The School of ISE Research Seminar Series
All Welcome: Refreshments Provided from 10:30 in staff room
Location: Fridays, 11.00-12:00pm, UC building 11, level B, room 24
Contact: Masoud.Mohammadian@canberra.edu.au
Introduction to Second Life and Virtual Worlds
Robert Cox
This is an informational seminar. Its goal is to introduce Second Life and virtual worlds.
This presentation will cover:
- What Second Life is, its features and what makes it different from other games;
- I will cover why it is a turning point in MMOG’s and why some believe it has a more general future as a UI for the web;
- I will also look/examine and propose research opportunities both for UC and the world in general;
- The levels of immersion that it gives;
- I will very briefly cover the emerging competitor products and some of the controversies surrounding second life; and
- Most Importantly I will have some screenshots and videos of SL in action so people can get a feel for what it is.
3 October
Artificial Intelligence Modelling of the Relationship Between Training and Performance in Athletes: a case study of elite cyclists
Tania Churchill
Training programs attempt to enhance cycling performance by manipulating the volume, frequency and intensity of training. The eternal question for coaches and sports scientists is how to design a training programme to elicit a maximal performance on a given day. Models have been developed to describe the relationship between training and performance, ultimately allowing prediction of performance from training input.
This PhD progress seminar will review attempts to use mathematical models to quantify the relationship between training and performance. The limitations to these existing systems modelling approaches will be discussed. The case for using different modelling techniques (e.g. Artificial Neural Networks, stochastic optimisation, and mixed models) will be presented, as well as some initial results from this research.
26 September
Sergey Sergeev
By “integrable”, or “solvable” systems a wide range of subjects is usually understood:
- Integrable systems of Newtonian Mechanics
- Integrable systems of Quantum Mechanics, like the Hydrogen atom
- Non-linear PDEs with sufficiently high hidden symmetry group like Darboux-type PDEs of differential geometry and soliton equations
- Classical difference equations
- Equations of discrete-differential geometry
- Spin chains -- Heisenberg magnet and various similar models
- Solvable lattice models of Statistical Mechanics
- Quantum groups
- Integrable quantum field theories
In my talk I will overview relations between these subjects and their conceptual integrity.
19 September
High Quality Speech Recordings via the Internet
Christoph Draxler
"Until recently, high quality speech recordings required that either a recording team visits the speakers, or that the speakers come to a recording studio. With speech recordings via the Internet, recordings can be performed in parallel, in high signal quality, and without any time-consuming and expensive travel.
In the Ph@ttSessionz project, more than 1000 pupils have been recorded in secondary schools in more than 40 locations in all dialect regions of Germany. A uniform signal quality was achieved by using the same equipment at every recording location. This equipment consists of a Beyerdynamic opus 45 close-talk microphone and an AudioTechnica AT
3031 condenser microphone connected to the PC via an M-Audio mobile pre USB A/D adapter. The signal quality is 22.05 kHz, 16 bit stereo.
The recording software is embedded into a web application and can be accessed via a normal Internet browser. During the recording session, the recorded signal files are uploaded to the server where they are immediately available for further processing, e.g. annotation.
The Ph@ttSessionz web application has now been generalized and extended to allow an easy adaptation to other speech database collection projects. Wherever there is an Internet connection, we can now record speech - in any language, in any quality!
29 August
Scoring System for Tennis with Time Limits
Ian Lisle
Apart from the professional tour, many tennis matches are played with a fixed limit on court time. For such time-limited fixtures, standard tennis scoring systems may lead to undesirable outcomes such as wasted court time as teams wait for changeover, or a high proportion of matches unfinished. In this talk we investigate possible modifications of the scoring system for the Sydney Badge competition. By running millions of simulations, a variety of scoring systems can be tried for effectiveness. One especially promising modification is presented in detail.
This is joint work with Professor Graham Pollard.
22 August
Observable Markov Modelling and Automated Feature Weighting-Based Subspace Methods for Cell Phase Classification
Dat Tran
Studies of drug effects on cancer cells are performed through measuring cell cycle progression such as interphase, prophase, metaphase and anaphase in individual cells. Such studies require the processing and analysis of huge amounts of image data. Manual image analysis is very time consuming thus costly, potentially inaccurate and poorly reproducible. Stages of an automated cellular imaging analysis consist of segmentation, feature extraction, and classification of individual cells in a dynamic cellular population. The segmentation stage was performed at Harvard medical School in which cell nuclei in different mitotic phases were recorded over a period of twenty-four hours at every fifteen minutes with a time-lapse fluorescence microscopy. In this seminar, we present observable Markov modelling and automated feature weighting-based subspace methods proposed for feature extraction and classification. We also present experimental results for the proposed methods.
15 August
CSI Canberra: experiences in offering Forensic Statistics online
Alice Richardson
The units Forensic Statistics (level 2) and Forensic Statistics G (level G) were offered at the University of Canberra for the first time in semester 1, 2008. To align with other units in Forensic Studies courses, the units were taught fully online. In this talk, we will describe the unit’s content and assessment, the results achieved by the first cohort of students, the experiences of the teaching staff during the semester, and our plans for semesters to come.
8 August
Wolfgang Polasek
We describe new approaches of making long-term forecasts in dynamic panel models of cross-sectional growth rates. Since the variance of the forecasts will depend on number of updating steps, we compare the forecasts behavior of a aggregated and a disaggregated updating procedure. We estimate spatial AR (SAR) or Durbin models allowing for heteroscedasticity. Since the forecasts are non-linear functions of the parameters MCMC based approach will produce the best results. We demonstrate the approach by a regional model where we predict 20 years ahead of regional growth in 99 Austrian regions in a space-time dependent system of equations.
Key words: Dynamic panel models, spatial econometrics, MCMC, regional forecasting.
1 August
Games, Sets and Matches: are we still teaching statistics?
Alice Richardson & Felicia Zhang
In this seminar we will report on the results of an experiment conducted in the unit Introduction to Statistics at the University of Canberra. A variety of strategies, referred to as games, sets and matches, were employed, many of which emanate from the teaching of foreign languages. These included personal strategies for the lecturer such as recording lectures, individual strategies for the students such as the use of Hot Potatoes software, and group strategies for the students such as vocabulary cards and in-class activities and discussions. The aim of the experiment was firstly to improve students’ use of statistical language, and secondly to see if we were indeed still teaching statistics, and improving overall student performance in the unit.
25 July
Software Team Formation Model for Cooperative Multi-agent Systems
Ebrahim AlHashel
The agent team formation process or what is commonly known as agent cooperation is the key advantageous feature of agent-oriented software (AOS) in comparison with all other software approaches. The Agent cooperation process has the potential to dynamically form a team of agents, and then execute a plan to achieve a common goal.
The purpose of this seminar is to demonstrate the progress of my research including the research in the field of Multi-agent systems’ cooperation concepts. The seminar will also illustrate why the agent-oriented software approach is essential to modern software engineering. Finally the seminar will explain the architecture of the proposed conceptual agent cooperation model (ACM).
6 June
Usability Modelling for Requirements Engineering
Sisira Adikari
For over two decades user-centric methods and techniques have been proposed to assist the production of usable, useful, and desirable software products. Despite these approaches, usability problems are still identified in finished software products creating problems for end users and systems acceptance testers and requiring expensive rework. Part of the reason for these continuing problems is that user-centric approaches are not part of the traditional software engineering process. The literature review shows that software engineering and human-computer interaction are largely different communities.
The aim of this research was to investigate whether the incorporation of user modelling and usability modelling into software requirements specifications would improving design quality and usability of software products. This research used a Design Science dominant mixed research methodology consisting of case study and action research for creating, analysing, and evaluating artefacts. Using the functional specification of an existing system in a government agency, 10 designers created screen and interaction designs. The specification was then enhanced with usability specifications and the designers redeveloped their designs in the light of the enhanced specification. Both designs were subject to pre-defined usability tests and designers described their design experience as they worked.
The results of the research demonstrated that enhancing traditional software requirements specifications with additional specification of user modelling and usability modelling made a positive difference to both designer perception as well as design quality of user interface artefacts. The theoretical and practical values of these findings are explored.
30 May
Majority Spanning Trees, Cotrees and their Applications
Mohammad Kaykobad
We define a new class of spanning trees, called majority spanning trees, and cotrees, called majority cotrees, in directed graphs. We prove their existence in any digraph with non-negative weights on edges. We prove that there exists a majority spanning tree T of G such that G-T is a majority cotree. We show its applications in scheduling period transportation system by criterion of minimizing waiting times in connections, and in settling multiple debts. We also devise an algorithm for ranking players of a round robin tournament using the same structure in tournament digraph.
23 May
Speaking with Different Voices: knowledge legitimation codes of mathematicians and mathematics educators
Steve Thornton
This seminar uses a textual analysis of two documents prepared by the mathematics community and the mathematics education research community to the National Numeracy Review in 2007 to uncover and compare knowledge legitimation within these two fields. The analysis shows that knowledge within these disciplines is based on different epistemic devices, and hence that debates surrounding mathematics education in schools arise, at least in part, from differing ways of viewing knowledge.
May 16
Integrating Automated Speech Recognition and Natural Language Processing
Dave Davies
When attempting to understand a speech utterance humans draw on knowledge from many sources: a diverse range of cues from the acoustic signal; lexical part-of-speech options; common word ordering or grammar; semantic and ontological information; the utterance context and expectations within a dialogue context. This talk will provide an overview of how the quantification and integration of such diverse cues can be achieved in automated recognition.
May 9
Knowledge-gap driven Continuing Medical Education (CME)
Peter Main
Today's medical practitioners are faced with a medical knowledge flood. This demands effective IT tools to ensure new knowledge can be rapidly found, learnt and used at the clinic. CME is now taking new directions worldwide, with electronic-CME showing rapid growth, however the effects of sponsor bias has generated serious concern.
Every medical research report creates knowledge-gaps that then propagate haphazardly through the clinical community. When this knowledge is applied, the experience acquired needs to filter back to researchers. Failure to deal with this can lead to events such as the Vioxx medication withdrawal. Computers can help bridge these bi-directional knowledge gaps.
Rare or unusual cases such as Toe-Tourniquet syndrome, or unusual vitamin B12 deficiency, need a computer-assisted, Self-Directed Learning paradigm (SDL), due to the large search-space problem. Testing some of the basic principles of such a framework is the focus of this pilot study.
May 2
Bala Balachandran
Agent-based computing is emerging as a powerful new paradigm that might be the cornerstone for the next generation of computing that seeks to address issues in Ambient Intelligence, Pervasive and Ubiquitous Computing, Grid Computing, Services Oriented Computing, the Semantic Web, E-Commerce and many other areas. However, one of the limiting factors in industry take up of agent technology is lack of adequate software engineering support, and knowledge in this area. This seminar reviews the current state of agent technology and discusses future challenges of this important area of research. The seminar will also outline some of the agent technology research projects that are currently under way.
April 4
Kim Le
This seminar is composed of two parts.
First, Kim will present an overview of some plenary lectures and other conference papers, including:
A) Category: soft computing
Toward Human-Level Machine Intelligence—Is it Achievable?
www.wseas.org/wseas-zadeh-2008.pdf
LotfiA. Zadeh, Computer Science Division Department of EECSUC BerkeleyWSEAS
B) Category: mathematics
Formalisation and Verification in a Type-theoretic in a Type-theoretic Frame work.
Zhaohui Luo, Dept of Computer Science, Royal Holloway, Univ. of London, http://www.cs.rhul.ac.uk/home/zhaohui/type.html
C) Category: soft computing and mathematics
Probability Measures of Fuzzy Events and Linguistic Fuzzy Modelling – Forms Expressing Randomness and Imprecision.
Anna Walaszek-Babiszewska, Dept of Control and Computer Engineering, Opole University of Technology, Poland.
In the second part, Kim will present his SEPADS conference paper and his continuing work toward an extended version for a WSEAS Journal paper, An Active Distributed Medical Advisory System, SEPADS Conference, Cambridge Univ., UK, 20-22 February 2008.
For the last few decades, medical practitioners and specialists have been waiting for efficient medical advisory systems. Most of current computer-aided diagnosis systems are rather passive. Health data, a rich source of medical knowledge, often have not been used efficiently for medical treatment as well as financial management. Their accesses are very limited. This paper introduces the design of a well-structured medical advisory system based on the advanced Internet. The system is also equipped with data mining and AI techniques to make it becoming an active distributed medical advisory system
21 March
