2012
Seminar Cancelled, sorry for any inconvenience
Ninth ISE Seminar for 2012 to be held on 18 May from 11:00am to 12:00pm in room 11B24
ALL WELCOME
-oOo-
“UC-ACM Student Chapter – A Brief Introduction”
By
Lavneet Singh
Abstract:
ACM stands for Association of Computer Machinery and it’s the oldest and biggest computing society among whole world. Thus, to connect through ACM, we have bring ACM chapter here first time in Australia. The objective and motto of the chapter is to promote for joining as members and to increase awareness of computing and ICT among broad ranges of audiences. This chapter is open for everyone to promote chapter university wide in broad range of students, not only targeting ISE students. The objective and motto of chapter is to increase awareness about computing and ICT to our audiences and to provide the people resources which they can’t get in their university degree. So in this seminar, I will describe the objectives and its future strategic plan and also misconceptions too. IN brief, it’s an whole seminar giving details on UC-ACM chapter and its working strategy. Please come and join us.
Speaker:
Lavneet has degrees in Bachelors of Computer Science and Engineering, Bachelor of Computer Applications, Master of Computer Applications, Master of Business Administration, and Grad Cert in Project Management. He has received several scholarships and awards as a young researcher in India , and has published widely in several areas including machine learning and pattern recognition, Last couple of years, he has been a visiting researcher in several highly reputed institutions in Switzerland and Australia, including University of Melbourne, NICTA, and ANU. He was a visiting researcher in Faculty of ISE, UC, from July 2011 to Dec 2011, and worked on a collaborative project with India and Australia, and some of this work has been recently accepted as a ERA/CORE ‘A‘ level Conference (WCCI 2012), to be held in Brisbane Australia. He has recently commenced PhD in Information Sciences and Engineering and working on pattern recognition, machine learning and algorithm optimization.
Eighth ISE Seminar for 2012 to be held on 11 May from 11:00am to 12:00pm in room 11B24
ALL WELCOME
-oOo-
“Automatic Speaker Verification with Information from Glottal Waveforms”
By
David Vandyke
Abstract:
The majority of work in the field of automatic speaker verification (SV) focuses on modelling short term magnitude spectrum characteristics of speech, and some prosodic features such as pitch, energy and formants. Under certain conditions, there exist very accurate ASR systems based on these techniques, most commonly Gaussian Mixture Modelling of Mel Frequency Cepstral Coefficients with a Universal Background Model. A smaller body of work focuses on speaker discriminatory information found in glottal waveforms, receiving more consideration over the past decade due to several published high quality results and the lack of such a baseline method in this area. The term glottal waveform describes the air flow through the glottis and relates to the vibration pattern of the vocal folds during phonation of voiced speech. This research looks at developing theory and methods to use the information in these glottal waveforms to improve and/or develop stand alone ASR systems.
Speaker:
David Vandyke is a PhD candidate from the faculty of Information Sciences and Engineering at the University of Canberra. He graduated from Monash University in 2008 with a Bachelor of Advanced Science with Honours (H1), and commenced his doctoral studies in February 2011. He is a current recipient of the Australian Postgraduate Award from the Australian Government.
Seventh ISE Seminar for 2012 to be held on 20 Aprilfrom 11:00am to 12:00pm in room 11B24
ALL WELCOME
-oOo-
“A Novel Approach to Protein Structure Prediction Using PCA Based Extreme Learning Machines and Multiple Kernels”
By
Lavneet Singh
Abstract:
In the area of bio-informatics, large amount of data is harvested with functional and genetic features of proteins. The structure of protein plays an important role in its biological and genetic functions. There is an urgent need of transforming the large harvested human genomic sequences data using effective and efficient computational algorithms to extract the biological knowledge. The structure of protein plays an important role in its biological and genetic functions. Past studies tried to predict protein folding patterns based on an approach involving comparison of between the unknown protein sequences, and the known protein sequences by computing sequence similarities. In this study, we propose a protein structure prediction scheme based novel learning algorithms – the extreme learning machine and the Support Vector Machine using multiple kernel learning, The experimental validation of the proposed approach on a publicly available protein data set shows a significant improvement in performance of the proposed approach in terms of accuracy of classification of protein folds using multiple kernels where multiple heterogeneous feature space data are available. The proposed method provides the higher recognition ratio as compared to other methods reported in previous studies
Speaker:
Lavneet has degrees in Bachelors of Computer Science and Engineering, Bachelor of Computer Applications, Master of Computer Applications, Master of Business Administration, and Grad Cert in Project Management. He has received several scholarships and awards as a young researcher in India , and has published widely in several areas including machine learning and pattern recognition, Last couple of years, he has been a visiting researcher in several highly reputed institutions in Switzerland and Australia, including University of Melbourne, NICTA, and ANU. He was a visiting researcher in Faculty of ISE, UC, from July 2011 to Dec 2011, and worked on a collaborative project with India and Australia, and some of this work has been recently accepted as a ERA/CORE ‘A‘ level Conference (WCCI 2012), to be held in Brisbane Australia. He has recently commenced PhD in Information Sciences and Engineering and working on pattern recognition, machine learning and algorithm optimization.
Sixth ISE Seminar for 2012 to be held on 13th April from 11:00am to 12:00pm in room 11B24
ALL WELCOME
-oOo-
“Classification and feature selection using Extreme and Cross Domain Transfer Machine Learning”
By
Lavneet Singh
-oOo-
Abstract:
The conventional machine learning methods usually assume that the training and test data are drawn from the same data distribution. Meanwhile, classifiers trained with only a limited number of labelled patterns are usually not robust for pattern recognition tasks. Cross Domain Learning methods have shown promising results by leveraging labelled pattern from the auxiliary domain to learn a robust classifier for the target domain which has only a limited number of labelled samples.
It is clear that the learning speed of feed – forward neural networks is in general far slower than required and it has been a major bottleneck in their applications for past decades. Two key reasons behind may be: first is the slow gradient-based learning algorithms are extensively used to train neural networks, and secondly all the parameters of the networks are tuned iteratively by using such learning algorithms. Unlike these conventional implementations, our study proposes a new learning algorithm called Extreme Learning Machine (ELM) for single-hidden layer feed - forward neural networks (SLFNs) which randomly chooses hidden nodes and analytically determines the output weights of SLFNs.
In this seminar, We try to deal with the considerable change between feature distribution of different domains in finding the patterns of genetic sequencing, Cross Domain Machine Learning and Extreme Machine Learning minimizes the structural risk functional and Maximum Mean Discrepancy. We are working towards developing a reduced gradient procedure to efficiently and effectively lean the linear combination coefficient's of multiple base kernels as well as target classifiers for finding the patterns in different datasets.
Speaker:
Lavneet has degrees in Bachelors of Computer Science and Engineering, Bachelor of Computer Applications, Master of Computer Applications, Master of Business Administration, and Grad Cert in Project Management. He has received several scholarships and awards as a young researcher in India , and has published widely in several areas including machine learning and pattern recognition, Last couple of years, he has been a visiting researcher in several highly reputed institutions in Switzerland and Australia, including University of Melbourne, NICTA, and ANU. He was a visiting researcher in Faculty of ISE, UC, from July 2011 to Dec 2011, and worked on a collaborative project with India and Australia, and some of this work has been recently accepted as a ERA/CORE ‘A‘ level Conference (WCCI 2012), to be held in Brisbane Australia. He has recently commenced PhD in Information Sciences and Engineering and working on pattern recognition, machine learning and algorithm optimization.
Fifth ISE Seminar for 2012 to be held on 23 Marchfrom 11:00am to 12:00pm in room 11B24
ALL WELCOME
-oOo-
“LINK: development of a human-centred problem solving method
initial findings”
By
Paul Dowling
-oOo-
Abstract:
The seminar will present a human-centred problem solving method called LINK. Incorporating elements of creative problem solving, analogical reasoning, and neural theory of language, LINK assists individuals to address ill-structured problems in domain A by systematically evoking and using their own knowledge from domain B. The seminar will cover the study method and initial findings.
Speaker:
Paul is in the final year of his part-time PhD. In his work career he has spent the last two years working part time as a strategy and evaluation consultant. Prior to that Paul has a worked in a variety of sectors including ten year in the corporate sector for IBM, five years in the private sector as a consultant in the business incubation industry and six years in the drugs in sport field.
Fourth ISE Seminar for 2012 to be held on 9 Marchfrom 11:00am to 12:00pm in room 11B24
ALL WELCOME
-oOo-
“Structure and Motion without Correspondences”
By
Stefan Lehmann
-oOo-
Abstract:
I will highlight key research ideas and important findings of my PhD thesis entitled ‘Structure and Motion without Correspondences’ (The University of Queensland). The aim of this research was to develop and evaluate a new approach to structure and motion. A novel paradigm is investigated to estimate three-dimensional motion and epipolar geometry constraints from images of a three-dimensional scene, without requiring correspondences and without applying any additional constraints. This new paradigm avoids the limitation of conventional methods due to their reliance on point correspondences. The fundamental estimation problem is re-formulated in the frequency and Radon domains. Mathematical models are derived that lead to a number of robust estimation schemes. Significant results of my PhD research have been published in the January 2007 edition of the IEEE Transactions on Pattern Analysis and Machine Intelligence which can be accessed online through IEEE Xplore (http://ieeexplore.ieee.org).
Speaker:
Stefan Lehmann graduated from the Darmstadt University of Technology with the German University Degree (Diplom) in Electrical Engineering. Following graduation, he moved to the United States, where he worked as a Researcher at the Fraunhofer Center for Research in Computer Graphics. During this time, he was involved in the SysCoP (System for Copyright Protection) project, a patented digital watermarking encryption technology recognised worldwide. Additionally, he was engaged as a Guest Professor in the International Certificate Program for New Media at the Rhode Island School of Design. When he returned to Germany, he worked as a Research Associate in the Department for Virtual Reality and Scientific Visualisation at the Fraunhofer Institute for Computer Graphics in Darmstadt. During this time, he participated in the ARVIKA project, the world’s largest project on Augmented Reality. Additional duties included student supervision and lecturing at the Darmstadt University of Technology.
In March 2007, he was awarded the PhD in Electrical Engineering from the University of Queensland. His PhD research focused on image processing and analysis, computer vision, and digital signal processing, and was supported by a scholarship (IPRS) from the Commonwealth Government of Australia and a scholarship (UQIPRS) from the University of Queensland. During this time, he was also engaged as a Tutorial Assistant and subsequently received the Excellence in Tutoring award from the School of Information Technology and Electrical Engineering in 2004. Following his PhD candidature, he worked as a Teaching/ Research Assistant in the School of Information Technology and Electrical Engineering, The University of Queensland, before taking up a position as a Design Engineer with Filtronic Pty Ltd in Brisbane, where he conducted research and development in the areas of microwave imaging, radar technology and digital signal processing. In November 2008, he was appointed as a Researcher with National ICT Australia (NICTA), in the Human Performance Modeling (HPM) program. The HPM program focuses on modelling complex human decision making in the Air Traffic Management (ATM) domain. In parallel, he has been performing teaching duties within the scope of his additional appointment as an Honorary Lecturer, School of Information Technology and Electrical Engineering, The University of Queensland. He is a Senior Member of the IEEE.
Third ISE Seminar for 2012 to be held on 10 February from 11:00am to 12:00pm in room 11B24
ALL WELCOME
-oOo-
“Imaging Spectroscopy for Scene Analysis”
By
Antonio Robles-Kelly
-oOo-
Abstract:
In this seminar, I will introduce imaging spectroscopy, its capabilities and complexities. We will depart from the differences and similarities between imaging spectroscopy and trichromatic technologies. I will present how the generally ill-posed problem of recovering the illuminant power spectrum can be rendered tractable using ultispectral imaging. I will then go on to show how specularities, shading and relfectance can also be recovered once the illuminant power spectrum is at hand. Along these lines, I will present a method for specularity removal based upon an entropy minimisation scheme. I will finalise the talk by show how multispectral imagery can be used for object material recognition, re-illumination, re-colouring and plant pest detection.
Speaker:
Antonio Robles-Kelly
· Antonio Robles-Kelly received his B.Eng. degree in Electronics and Telecommunications with honours in 1998. In 1999, after a year in industry, he enrolled in a PhD programme at the University of York. During his doctorate, in 2001, he visited the University of South Florida as part of the William Gibbs/Plessey Award to the best research proposal to visit an overseas research lab. After receiving his doctorate, Dr. Robles-Kelly remained in York until Dec. 2004 as a Research Associate under the MathFit-EPSRC framework. In 2005, he took a research scientist appointment with National ICT Australia (NICTA) at the Canberra Lab.
· After working on surveillance systems with query capabilities, in 2006 he was appointed project leader of the Spectral Imaging project and promoted to Senior Researcher. From 2007 to 2009, he was a Postdoctoral Research Fellow of the Australian Research Council. In 2009, he became a conjoint senior lecturer at the University of New South Wales at the Australian Defense Force Academy (UNSW@ADFA) and, in 2010, he was elevated to the Senior Member Grade of the IEEE. He is an Adjunct Associate Professor at the ANU and a Principal Researcher at NICTA.
· He has been a technical committee member of mainstream computer vision and pattern recognition conferences and an Associate Editor of the IET Computer Vision Journal. His research interests are in the areas of Pattern Recognition, Computer Vision, Spectral Imaging and Computer Graphics.
Second ISE Seminar for 2012 to be held on 20 January from 11:00am to 12:00pm in room 11B24
ALL WELCOME
-oOo-
“Reflection subgroups of complex reflection groups”
By
Don Taylor
The University of Sydney
-oOo-
Abstract:
The usual notion of a reflection in Euclidean space has a natural analogue in vector spaces over the complex numbers.
In this talk I will give a brief introduction to finite groups generated by complex reflections and report on some new results characterising their reflection subgroups
Speaker:
Don Taylor, University of Sydney
Don Taylor is currently a Senior Research Fellow at the University of Sydney working with the Computational Algebra Group on the development of the Magma computer algebra system.
From 2008 until 2011 he was the Editor of the Bulletin of the Australian Mathematical Society and prior to that he was Head of the School of Mathematics and Statistics at the University of Sydney.
Recently, in collaboration with Gus Lehrer, he published the book "Unitary Reflection Groups" (Cambridge University Press, 2009).
His previous book "The Geometry of the Classical Groups"
(Heldermann Verlag, 1992) is still in print.
First ISE Seminar for 2012 to be held on 13 January from 11:00am to 12:00pm in room 11B24
ALL WELCOME
“Symmetry properties from defining equations”
by
Tracy HUANG, ISE PhD research student
(PhD Research Proposal –Confirmation Seminar)
Faculty of Information Sciences and Engineering
University of Canberra
-oOo-
Abstract:
Symmetry analysis is a method to study differential equations (DEs), and is used in fields such as physics, engineering and geometry. Symmetry analysis gives a better chance to solve DEs, also it can extract information about solutions of DEs without solving them.
The problem is linearised by seeking symmetry vector fields, which form a Lie algebra. This method involves systematic algebraic calculation, so is well-suited to computer algebra; various packages are available to do this. The method proceeds by setting up a system of defining equations for the vector fields, which is then solved. However, integrating the defining equations is not algorithmic, and can be difficult to achieve.
My supervisor, Dr Ian Lisle, has recently developed methods for finding algebraic properties of Lie algebras of vector fields (LAVF) at level of defining equations, however, these methods have not yet implemented. Therefore, my objectives are, first, implement Lisle's methods in computer algebra system Maple as a package. Second, develop methods for allowing partial/fully integrated system, so this type of system can get benefit from our methods too. Finally, develop methods for finding "geometric" properties of LAVF at level of defining equations. All developed methods will be also implemented.
In this initial seminar, I will describe my research studies in more detail, plus show progress so far by presenting a demonstration of the package (pre-alpha version).
Speaker:
Tracy HUANG, ISE PhD research student (Maths) since Feb, 2011
Qualifications
BSc(2004), MIT(2006), MEng(2009), all at UC
Employment
2009 - 2010 Full-time river model programmer, eWater.
2006 - Present Casual tutor in ISE, Math & Stats.
Publications
Two joint papers, one in Applied Maths one in Computer Graphics