Contents

1. 1 November 2011
2. 2 October 2011
3. 3 September 2011
4. 4 August 2011
5. 5 July 2011

November 2011

Title: Co-authorship and other publication issues

Abstract:

This will cover matters such as:

- what kind of contribution merits a co-authorship of a paper?

- how is order of authors determined?

- how do these things vary from one discipline to another?

- how can disagreements about co-authorships be resolved?

Panel Members:

- Prof Kim Marriott (Head, Clayton School of IT)

- A/Prof Graham Farr (HDR Co-ordinator, Clayton School of IT)

- A/Prof Maria Garcia de la Banda (Head, Caulfield School of IT)

- Dr Arun Konagurthu (Larkins Fellow)

The discussion will be chaired by Prof Kim Marriott (Head, Clayton School of IT). Panelists will talk for about 5 mins each and then there will be plenty of time for discussion and questions.

The discussion will focus on computer science but will touch on related issues in other disciplines.

October 2011

Host: Hiran Ganegedara
Title: Scalable Data Mining: A Sammon's Projection Based Techniqe for Merging Self Organising Maps
Abstract:
Self-Organizing Map (SOM) and Growing Self-Organizing Map (GSOM) are widely used techniques for exploratory data analysis. The key desirable features of these techniques are their applicability to real world data sets and their ability to visualize high dimensional data in low dimensional output space. One of the core problems of using SOM/GSOM based techniques on large datasets is the high processing time requirement. One possible solution is the generation of multiple maps for subsets of data where the subsets consists of the entire dataset. However the advantage of topographic organization of a single map is lost in the above process. I will be presenting a new technique where Sammon's projection is used to merge an array of GSOMs generated on subsets of a large dataset.I will be discussing cluster accuracy and performance analysis for several datasets. This technique is ideally suited to harness the processing power of parallel computing resources.

Host: Dror Cohen
Title: "Computational Neuroscience, Physics envy and the Free-Energy Principle"
Abstract:
The biological sciences are increasingly utilising computational approaches for data analysis, as well as to better understand the governing mechanisms. Computational insights are particularly valuable in the neural sciences where the relationship between function and physiology is intricately coupled and difficult to discern. A recently proposed Free-Energy principle attempts to provide a unifying framework for the understanding of computational mechanisms throughout the cortex. We demonstrate how this principle can produce topography preservation, a feature that has been well observed in the cortex.

September 2011

• Tuesday, 6 September 2011, 1:00-2:00pm (Room 115/63)

Host: Sara Miranda

Title: "The Library – your new best friend and partner in your postgraduate degree"

Abstract: 

The Hargrave-Andrew Library has staff dedicated to helping academics and students in the Clayton School of Information Technology.

Sara Miranda, the information research skills librarian can assist in effective use of library services and resources, including databases, finding information, citing and referencing.

Noriaki Sato, the learning skills adviser, can assist with thesis writing, oral communication and presentation, and writing for research projects. Postgraduate students can arrange individual sessions with Nori, or participate in group sessions tailored to your needs. 

In this session we will present an overview of what we do, and go into some detail on how you can use our services and facilities. If there is anything you wanted to know about the library but were reluctant or didn’t have time to ask,this is your opportunity.

August 2011

• Thursday, 25 August 2011, 1:00-3:00pm (Room 145/63)
  
  
  Host: Brian Jenney
  Title: "The Weibull Distribution"
  Abstract: 
  Reliability is invariably at the 'sharp end' of engineering. Here you are dealing with complex systems that cost a lot of money and are very expensive to test to destruction. This you need to do for renewal and replacement strategies to keep the system going. Consider a case where a sub-system in a machine is being analysed to estimate its reliability. A random sample of 8 is taken and their performance recorded with care. After 230 hours in operation the first one fails but, with pressure from the sales department, another is lost and sold after 362 hours.Another serviceable one is lost in a fire after 385 hours, but the next two recordings are genuine failures at 415 and 496 hours respectively. 

  Then another goes to an overseas customer at 524 hours and we lose trace of that one. A future 'true' subsystem fails at 547 hours.So we have recorded 4 proper failures, lost three and still have the eighth running at 600 hours and have to consider whether to give up and call the test a waste of time.
  There is a simple analytical technique that can use these data and give us an idea of whether we are improving with time, observing random failure or experiencing wear out. This is based on a three parameter distribution that has been simplified and converted to graphical methods that are best explained by examples adding in complications iteratively.
  
  
  
  
• Thursday, 4 August 2011, 1:30-2:30pm (Room 135/26)

Host: Marc Cheong

Title: "Twitter: from 'What are you doing?' to #ThingsWeLearnedOnTwitter"
Abstract: 

In this talk, I'll give an overview of microblogging, and someinteresting observations from my published work and from current literature.
This talk will be non-technical and interactive, where the audience gets to learn more about:
- the concept of microblogging (and influences on other social networks)
- the idiosyncrasies of microblogging habits amongst its user base
- memes and their evolution on Twitter
- social usage habits and privacy implications
- hidden in plain sight: why combined user+message metadata can reveal much about social behavior
Last but not least, a live demonstration on Twitter (time permitting) will help illustrate the above points.

July 2011

Host: Carolina Segovia

Title: "Integration of Data Mining, Game Theory and Optimization for Tackling Crime"

Abstract:

Classical Data Mining models for crime prediction give only a partial solution to the problem, because they do not consider the game between the police and the criminals. By incorporating Game Theory into the models it is possible to consider this adversarial behavior, through modeling the problem as a Bayesian Stackelberg game. We propose a methodology to predict crime, by integrating dynamic crime prediction Data Mining models with a game theoretical approach that enables police resources to be optimally allocated.

Panel Discussion on "Life after PhD"

This weeks session will take the form of a panel discussion, on employment opportunities and pathways to take after graduation. The event will be hosted by a special panel of successfully employed PhD graduates. This is a great opportunity for current students to get advice and information from experienced PhD graduates.

Panel Members:

Kerri Morgan

Kerri Morgan recently returned to Monash University as an ARC Australian Post-doctoral Research Fellow. She is currently working on the algebraic properties of the chromatic polynomial, a polynomial that gives the number of proper colourings of a graph in at most k colours. She previously worked as a research fellow at the University of Melbourne in the area of efficient synchronisation of large data repositories. During this time she co-founded the Theoretical Research In Computer Science (TRICS) reading group. Kerri is interested in all areas of theoretical computer science with particular emphasis on graph theory.

Leelani Kumari Wickramasinghe

Leelani Kumari Wickramasinghe obtained a BSc (Hons) in Computer Science and Engineering with a first class honours from University of Moratuwa, Sri Lanka in 2000 and a PhD in Information Technology from Monash University, Australia in 2007. During her PhD tenure she worked as an assistant lecturer in the same faculty. Prior to her PhD, she worked as a software engineer for Millennium IT, Sri Lanka. Since 2007 she is working as a research fellow in the Faculty of Medicine, Nursing and Health Sciences, Monash University. Her current research is on mining data collected from a large health care system. Prior to this, she worked on developing intelligent agent-based techniques for consumer life cycle management.

Matthew Butler

Matt is a lecturer and researcher within the Berwick and Caulfield schools of Information Technology. His research interests lay in the visualisation of information using multimedia technologies, as well as the use of emergent technologies for education, in particular the learning challenges faced by novice programming students. This was the topic of his PhD thesis, which was undertaken part-time and completed in 2010. Matt's teaching areas include programming, game development, and information visualisation.

Matthias Vigelius

Matthias Vigelius is currently a research fellow at the Faculty for Information Technology, Monash University. Born in Germany, he did his undergrads in Berlin (2000) and obtained a Masters degree in physics from the University of Heidelberg (2005). He completed his PhD in astrophysics at Melbourne University in 2009. Current research interests are focused on developing efficient implementations of algorithms to solve reaction-diffusion equations on graphics processing units.

Yee Ling Boo

Yee Ling Boo received Bachelor of Information Technology (Honours) from Multimedia University, Malaysia and Ph.D in Information Technology from Clayton School of Information Technology, Faculty of Information Technology, Monash University. Before the pursuit of her Ph.D degree, she worked as a software engineer in Malaysia. She is currently an associate lecturer at School of Information Systems, Faculty of Business and Law, Deakin University. She is also a lab member of Cognitive and Connectionist Systems Lab (CCSL) at Clayton School of Information Technology.