CHAI Seminar Schedule
2005
Here is the schedule for the Computer Human Adapted Interaction Research Group seminars.
They range from formal presentations of mature work by members of the group or visitors to less
formal reports of work-in-progress. The regular meeting time is 11:00am Wednesday in the School of IT Building Seminar Room 123.
Subscribe to the CHAI Seminar Calendar
Weekly Seminars
Special Seminars
:: FileNet: a novel 3-D file system and e-mail viewer
Speaker:
Anthony Collins
Description:
This presentation will discuss FileNet, a novel 3-D file system and e-mail viewer that displays items relevant to a particular task being most visible. For example, a user selects a focus item (such as a university assignment), and is presented with other relevant items (such as e-mails from group members, and source code related to the assignment) in a natural way. This is achieved by using 3-D drawing techniques that display less relevant items further away (appearing smaller), and more relevant items closer to the user (appearing larger). FileNet also implements task-oriented features such as "To Do" lists to simplify the management of many files or e-mails related to a task, and uses interactive animation to better convey relationships between content.
Speaker Bio:
Anthony Collins is a summer student working on the Keep-in-Touch project with the Smart Internet Technology Research Group. He has just completed his Bachelor of Computer Science and Technology degree, and aims to further develop the FileNet application for his honours thesis next year.
:: Walkabout: An Asynchronous Internet Messaging Architecture Tailored to Mobile Devices
Speaker:
Adam Hudson
Description:
Walkabout is an architecture that provides network support for message transfer between mobile devices communicating across the Internet. In order to cope with changing addresses and device disconnections, devices send messages to each other across a peer-to-peer overlay network, via peers which are located in the same local network segments as they are.
This enables them to transfer messages at local network speeds, maximising their usage of transient network connectivity. The overlay uses a transfer protocol that makes use of parallel downloads wherever possible. Messages migrate to follow destination devices as they move, and are cached when the destination is not available. We have designed an approach that is robust and efficient and intend to prove this through simulation and a series of experiments.
Speaker Bio:
Adam graduated from the University of Sydney in 2003 with first class honours. Over the course of his studies, he also spent six months on exchange to the University of California, Los Angeles, and three years working on the development of a secure online health application. After graduating, he returned to full-time study to pursue a Ph.D. in Computer Science, also at the University of Sydney.
:: Implicit group messaging in peer-to-peer networks
Speaker:
Daniel Cutting
Description:
We describe "implicit group messaging": a form of many-to-many message delivery that does not require members to join explicit groups. Instead, publishers specify the characteristics of intended audiences and the system dynamically attempts to deliver messages to all matching members. We study this type of message delivery in the context of Internet-based peer-to-peer networks, presenting a novel model based on a structured overlay that encodes a peer's description in its address and uses geometric routing protocols to construct multicast delivery trees.
Speaker Bio:
After graduating from the University of New South Wales in 1998, Dan worked at several web development companies, both in Sydney and London. In April 2003 he returned to full-time study to pursue a Ph.D. in Computer Science at the University of Sydney.
:: Problems in personal information retrieval
Speaker:
Paul Thomas
Description:
Each of us has access to a vast amount of information in electronic form: personal files, source code, calendars, web sites, email, and so forth. Personal information retrieval (PIR) is the task of finding relevant and useful information amongst this. The talk will introduce PIR and discuss a few particular subproblems: server selection, results presentation, and evaluation. A prototype tool will also be presented.
Speaker Bio:
After graduating from Victoria University of Wellington (NZ), Paul worked there for a time before joining CSIRO. He is now a PhD student at the Australian National University.
:: Project Entropia - Virtual World, Real Economy
Speaker:
David Storey
Description:
Project Entropia is an online Virtual World which, unlike other online games, has an economy that is based on real currently. In Project Entropia the player joins numerous other colonists from around the globe on the distant (and fictional) planet of Calypso. Once online players are free to deposit money with which to purchase equipment and head out for fun, and sometimes profit ;)
Speaker Bio:
David Storey is a 2nd year PhD student working on methods of applying User Modelling, Data Mining and Machine Learning to computer games. Entering Project Entropia you will meet his avatar, "Deathifier", an inhabitant of the virtual world for almost three years and the owner of the USD$26,500 Treasure Island - purchased last year and gathering widespread media coverage.
:: Unobtrusive Personalization for E-buyers Using Layered User Profiles
Speaker:
Oshadi Alahakoon
Description:
She is a third year PhD student from Monash University, Melbourne. Her supervisors are A/Prof Arkady Zaslavsky and Dr. Seng Loke.
:: Distributed Caching and Replication with Relaxed Data Freshness Guarantees
Speaker:
Uwe Roehm
Description:
In order to achieve to the high scalability and performance goals of modern e-business sites, data caching and clustering at all system layers is essential. While caching is beneficial for readers, in the presence of updates caching and clustering together induce a massive consistency and data currency problem. The typical approach is to guarantee consistency via a central back- end database, but to somewhat ignore the data currency issues: Access to cached data is widely accepted to be potentially outdated.
In this talk, we describe approaches to overcome this situation by giving explicit data freshness guarantees to clients when accessing cached or replicated data. We first look at the underlying replication problem and present freshness-aware scheduling for trading data freshness against read performance when accessing a cluster of replicated databases. We then discuss middle-tier database caching.
And finally, we give an outlook on freshness guarantees for distributed caching in a cluster of application servers.
:: The Keep-in-Touch system (Mark)
Speaker:
Mark Assad
Description:
We all have a basic need to communicate with family and friends. Message-based communication, based on the written word has been important since earliest civilisations. More recently, email has been a killer application of the internet and SMS the corresponding winner for mobile telephony, precisely because they are message-based and asynchronous. They complement the critical direct communication by enabling people to send messages to each other.
Our ``Keep-in-Touch" system consists of messaging appliances that facilitate message communication within the family. The ``Keep-in-Touch" devices are appliances because they carry out a single function in a convenient, efficient manner. They are not general purpose computing systems, that allow general email or web browsing, they simply provide an easy to use voice messging service in a device that fits into the family environment unobtrusively.
This talk gives an overview of the Keep-in-Touch system.
Speaker Bio:
Mark Assad is currently doing his PhD at the University of Sydney, in the School of Information Technologies. His research is in the area Ubiquitous Computing. He is supported by a scholarship from the Smart Internet CRC and an Australian Postgraduate Award.
Mark received first class honours in his undergraduate degree of a Bachelors of Computer Science and Technology (Advanced). His honours thesis was on the development of a distributed Context Aware Framework. Mark completed his undergraduate degree at the Univeristy of Sydney. He also participated in a University Exchange program at the Univerisity of California - San Diego.
:: Sharing Digital Media on Collaborative Tables and Displays (Trent)
Speaker:
Trent Apted
Description:
We are just beginning to explore the possibilities afforded by tabletop interfaces, including the provision of some interactions that are not well supported by traditional computer displays. Face-to-face collaboration is possible, with multiple users sitting around the tabletop, each able to access computing resources. A table involves a social dimension and an environment where users can interact with a computer and each other in an informal setting. A tabletop also has the natural concept of a user?s personal space ? an area close to them in which they can work.
This talk describes the design and early experience with Cruiser ? a multi-user, gestural, collaborative digital photograph sharing interface for a tabletop ? and the techniques we use to share information with devices and other displays in its pervasive computing environment. The design is strongly influenced by the metaphor of physical photographs placed on the table and we have concentrated on the provision of an effective UbiComp interface that does not use a keyboard, a mouse or traditional WIMP (Windows, Icon, Mouse, Pointer) interface widgets. That is, with an emphasis on seamlessness.
Aspects of the interface include the ability to interactively attach audio, handwriting or drawings, as well as other photographs to the ?back? of an image after flipping it over; direct sharing of images with digital cameras, large displays and other UbiComp devices; and the provision of personal spaces ? an area close to a user in which only they can work, which is enforced by the interface.
Speaker Bio:
In March 2005, Trent Apted started his PhD at the University of Sydney after working with NICTA for one year as a research engineer on Project Nightingale; a joint research project with the University of Sydney and the Smart Internet CRC. In his PhD work, Trent is exploring the potential of the "tabletop interface" as a platform for formal and informal face-to-face collaboration, multimedia display and exchange, information visualisation and entertainment.
After studies in 2003, Trent was awarded the University Medal and first class honours in the Bachelor of Computer Science and Technology (Advanced) at the University of Sydney. For his honours thesis he investigated learning techniques for interactive tangible user interfaces, or "Smart Toys", and graduated with the highest fourth-year weighted average mark in the Faculty of Science. While undertaking his degree he did tutoring and gained experience in a range of computer science disciplines including machine learning, knowledge representation, database design, computer security, networking and user interfaces as a result of coursework and research projects while participating in the Science Talented Student Programme.
:: Adaptively Recommending Museum Tours (William)
Speaker:
William Niu
Description:
We believe that a user should be able to scrutinise the beliefs a system has about him/her, namely the user model, as well as be able to update that model. A MyMuseum prototype of a scrutable and personalised museum guide has been built to explore ways to visualise how the system adapts to the user and the reasons for the adaptation. As a result of the evaluation, we propose to improve the process of recommending a touring path by using an automatically built dynamic ontology.
Speaker Bio:
William Niu is a PhD student in the Smart Internet Technology group at the School of Information Technologies at the University of Sydney. His research interests include ubiquitous computing, user modelling, ontologies, information retrieval, and computer music. He currently focuses on user modelling, in particular, he is working on group modelling in museum and home settings. His research is supported by the Smart Internet Technology CRC.
:: User modelling, data mining and visualisation to support groups in long term complex activities.
Speaker:
Peter Goodyear, Judy Kay Peter Reimann, Kalina Yacef
Description:
This is an introduction to an ARC funded project that is joint research between Education and IT. In essence, the project converts a rich collection of data from group activity to build user models, mine for interesting patterns and transform these into useful visual representations of information that can improve group processes and products.
:: Integration of Tada-Ed's decision tree classification results into the Logic Tutor.
Speaker:
Morgan Cugerone
Description:
Tada Ed stands for Tool for Advanced Data Analysis for Education. It is used to mine students data in order to find out pedagogically relevant patterns. This tool provides many data mining methods such as Association Rules, Kmeans Clustering and Decision Tree. This talk presents how results produced by this last method are used as a contribution to the Logic Tutor, and focuses on the implementation and integration parts.
Speaker Bio:
Student in 5th year of the Ecole Sup rieure d'Ing nieurs L onard de Vinci(Engineering School) in Paris, La Defense. Morgan has spent half a year as a trainee in the SITRG, bringing contributions to TADA-Ed(Tool for Advanced Data Analysis for Education). I hope it's enough, it's weird to talk about me though :)
:: WP Programming Language: A Language That's Missing Some Features
Speaker:
Geoff Langdale
Description:
The YAWP programming language is a new general-purpose programming language. It is a strongly-typed imperative language with parameterized types. It is designed with a bias towards being compiled into machine-dependent object code as opposed to being interpreted.
YAWP is unusual in that it lacks most of the amenities that are normally considered essential to a programming language, including usable built-in types and structured control flow. Instead, the abstraction mechanisms in the YAWP language are used to recreate familiar types and control flow structures - for example, "if", "while", "int" and "float" are in library code. YAWP has a generic type mechanism similar to that of C++, with substantial extensions.
Another important feature that YAWP lacks is a concrete syntax. The canonical representation of YAWP source code is a parse-tree like data structure. Multiple different front-ends will be used to read and write YAWP source code, possibly including some quite non-traditional displays of code and data. A huge range of theological disputes about programming language design are thus avoided. Further, syntactic problems in the language do not become permanently encoded into thousands or millions of lines of existing code.
YAWP is very much a work in progress. Work is underway on a proof of concept implementation of the language for x86, PowerPC and ARM, as well as enough YAWP code to demonstrate all the features of the language.
:: Adapting Information Delivery to Groups of People
Speaker:
William Niu
Description:
When people engage with information, they are often in social groups. This applies, for example, in the case of museum visits, where people typically attend the museum and view the exhibits in small groups. For the thesis, I will look into approaches and challenges in terms of group interaction models, knowledge-based reasoning across models, and possibly, applying clustering techniques on individual models. As a start, I intend to build a system GM, Group Modeller, which will create group models from a set of individual user models.
Speaker Bio:
William Niu is a PhD student in the Smart Internet Technology group at the School of Information Technologies at the University of Sydney. His research interests include ubiquitous computing, digital libraries, information retrieval, user modelling, and computer music. He currently focuses on user modelling, in particular, he is working on group modelling in museum and home settings. His research is supported by the Smart Internet Technology CRC.
:: b-AMuSEd
Speaker:
Josiah Poon Ronnie Ma Phoebe Mong
Description:
How do you manage your large collection of multimedia items?
How do you even manage to create a playlist of all those video, audio, and picture files?
The b-AMuSEd (basic AMUSE development) team is exploring the ideas to these problems with the Adpative Multimedia Selection Engine - a clever, adaptive multimedia system that allows you to manage and present multimedia items not just to your preferences, but to your lifestyle as well.
This presentation will highlight the ideas of AMUSE and the design of the current prototype.
:: Highlights of the 2-day Gestural Interaction Workshop held at NICTA
Speaker:
Trent Apted
Description:
The ability to recognise human gestures holds great promise for the development of natural human machine interfaces. However, as a multi-discipline research field, gesture recognition is still in its infancy. The purpose of the Gestural Interaction Workshop is to bring together researchers working on human gesture recognition, computer vision and their applications to human-machine interfaces, and to provide a forum for the presentation and exchange of ideas and research currently in progress. The topics covered at the workshop include: Gesture-based human-machine interaction, the state-of-the-art in research; Analysis, segmentation and recognition of gestures; Thermal imaging and its uses in gestures, facial expressions, emotions; Computer vision; Speech and gesture-based multimodal user interaction; and Robustness of gesture recognition in real-world applications.
Speaker Bio:
In March 2004, Trent Apted joined the Networks and Pervasive Computing program at National ICT Australia (NICTA) as a research engineer on Project Nightingale, a joint research project with the University of Sydney and the Smart Internet CRC. In conjunction with researchers from NICTA and SITCRC, Trent has developed a novel digital photograph sharing application for use in a pervasive computing environment, targetted at allowing an elder to engage in reminiscence activities without the need for a keyboard, mouse or computer monitor.
After studies in 2003, Trent was awarded the University Medal and first class honours in the Bachelor of Computer Science and Technology (Advanced) at the University of Sydney. For his honours thesis he investigated learning techniques for interactive tangible user interfaces, or "Smart Toys", and graduated with the highest fourth-year weighted average mark in the Faculty of Science. While undertaking his degree he did tutoring and gained experience in a range of computer science disciplines including machine learning, knowledge representation, database design, computer security, networking and user interfaces as a result of coursework and research projects while participating in the Science Talented Student Program.
During the last two years of his undergraduate degree, he published papers in four international conferences and submitted a paper to the International Journal on Continuing Education and Lifelong Learning (special issue on Concepts and Ontologies in WBES), which appeared in 2004. He has since been working for NICTA on Project Nightingale and is about to commence a PhD in Computer Science with the Networks and Systems Laboratory at the University of Sydney.
:: Cellerator
Speaker:
Bob Kummerfeld
Description:
Cellerator is a very light weight web framework developed over the last few years by Bob Kummerfeld and Piers Lauder. Features include support for embedding python in pages, automatic layout of cells based on cell naming, sessions and authentication. It has a web based site administration system. This talk will describe the design and implementation in detail and show the system in action.
:: Seminar - Exhibit Anywhere, Anytime
Speaker:
Simon Goldrei
Description:
This talk provides an overview of the Seminar system. Seminar is a three part system for the broadcasting and archival of seminars such as conference proceedings, research group meetings, student presentations and lectures. Seminar provides a simple, uniform method to capture a computer-based presentation regardless of the software application used in the presentation. Seminar captures the video and audio of the presenter, as well as a motion capture of the computer?s main display.
Using this approach the user is free to present slides from applications such as PowerPoint or Keynote, demonstrate software, Websites or exhibit to the world anything displayed on their computer. The Seminar system consists of an application written for Mac OS X, a streaming media server and a web-accessible database. Presentations captured with Seminar can be viewed on any platform that Apple?s QuickTime Player supports.
This talk will provide implementation level details of recent improvements in achieving better hardware based performance particularly on existing and future G4 processor based systems by effective utilisation of the GPU.
The talk will also highlight ongoing research efforts to include the ability to create Learning Objects, which can be thought of as reusable 'segments' of a presentation. Our approach, in creating Learning Objects, attempts to automatically capture the meta-data pertaining to each Learning Object in order to provide relevant searching of the Learning Objects stored in an archive. Current standards for describing Learning Objects are discussed.
This talk will be live web-cast and available on-demand afterwards at SydneySeminars.com. Requirements are the Apple QuickTime v6.0 browser plug-in or better.
Speaker Bio:
Simon Goldrei is a Honours student in the Smart Internet Technology research group at the School of Information Technologies at the University of Sydney. His research focuses on online learning, in particular, he is working on the Seminar system. His research and development is being jointly funded by Apple Computer Australia and the Apple University Consortium.
:: JAM - Dynamic subscriptions in a Publish/Subscribe messaging system
Speaker:
Michael Avery
Description:
This talk focuses on the JAM Publish/Subscribe messaging system. In a Publish/subscribe messaging system, subscribers make requests to receive certain types of data from the network. The subscription language in JAM differs from other systems because it allows subscribers change their requests over time. This talk will describe this subscription language, and why this language is beneficial. It will also discuss methods of efficiently evaluating these dynamic subscriptions.
Speaker Bio:
Michael Avery is a PhD student in the Smart Internet Technology group at the School of Information Technologies at the University of Sydney. His research focuses on computer networks, in particular, he is working on the JAM Publish/Subscribe messaging system. His research is being funded by the Smart Internet Technology CRC.
:: Fusion of speech and manual gesture recognition-based inputs for multimodal HCI
Speaker:
Julien Epps
Description:
Speech and manual gesture can facilitate more natural human-machine interaction in applications for which the traditional keyboard and mouse input mechanisms are inappropriate, however the possibility of their concurrent use raises the issue of how best to fuse the two inputs. This seminar gives a brief overview of possible approaches to multimodal fusion, describes the results of a user study into speech and gesture constructions, and outlines some of the practical considerations in applying these results to a live speech and gesture-based multimodal prototype.
A brief introduction to the NICTA VICAT (Visualization and Interaction on a Collaborative Access Table) project will also be given. This project explores methods for intense collaboration between multiple remote users via tables comprising large vertical and horizontal screens.
Speaker Bio:
In early 2004, Julien Epps accepted a Researcher position at the Australian Technology Park Laboratory of National ICT Australia, within what is now known as the Interaction, Machines and Graphic Environments (IMAGEN) Program. His research in this position has concentrated on speech recognition and practical methods for fusion of multimodal inputs, mainly for desktop-style devices. Julien produced one of the first studies into the integration patterns of speech and manual gesture, and has developed a number of novel techniques for increasing the robustness of speech recognition.
After completing his PhD with the School of Electrical Engineering and Telecommunications at UNSW in 2000, Julien Epps joined Motorola Labs in Sydney as a Senior Research Engineer. During his two and a half years with Motorola, he worked on speech recognition, speaker adaptation and multimodal interfaces for handheld devices. Julien holds Honorary Associate and Conjoint Senior Lecturer positions with the Universities of Sydney and NSW respectively, and is lecturing in courses at both institutions in 2005.
Slides
:: Using technologies of ubiquitous computing to assist caregivers of children with autism
Speaker:
Asso. Prof. Gregory D. Abowd
Description:
Parents and teachers of children with autism (CWA) often use several therapeutic interventions, keeping vast records to assess improvement in behavior and learning. Automated capture technologies and the associated access interfaces for exploring past experiences are particularly promising for monitoring the effectiveness of these interventions for behavioral and learning disabilities in children.
Behavioral and learning data can be captured, analyzed, and mined over time to provide valuable evidence to track the progress of any intervention. Prototypes developed for this problem must address both technical and social factors to be successful. These factors include providing for all elements of the care cycle, understanding the need for qualitative richness of collected data, minimizing the effort required to use capture technology, addressing privacy concerns, and considering financial constraints. In this talk, I will discuss some of the work done by my research group over the past year in formative studies of this application domain and some deployment studies in structured and unstructured settings, both in homes and in schools.
Speaker Bio:
Gregory D. Abowd
Associate Professor, College of Computing & GVU Center, Georgia Tech Founding Director, Aware Home Research Initiative abowd@cc.gatech.edu http://www.cc.gatech.edu/fac/Gregory.Abowd
Gregory D. Abowd is an Associate Professor in the College of Computing and GVU Center at the Georgia Institute of Technology. His research is in the area of Human-Computer Interaction, with particular focus on the engineering and evaluation of mobile and ubiquitous computing applications. He leads a research group in the College of Computing focussed on the development of prototype future computing environments which emphasize mobile and ubiquitous computing technology for everyday uses. The general themes he investigates include automated capture environments, context-aware computing, and natural interaction. He has focussed his applications work in the domains of university education (the Classroom 2000 and eClass projects), the office (CyberDesk, TeamSpace) and home (the Aware Home). His most recent work has extended into the domain of early childhood interventions for children with autism. Dr. Abowd has affiliations with several campus research groups, including the GVU Center and the Broadband Institute, for which he is Associate Director in charge of the Residential Laboratory research.
Dr. Abowd received a BS in Mathematics from the University of Notre Dame in 1986 and the degrees of M.Sc. (1987) and D.Phil. (1991) in Computation from the University of Oxford, where he attended as a Rhodes Scholar. Before coming to Georgia Tech in 1994, Dr. Abowd held post-doctoral positions with the Human-Computer Interaction group at the University of York in England and with the Software Engineering Institute and Computer Science Department at Carnegie Mellon University.
:: Project Nightingale: from SharePic and MEMENTO to Anotepic
Speaker:
Trent Apted David West
Description:
Project Nightingale is a joint project between the National ICT Australia, the Smart Internet Technology Cooperative Research Centre and the University of Sydney and explores the needs of Australia's aging population and the role of smart wireless networks and pervasive computing in memory sharing and reminiscing. The project is now coming to a close. This talk will present the main demonstrator applications developed through the course of the project and results of our initial user trials.
SharePic is a ubiquitous, multi-user, multi-touch, gestural collaborative digital picture sharing application. It was initially developed to use the DiamondTouch technology from Mitsubishi Electronic Research Laboratories (MERL), but has since been adapted to the Mimio pen system on a regular tabletop for (single-touch) multi user interaction. Users can share digital photographs in an interactive, face-to-face setting without the need for a keyboard and mouse.
MEMENTO is a physical/digital scrap book, implemented with our own drawable user interface (DUI). Using the Anoto pen system and regular paper, we produce a digital copy of physical information entered into the scrapbook; including images, linked audio and handwritten annotations. The digital copy is represented as a web page that can instantly be shared across the globe.
Recently, these two applications were combined to create Anotepic, an annotation system for digital images. Still without the need for a keyboard and mouse, users can collaborate on a design or reminiscing task centred around digital photographs on a tabletop interface. Audio can be added to the images and they can be digitally labelled with handwritten annotations using a pen and paper in a natural way.
Speaker Bio:
David West -
David West has been working as a Research Associate at the University of Sydney, in conjunction with the Smart Internet CRC and NICTA, on projects Nightingale and Bluestar, since December 2003. During this time David has researched a number of areas in Pervasive Computing, with the emphasis of supporting reminiscence activities among the elderly, and the seamless interaction of different modalities distributed across multiple devices in a virtual personal server space. In addition, David has explored a hybrid approach to location awareness, incorporating position estimation using the GSM network, and more exact positioning using Bluetooth beacons.
David graduated from the University of Dublin, Trinity College in 2003, where he completed his Masters degree in Networks and Distributed Systems. Along with graduating at the top of both his postgraduate and undergraduate classes, David is a University Medalist and scholar of the University. For his Bachelors final year project, David researched and implemented a system for face recognition by modelling the colour distribution in an individual's face. For his Masters thesis he implemented and evaluated the mobile ad-hoc networking protocol, AODV, for the Windows CE and XP operating systems.
David has worked for a web-design company in his home county of Kerry in Ireland. He has also worked in the technology division of Barclays Capital investment bank in New York.
David has travelled widely throughout many countries in Europe, as well as having spent a number of months traveling the USA. He is about to embark on an extended period travelling Australia and Asia.
Trent Apted -
In March 2004, Trent Apted joined the Networks and Pervasive Computing program at National ICT Australia (NICTA) as a research engineer on Project Nightingale, a joint research project with the University of Sydney and the Smart Internet CRC. In conjunction with researchers from NICTA and SITCRC, Trent has developed a novel digital photograph sharing application for use in a pervasive computing environment, targetted at allowing an elder to engage in reminiscence activities without the need for a keyboard, mouse or computer monitor.
After studies in 2003, Trent was awarded the University Medal and first class honours in the Bachelor of Computer Science and Technology (Advanced) at the University of Sydney. For his honours thesis he investigated learning techniques for interactive tangible user interfaces, or "Smart Toys", and graduated with the highest fourth-year weighted average mark in the Faculty of Science. While undertaking his degree he did tutoring and gained experience in a range of computer science disciplines including machine learning, knowledge representation, database design, computer security, networking and user interfaces as a result of coursework and research projects while participating in the Science Talented Student Program.
During the last two years of his undergraduate degree, he published papers in four international conferences and submitted a paper to the International Journal on Continuing Education and Lifelong Learning (special issue on Concepts and Ontologies in WBES), which appeared in 2004. He has since been working for NICTA on Project Nightingale and is about to commence a PhD in Computer Science with the Networks and Systems Laboratory at the University of Sydney.
Slides
:: Context modelling for Ubiquitous Computing
Speaker:
Thomas Strang
Description:
Context-awareness is one of the drivers of the ubiquitous computing paradigm, whereas a well designed model is a key accessor to the context in any context-aware system. This talk provides a survey of the the most relevant current approaches to modelling context for ubiquitous computing. This includes (but is not limited to) approaches based on graphics, logics and ontologies. Numerous approaches are reviewed, classified relative to their core elements and evaluated with respect to their appropriateness for ubiquitous computing.
Speaker Bio:
Thomas Strang was born in Eschweiler near Aachen, Germany, in 1972. He studied computer science at the University of Technology (RWTH) in Aachen and received his Dipl.-Inform. degree in 1998. During his studies his special interests have been on communication and distributed systems, including high speed networks and telecommunications, local area networks, multimedia communications, computer graphics and security. His diploma thesis was about a video gateway to support video streaming to mobile clients. Parallel to university he has been working in industry since 1988, where he gained experiences in design and development of large scale security projects. Since July 2000 he is working as a researcher with the German Aerospace Center (DLR) in Oberpfaffenhofen, Germany. Here his research focus is on ubiquitous and pervasive computing, location- and context-awareness, service discovery and execution frameworks, Semantic Web and smart mobile devices.
Slides
:: Virtual Characters in a Physical World
Speaker:
Michael Kruppa
Description:
The talk will focus on two projects, namely PEACH [Personal Experiences with Active Cultural Heritage] and VRI [the Virtual Room Inhabitant]. Both projects deal with virtual characters being capable of moving within real, physical space (i.e. in PEACH [a museum guide project with localized information service] characters may "jump" from stationary information systems onto a PDA to accompany the user, while in VRI, the character is able to move freely along all surfaces within a room). Also, both projects make use of external user models to adapt to the users preferences and previous knowledge. The overall idea is to develop characters which establish "a long time relationship" with their users by accompanying them in very different situations, and in doing so, improve the overall acceptance by the users. I will also mention a user study I have recently conducted in Saarbruecken, regarding multiuser interaction (with inhomogenous groups) with public displays.
Speaker Bio:
1996-1999: Undergraduate studies in Computer Science, University of Bremen 1999-2001: Graduate studies in Computer Science, Saarland University 2001: Masters Degree(Diplom) in Computer Science at Saarland University 2001-2003: Research Assistant (PhD) at DFKI Saarbr?cken (the german research center for artificial intelligence, Prof. Wahlster) since 12.2003: Member of the European Post-Graduate College Language Technology and Cognitive Systems For further information, please have a look at: http://www.it.usyd.edu.au/~michaelk
|
![[Cellerator]](Images/cellerator.png "Cellerator Home"){kind=small}