Study
For
About
See Also
Computer Policies
CSSE Summer Research Showcase
Over the 2009-10 summer a number of students have been carrying out research projects in the Department of Computer Science and Software Engineering and with our colleages in Human Interface Technology Lab NZ. Some of these have been supported by UC or CSSE Summer Scholarships and others are COSC366 research projects.
In order to celebrate the achievements of all our summer research students and to share them with staff and students, we will be holding a summer research showcase on Monday, 15th February. The showcase will take place in room 031 and will feature presentations in the style of an academic conference.
A BBQ will take place at the end of the showcase sessions.
Here is the draft timetable for the day. Updates will be posted as further detail comes to hand.| Time | Speaker | Title |
|---|---|---|
| 9:30 | Carl Cerecke | Yakyacc for all! |
| 9:50 | Chris Deaker | Tag Clouds for Information Visualisation |
| 10:10 | Myse Elmadani | Exploring concept tagging in Moodle |
| 10:30 | Break | |
| 11:00 | Marina Filipovic | A machine-learning decision support system for prescribing Warfarin |
| 11:20 | Constantin Naghi | |
| 11:40 | Ansar Ali | Deploying Intelligent Tutoring Systems in MOODLE |
| 12:00 | Break | |
| 1:00 | Colin Fagg | An OO Metrics Framework for JST |
| 1:20 | Josh Oosterman | Code Critick: Informing Software Development |
| 1:40 | Linda Pettigrew | Empirical Software Engineering |
| 2:00 | Break | |
| 2:30 | Paul Williams | An intelligent educational game platform |
| 2:50 | Martin van Zijl | Possible Uses for a Ball-Shaped Interface |
| 3:10 | Alexander Chernoglazov | Development of a Visualisation Infrastructure for High Performance Computing |
| 3:30 | BBQ |
Abstracts
- Title: Yakyacc for all!
- Speaker: Carl Cerecke
- Abstract:
- This talk will present the parser-generator test framework developed
for Yakyacc.
- Title: Tag Clouds for Information Visualisation
- Speaker: Chris Deaker
- Abstract:
- Tag clouds, or word clouds, have been used on various websites as a means of visualizing information. Clouds often represent simple statistics about sites and their users’ interactions, such as the frequency of occurrences of particular keywords or search terms. In these cases, the domain is simple and easily visualized. However, such clouds are severely limited when visualizing domains with an arbitrary number of possible measurements, such as those found in software evaluation. This presentation will discuss the background research and development process used to create a tool capable of representing more complex domains, and discusses possibilities for further development.
- Title: Exploring concept tagging in Moodle
- Speaker: Myse Elmadani
- Abstract:
- The Moodle learning management system is becoming increasingly widespread due to its adoption by education providers at all levels. The system however does not provide educators with the ability to create connections between online course material and domain concepts easily and visibly. We chose concept tagging as the method for allowing teachers to provide these links in a new concept tags plug-in for Moodle. Students and teachers will mutually benefit from the use of such a mechanism, as it would aid information retrieval from content.
- Title: A Machine-Learning Decision Support System for Prescribing Warfarin
- Speaker: Marina Filipovic
- Abstract:
- This research project deals with the problem of prescribing accurate dosage of Warfarin to heart-valve transplant patients. The main goal of this research is to build a prototype for web-based Warfarin Predictor. This prototype will need to allow for a doctor to have enough flexibility to create and combine different attributes in order to create the model that would predict warfarin intake of the patient with the greatest accuracy. By giving doctors enough flexibility to self create the input attributes and combine them in order to make a prediction model, it is hoped that it would be possible to find the combination of attributes and algorithm that would best suit individual patients. This could be particularly important when the patient does not have a long history of taking Warfarin and other patients' histories could be used as a measure. The other goal of this research is to conduct some experimental work once the web-based predictor is created to give some insight to what type of attributes would work best with an algorithm as it is expected that some combinations would not give any value while others could enhance the accuracy of Warfarin prediction.
- Title:
- Speaker: Constantin Naghi
- Abstract:
- Title: Deploying Intelligent Tutoring Systems in MOODLE
- Speaker: Ansar Ali
- Abstract:
- Intelligent Tutoring systems (ITS) are being used for training purposes since late 70's. Ideally an ITS manages to model individual student’s learning behaviors and uses different techniques to compare the domain’s knowledge with student knowledge. On the other hand, Learning Management Systems (LMSs) are also being used by universities and colleges to manage course-ware and impart education through the Internet. There is huge number of learners who are using these LMS. One of these LMS is MOODLE which is internet based and customizable. During "Summer Research Project" structure and working of ASPIRE (an ITS authoring and tutoring system) and MOODLE (a LMS) were reviewed thoroughly. ASPIRE is re-implemented as plugin of MOODLE. This plugin is capable of taking domain model in XML form. It presents problems dynamically selected fom given domain to a learner. It returns appropriate feedback when the learner submits the solution. On the basis of performance of the learner, this system models the learner's knowledge and learning behaviors.
- Title: An OO Metrics Framework for JST
- Speaker: Colin Fagg
- Abstract:
- Software metrics are a useful tool for evaluating quality of software. Unfortunately many common metrics are ambiguous in how they are applied and this can lead to difficulty in evaluating results. Lydite is an OO metrics framework based on JST, its goal is to provide unambiguous results for metrics run on Java source code. It provides a library of common metrics that can be configured for more particular results as well as offering a base for the creation and application of more specific metrics.
- Title: Code Critick: Informing Software Development
- Speaker: Josh Oosterman
- Abstract:
- In the world of OO (object oriented) software engineering, many rules, heuristics and guidelines are available to help designers write good code. In this presentation, I will discuss a tool, Code Critick, to assist developers by automatically evaluating source code against design guidelines. We have developed a set of automated rules built on several existing technologies, including JST and CodeRank. Feedback is presented 'In Situ', using a plug-in for the Eclipse Java IDE. Due to the large amount of information usually returned by such tools, techniques of reducing information overload are also discussed.
- Title: Empirical Software Engineering
- Speaker: Linda Pettigrew
- Abstract:
- Software and information visualisations have a long history of being difficult to evaluate. This project explores the difficulties of evaluating visualisations and offers a framework, based on methods used in other disciplines, for future evaluations.
- Title: An intelligent educational game platform
- Speaker: Paul Williams
- Abstract:
- Exploring using the Open Simulator virtual world platform to create an intelligent educational game for teaching Te Reo.
- Title: Possible Uses for a Ball-Shaped Interface
- Speaker: Martin van Zijl
- Abstract:
- A wide variety of "tangible user interfaces" (TUIs) have been developed, including "Digital Foam", a "cubical mouse", or interactive tabletops. These are interfaces you can touch and manipulate with your hands. In this project, we explored possible uses for a soft, squeeze-able ball (roughly the size of a tennis ball). We identified two areas in particular: 1) Medical Rehabilitation, 2) Alternative input for a game/modeling environment. In medical rehabilitation, patients often need to re-train motor skills, which can be easily done with a ball-shaped interface. In a modeling environment, traditional GUI interfaces (such as Blender) are often hard to use with a mouse and keyboard. For a game environment, using a ball-shaped interface can be a fun alternative to the traditional inputs.
- Title: Development of a Visualisation Infrastructure for High Performance Computing
- Speaker: Alexander Chernoglazov
- Abstract:
- This project is aimed at developing and testing the infrastructure to visualise various scientific data (for example, results of a simulation conducted on a supercomputer) on a tiled display. Tiled displays, also known as display walls, are a collection of monitors driven by separate machines (known as a cluster) with the aim of creating a uniified high-resolution display area. Different issues looked at involve the structure of a visualisation cluster, the software required to enable distributed rendering, stability and performance of scientific visualisation applications on a tiled display and computational steering.