distributed uniformly in space and time and have complex nonlinear dynamics due to different physical feedback mechanisms. In this talk we present a general framework for both modelling and numerical simulations of complex multi-scale phenomena that takes advantage of multi-resolution wavelet analysis, a new mathematical concept, which allows one to represent a function in terms of basis functions, called wavelets. Wavelets are localized in both space and scale, and as a result functions with localized regions of sharp transition are well compressed using wavelet decomposition. This property allows us to construct an efficient adaptive numerical method, which employs wavelet compression as an integral part of the solution. The adaptation is achieved by retaining only those wavelets whose coefficients are greater than an a priory prescribed threshold. This property of the multilevel wavelet approximation allows local grid refinement up to an arbitrary small scale without a drastic increase in the number of grid points; thus high resolution computations are carried out only in those regions where sharp transitions occur. The prowess and computational efficiency of the adaptive wavelet collocation method are demonstrated for the solution of a number of test problems including both compressible and incompressible Navier-Stokes equations. The results indicate that the computational grid and associated wavelets can very efficiently adapt to the local irregularities of the solution. Furthermore, a solution is obtained on a near optimal grid, i.e. the compression of the solution is performed dynamically as opposed to a posteriori as it is done in data analysis. The efficiency of the adaptive wavelet collocation method can be greatly enhanced by combining it with the recently developed Coherent Vortex Simulation (CVS) approach, which is closely related to the standard Large Eddy Simulation (LES) method. In contrast to LES, in which the velocity field is decomposed into large- and small-scale fields, in CVS the velocity field is decomposed into coherent and incoherent components that represent the most energetic structures in the solution and the small scale quasi-random residual field respectively. This decomposition is achieved by using non-linear wavelet de-noising strategies. With this new decomposition one can follow the evolution of the coherent structures, while the effect of the incoherent components is modeled stochastically. To achieve further computational efficiency we developed a new approach, called Stochastic Coherent Adaptive Large Eddy Simulation (SCALES). The details of this approach and its relationship to CVS and LES are discussed. Preliminary analysis for homogeneous turbulence is given and potential applications of the new approach are discussed. Hide Abstract
Speaker: Emotional Stereotyping using Conceptual Graphs for Believable Agents Billy Yue  (Student, Mathematics and Computing, USQ, ) D109, Thursday 12 June, 2003, 12:00noon to 1:00pm Computing Show Abstract One field in AI (Artificial Intelligence) is the study and construction of intelligent agents. Intelligent agents are a key issue in computer game development. They are used to create computer controlled characters known as non-player characters (NPCs). As the acceptance by the player of a computer game can depend on their suspension of disbelief, NPCs should behave in a believable way. Implementing emotional concepts in an NPC could make the agent appear more believable to the player. The study proposed in this presentation uses stereotyping principles as a way to perceive and process emotions in an intelligent agent. The agent's attitude toward its environment and entities within its environment are used to generate emotional states which in turn generate the agent's behaviour.Hide Abstract
Speaker: Detection of Fraud and Inappropriate Practice Warwick Graco  (HIC Canberra, ) D109, Thursday 29 May, 2003, 12:00noon to 1:00pm Computing Show Abstract One of the major challenges facing society is the increasing cost of white-collar crime. Recent examples include collapse of Enron and Worldcom and the dot.com scams on Wall Street. This presentation gives an overview of issues involved in discovering and detecting fraud, abuse and other forms of regulatory non-compliance. The first part of the presentation covers issues to do with how new frauds and abuses are identified while the second part deals with issues about how regulatory breaches are detected using smart technology. Examples involving health fraud and abuse are used to illustrate issues.Hide Abstract
Speaker: Improving the Deviance Measure in Generalized Linear Models using Quantile Residuals Daniel Burrell  (Genetic Solutions; PhD student at University of Newcastle) D109, Thursday 15 May, 2003, 12:00noon to 1:00pm Statistics Show Abstract In 1972, Nelder and Wedderburn introduced the generalized linear modelling framework to bring together what was previously a group of distinct techniques for dealing with specific types of non-normal data. Along with this framework, they introduced a generalization of the analysis of variance, which they called the analysis of deviance. The analysis of deviance forms the base technique for model selection within the GLM framework, but is only valid under certain regularity conditions, since it is derived as a first-order asymptotic result. That is to say, the distributional properties of the analysis of deviance statistic are only approximate in as much as they are asymptotically convergent in distribution. Many have tried to strengthen the deviance measure with limited success. In 1996 Dunn and Smyth introduced the notion of quantile residuals, resulting in a working definition of a residual under the GLM framework, which exhibits, under ideal conditions, exact normality. It seems apparent, that with a slight alteration to the quantile residual definition, an improvement to the standard definition of the deviance can be made. It is theorized that such a deviance measure, defined in terms of quantile residuals, will exhibit much more accurate results. The exact normality of the residuals under ideal conditions, will allow for a likelihood-ratio type statistic to be developed as a deviance measure, which will necessarily be distributed according to an exact chi-square distribution. As a result, this 'quantile deviance' will allow for much stronger inference and hence, model selection. My proposed dissertation will focus on developing the quantile deviance measure and testing its relevance and usefulness. A number of potential problems are foreseen, and will be discussed and hopefully resolved. Hide Abstract
Speaker: Prediction of shear forces due to flood flow over sloped planes Ruth Mossad  (Faculty of Engineering and Surveying, USQ, ) D109, Thursday 1 May, 2003, 12:00noon to 1:00pm Computing Show Abstract Flood flow over sloped farmland is a cause of major concern to farmers. The floodwater imposes forces that cause erosion of the valuable topsoil. Understanding the different parameters that affect these forces can help in evaluating different ways of controlling the floodwater to reduce the soil erosion. To analytically predict these forces, a solution of the momentum equations for turbulent free surface flow, which include the effect of viscous and gravitational forces, and the continuity equation, is needed. In this work, a computational fluid dynamics (CFD) software called Fluent is used to solve these equations. This software uses a control volume based finite difference method. The parameters that affect the shear forces are Froude number, plane slope, length, and roughness. The effect of these parameters on the shear forces on the plane is investigated. The results show that plane length affects the solution particularly further along the plane. The results also showed that, apart from the early stages where the flow is developing, the local values of the shear stress over the plane increased until they reached a maximum before levelling out. It was also found that the higher the plane slope the higher the shear stresses and the sooner the levelling out occurred. Roughness, as expected, increased the shear stress over the plane and caused the levelling out to happen earlier.Hide Abstract
Speaker: Efficient Rule Discovery And Robust Rule-Based Predictions Jiuyong Li  (Mathematics and Computing, USQ) D109, Thursday 24 April, 2003, 12:00noon to 1:00pm Computing Show Abstract Data mining is a boosting research area due to the increasing popularity of dealing with a large number of data in various fields. The rule is one of the most expressive and human readable representations for knowledge, and hence rule mining is one of the central tasks in data mining. Rules are mainly used for the explanation and prediction. Association rules are very popular for the explanation due to that efficient algorithms are available for generating rules with high support. However, rules with low support play an important role in the prediction. Most association rule discovery algorithms are inefficient with low support. In this talk, we will discuss efficient association rule discovery algorithms for the predictive purpose, and how to use an association rule set to make robust predictions.Hide Abstract
Speaker: Training Strategic Problem Solvers Michael de Raadt  (Mathematics and Computing, USQ) D109, Thursday 10 April, 2003, 12:00noon to 1:00pm Computing Show Abstract When an expert programmer begins to create a piece of code which will solve a problem, they will usually apply knowledge that they have developed in the past through solving problems of a similar nature to the one that presents itself currently. A guarded loop here, an initialisation there, some testing and a function is complete. For a novice, who may understand the rules of syntax that apply to a programming language, and may have applied such rules in several examples, when such an example presents itself, they are lost without the canned solutions possessed by an expert. Can we give students expert knowledge for problem solving without years of experience?Hide Abstract
Speaker: A Statistical Approach to Rainfall Modelling Sarah Lennox  (Student, Mathematics and Computing, USQ) D109, Thursday 20 March, 2003, 11:00am to 12:00noon Statistics Show Abstract An ability to accurately or appropriately model rainfall promises numerous practical applications. The varied and multiple applications of rainfall modelling demands accurate and reliable models. There are numerous difficulties in the modelling of rainfall which make the finding of an appropriate model very difficult. This talk will touch on some of the more important difficulties. The talk will also focus on modelling the rainfall process by using generalised linear models, in particular the Tweedie distributions. Two separate models are usually required for modelling rainfall processes, a rainfall occurrence and a rainfall amounts (>0mm) model. The family of Tweedie distributions allows for rainfall amounts of exactly zero and can therefore model rainfall processes in one combined model. Modelling rainfall amounts greater than or equal to 0mm has not been performed in one model before, so the Tweedie distributions will be fitted and compared with previous models to determine if a viable avenue exists for further use of these models in rainfall modelling. The Tweedie distributions also allow for distributions more skewed than the gamma distribution and therefore alternative models will also be addressed for rainfall amounts >0mm. The suitability and predictability of some models will be discussed and will show that the Tweedie models improve upon previous models used for monthly and daily rainfall amounts.Hide Abstract
Speaker: Guaranteed performance in the Internet (cont) Ron Addie  (Mathematics and Computing, USQ) D109, Wednesday 19 March, 2003, 11:00am to 12:00noon Computing Show Abstract In the seminar last week, the DiffServ protocol was described in connection with the provision of guaranteed quality of service in the Internet. DiffServ is not a protocol in isolation. In order to (potentially) provide guaranteed quality of service, several other techniques will also be necessary: 1. Service Level Agreements; Agreements which define the service to be provided; 2. Admission Control; A potential barrier for an individual service, to make sure it fits the SLA; 3. RSVP; A protocol for reserving bandwidth on a path through the Internet; 4. Policing and Conditioning; and Mechanisms for dropping or marking packets which do not fit the agreed service profile. 5. Multi-Protocol Label Switching (MPLS). An architecture by which the Internet can make use of tunnels which use a different switching technology. Most of these were discussed in last week's seminar, to some extent. This week, we will consider these topics in a little more detail, although only in relation to the overall issue of guaranteed quality of service in the Internet.Hide Abstract
Speaker: Guaranteed performance in the Internet: Is it possible? Is it necessary Is it available already? Ron Addie  (Mathematics and Computing, USQ) D109, Wednesday 12 March, 2003, 11:00am to 12:00noon Computing Show Abstract Multiservice networkingA quick reviewThe Broadband ISDN and Asynchronous Transfer ModeThe Internet and the ascendancy of "best effort services"DiffServ - an architecture for providing different performance for different types of packets in the InternetRSVP - a protocol for reserving bandwidth in the InternetMPLS - Multi-protocol label switching - a protocol whereby several different types of switching technologies can work together, using the most efficient type of switching at each node along the path of a packet.How DiffServ/RSVP/MPLS can perhaps be used together to provide guaranteed performance in the Internet.Hide Abstract
Speaker: Blind Signal Separation: The cocktail party problem Yan Li  (Mathematics and Computing, USQ) D109, Thursday 27 February, 2003, 11:00am to 12:00noon Computing Show Abstract Blind signal separation is the process of extracting unknown source signals from sensor measurements which are unknown combinations of the source signals. The cocktail party problem is a classic example demonstrating the need for robust audio separation algorithms. In this seminar, we will introduce the blind signal separation problem and a theoretical approach - Independent Component Analysis. The performance of a novel blind signal separation algorithm using recurrent neural networks will be presented in the seminar.Hide Abstract
Speaker: Dynamic Password Scheme David Lai  (Mathematics and Computing, USQ) D109, Thursday 20 February, 2003, 11:00am to 12:00noon Computing Show Abstract The simplest and common password authentication protocol is Challenge and Response. It is free from Replay Attack but vulnerable to Dictionary Attack. Another family of password authentication protocol is the Encrypted Key Exchange EKE family. Both parties have to store some common shared secret. User has to stay with a particular workstation or terminal and is not suitable for mobile users. Again the Asymmetric Key Exchange family requires both parties to store some common shared secret. Again it is not suitable for mobile users although it involves less authentication steps than Encrypted Key Exchange. One Time Password is great for mobiles users but we must choose between lower performance or having to reinitialise the scheme frequently. In this paper, we present Dynamic Password Scheme (DPS), a direct password authentication scheme that resists Replay and Dictionary Attacks; portable for mobile users and take less steps in the authentication process. It can determine if spoofing has occurred or not. DPS is also resistant to Social Engineering.Hide Abstract
Speaker: The Tale Of The Misleading Tail Ron Addie  (Mathematics and Computing, USQ) D109, Thursday 23 January, 2003, 11:00am to 12:00noon Computing Show Abstract Recently, results based on Large Deviations Theory and other methods have shown that the M/G/infinity single server queue (SSQ) or its discrete time version the Poisson Pareto Burst Process (PPBP) has a tail which can be bounded above and below by hyperbolic functions. In this paper we show that the hyperbolic upper and lower bounds for the stationary waiting time in a PPBP queue can be very misleading. We show that this hyperbolic tail result is relevant only from a certain buffer threshold onwards, and the magnitude of this threshhold may be very large, and we also show that any hyperbolic upper and lower bounds for a tail of the stationary waiting time complementary distribution necessary diverge from each other as the rate of the PPBP input increases.Hide Abstract
Speaker: Intelligent Agents For Network Management Peter Rowlings  (Student, Mathematics and Computing, USQ) D109, Thursday 12 December, 2002, 11:00am to 12:00noon Computing Show Abstract As the number of networks within an organisation grows, along with the diversity of systems comprising this internet, for instance, routers from various vendors, hosts with embedded router functionality, terminal servers, etc, managing all these systems within a coherent framework becomes important. Intelligent agent technology can delegate some intentions of network management stations (or human manager) into the network elements. In this research, we explore how to apply the intelligent agent in the network management. The objectives of this project are to examine how much delegation the intelligent agent can do and under what context the intelligent agent will perform its delegation. In this presentation, we propose two techniques, decentralised polling strategy and reliable multicast. We also describe how to use these techniques in conjunction with intelligent agent technology for the network management.Hide Abstract
Speaker: Recovering Information From Plant Biology Data R.S. (Bob) Anderssen  (CSIRO Mathematical and Information Sciences, Canberra, ) D109, Wednesday 6 November, 2002, 12:00noon to 1:00pm Bioinformatics Show Abstract When recovering information from experimental (biological) measurements, one must take into account that such observational data will not only be perturbed by stochastic fluctuations but also be indirect measurements of some underlying phenomenon. (For example, when one weights samples in the laboratory, one actually measures the extension or compression of a spring that has been calibrated against standard weights.) Consequently, there are various statistical, algorithmic and modelling issues that must be addressed including 1. (statistical) the choice of the measurements and their model to best reflect the information being sought and subsequently utilized, 2. (algorithmic) the careful assessment, in terms of the application, of the available methodologies, and 3. (modelling) the stabilization of the recovery of information from the indirect measurements. This talk will illustrate the importance of these issues with three applications from wheat biology; namely, 1. (statistical) the assessment of wheat grain hardness, 2. (algorithmic) the estimation of the number of active genes in endosperm development in wheat, and 3. (modelling) the application of near infrared (NIR) spectroscopy in the prediction of wheat properties such as protein content and flour yield. The major emphasis of the talk is the role of mathematics and statistics in recovering information from experimental measurements such as plant biology data.Hide Abstract
Speaker: Unmasking The Theta Method Baki Billah  (Department of Econometrics and Business Statistics, Monash University, ) D109, Thursday 26 September, 2002, 12:00noon to 1:00pm Mathematics Show Abstract The "Theta method" of forecasting performed particularly well in the M3-competition and is therefore of interest to forecast practitioners. The original description of the method given by Assimakopoulos and Nikolopoulos (2000) involves several pages of algebraic manipulation. We show that the method can be expressed much more simply and that the forecasts obtained are equivalent to simple exponential smoothing with drift.Hide Abstract
Speaker: Stable Pulse As A Model For Elementary Particle (or What Partial Differential Equations Can Do) Dmitry Strunin  (Mathematics and Computing, USQ) D109, Thursday 5 September, 2002, 12:00noon to 1:00pm Mathematics Show Abstract Dr Strunin will give a brief overview of topics of the 5th International School/Conference on Chaos and Nonlinear Dynamics held at the Unviersity of Maribor, Slovenia, in July 2002. He will then present in more detail his contributed paper dedicated to a solitary pulse as a model for elementary particle. In contrast to classical quantum mechanics where the particle presents point in space (and therefore has no width), Dr Strunin's model gives stable particle with finite width. The model develops previous ideas of Sivashinsky to draw a parellel between combustion front equations and the classical Hamilton-Jacobi (HJ) equation for the particle. Dr Strunin will extend the HJ formalism to obtain the stable steady pulse.Hide Abstract
Speakers: Spatial and temporal variation of Australian continental dust (1960-1999) Adrian Chappell, Marie Ekstrom, Expand Adrian Chappell  (School of Environment and Life Sciences, University of Salford, Manchester, ) Marie Ekstrom  (Climatic Research Unit, University of East Anglia, UK, ) Simplify D109, Thursday 29 August, 2002, 12:00noon to 1:00pm Mathematics
Speaker: Association Rules - Algorithms and Applications Markus Hegland  (Mathematical Sciences Institute, Australian National University, ) T110, Thursday 15 August, 2002, 1:30pm to 2:30pm Mathematics Show Abstract An association rule is a triple consisting of an "if-then rule" together with 2 numbers which characterise how well the rule is supported by a dataset. Association rules belong to the most popular discovery tools in data mining. This popularity is partly due to the fact that efficient algorithms are available to find rules with high support and confidence. In this talk, we will review the data mining process, some challenges, applications and several algorithms for association rule discovery. Some familiarity with elementary probabilistic concepts is assumed. Hide Abstract
Speaker: Non-blocking Optical Multistage Interconnection Networks Hong Shen  (Japan Advanced Institute of Science and Technology (JAIST), ) D109, Wednesday 7 August, 2002, 12:00noon to 1:00pm Computing Show Abstract In this talk, Prof Hong Shen will introduce some recent works on design and analysis of non-blocking optical MINs. The talk will focus on MINs designed by applying the vertical stacking technique under the packing strategy and show some analytical results on their performance. Prof Hon Shen has published extensively in the areas of parallel computing, networking and algorithms, and served on the editorial board of several academic journals.Hide Abstract
Speaker: Quantum Simulation of WAN Traffic Shanthi Joseph  (Student, Mathematics and Computing, USQ) D109, Thursday 1 August, 2002, 12:00noon to 1:00pm Computing Show Abstract Internet traffic, LAN traffic, and variable bit-rate video traffic are all known to have the interesting features of long-range dependence and self-similarity. This means that the traces of such traffic shows burstiness at all time scales. Sometimes the term fractal traffic is used to describe this phenomenon. In this seminar, a project which will study the traffic from USQ to the Internet at large will be described. The traffic will be modelled and simulated. The theoretical model used will be the Poisson-Pareto Burst Process (PPBP) which is a natural model for this kind of traffic. Mathematical analysis of this model is difficult, so simulation is also often used. Because of long-range dependence and the heavy weight of the tails of the burst distribution, even simulation is very difficult for this model. For this reason, quantum simulation of this traffic will be used in this project. Quantum simulation is a technique whereby several simultaneous parallel simulations are conducted at once. In order to make best use of these parallel simulations, or threads, from time to time threads are cloned and thinned. Some work on a framework for quantum simulation will also be described. Hide Abstract
Speaker: Modelling Variation in Caesarean Section Rates for NSW Hospitals Paul Fahey  (Mathematics and Computing, USQ) D109, Saturday 20 July, 2002, 12:30pm to 1:00pm Statistics Show Abstract Using 1997/98 hospital data we found that caesarean section rates in NSW public hospitals vary between 0 and 29.8%. We hypothesise that the variation comes from 3 distinct sources: 1. Patient-based differences (such as maternal age and morbidities); 2. Hospital-based differences (such as hospital class and throughput); and 3. Random variation (with rates for smaller hospitals more likely to vary than rates for larger hospitals). Excluding hospitals with fewer than 25 births in the 1997/98 data, we sought to analyse 69837 records from 93 hospitals using a generalised linear mixed effects model. This paper presents our initial experiences and findings. Hide Abstract
Speaker: Generalised Linear Models for Continuous Data with Exact Zeros (Mathematics and Computing, USQ) D109, Saturday 20 July, 2002, 12:00noon to 12:30pm Statistics Show Abstract Well defined models exist for continuous data and for discrete data. Some data, however, is both continuous and discete. One way to analyse such data is using the Tweedie distributions. Using an example from the insurance industry and another from biology, this talk demonstrates the use of these models. The Tweedie densities form a family of exponential dispersion models (EDMs) with variances of the form var[Y] = \phi\mu^p for some index p not in (1, 2) which defines the distribution within the family, and where \phi > 0 is the dispersion parameter. Of particular interest are the distributions with 1 < p < 2, which are continuous for Y > 0 with a positive probability that P(Y=0). As EDMs, the Tweedie densities can be used with generalized linear models. However, any maximum likelihood estimation is very difficult since the density function and distribution function do not generally have a closed form. (The exceptions are p=0 (normal), p=1 (Poisson), p=2 (gamma) and p=3 (inverse Gaussian) cases.) In this talk, two methods for accurately computing the density are discussed, followed by discussion of a fast interpolation method based on the saddlepoint approximation. In addition, methods for computation of the distribution function are briefly considered. Using the methods, the two examples mentioned above are given demonstrating Tweedie generalized linear models, the maximum likelihood estimation of p and residual analysis. Hide Abstract
Speaker: Risk behaviour of the improved estimators of the mean vector of multivariate normal distribution under linex loss function Zahirul Hoque  (Mathematical and Physical Sciences, University of Newcastle, ) D109, Sunday 14 July, 2002, 12:00noon to 1:00pm Statistics Show Abstract This paper defines different estimators of the mean vector of multivariate normal distribution by using both the sample and non-sample prior information. The Bayes estimator has also been considered. The risk function of different alternative estimators with respect to linex loss function have been computed and analysed under different conditions. It has been observed that the unrestricted estimator is inadmissible compared to Bayes estimator Shrinkage estimator uniformly dominates the preliminary test estimator. But when the difference between the true value of the mean vector and that assigned by the null hypothesis is zero then the restricted estimator dominates both preliminary test and shrinkage estimators.Hide Abstract
Speakers: If you want it done right ...... Making the Web work for all of us Penny de Byl, , Richard Watson, Expand Penny de Byl  (Mathematics and Computing, USQ)   (Mathematics and Computing, USQ) Richard Watson  (Mathematics and Computing, USQ) Simplify D303, Tuesday 2 July, 2002, 12:00noon to 1:00pm Computing Show Abstract Prototype web tools will be displayed showing how the power of new technology can be put in the hands of academics.Hide Abstract
Speaker: Item Response Theory and Its Application to the Analysis of Introductory Statistics Examination Items Colin Carmichael  (Student, Mathematics and Computing, USQ) D109, Thursday 27 June, 2002, 12:00noon to 1:00pm Statistics Show Abstract In this talk, a brief review of item response theory from its roots in Classical Test Theory to present time will be presented. The mathematics involved in the implementation of the theory will be developed and in particular the use of the Rasch model and Birnbaum’s two-parameter logistic model. The results of the application of item response theory to the analysis of multiple-choice questions from four consecutive semesters of the introductory statistics unit Data Analysis will be discussed. A critical analysis of item response theory will then be presented, both from a mathematical and an educational perspective.Hide Abstract