CIC seminar June 2018
Use of cross-sectoral data linkage to predict high-rate offenders in Western Australia
Presented by Associate Professor Anna Ferrante
Studies have repeatedly found that a small number of offenders account for a disproportionate amount of crime. High-rate, persistent offenders (so-called ‘chronic’ or ‘prolific’ offenders) have a major impact on local crime rates and public perceptions of safety, and place a substantial financial and social burden on communities. Reducing (ideally, preventing) high rates of offending is a highly desired criminal justice objective.
Using population-level administrative data, our study identifies ‘prolific’ offenders in WA and describes their demographic and crime profiles. The official criminal records of all offenders born in WA between 1980 and 1995 were linked to administrative records from health, education and child protection databases. Linked data on families (parents and siblings) were also included. Using this information, the study identified factors that distinguish between prolific and non-prolific offenders. The study then examined whether correlates of prolific offending were similar between a) male and female offenders, and b) Indigenous and non-Indigenous offenders. The results and implications of the study will be explained in more detail.
Anna Ferrante has a background in both criminology and record linkage with a degree in Pure Mathematics. Anna worked in IT as an applications programmer before gravitating toward academia and research.
As a Computer Programmer/Analyst, Anna specialised in record linkage (RL) and statistical applications. In the early 1990s, she used RL to establish a Road Injury Database for the School of Population Health, UWA. Following this, she set up data linkage infrastructure within the WA justice system - the Integrated Numerical Offender Identification System or INOIS (Ferrante, 1993). The development of INOIS enabled research into ‘criminal careers’, recidivism patterns, and the evaluation of criminal justice programmes.
As an academic, Anna has conducted criminological research and published papers on many issues including drugs and crime, driving and traffic related crime, Aboriginal justice issues, domestic violence, juvenile justice and criminal careers. Anna has co-authored two books – one of the measurement of domestic violence and the other on the over-representation of Aboriginal people in the criminal justice system.
In 2009 Anna was seconded to Curtin University to establish the Centre for Data Linkage (CDL). The CDL is part of the Population Health Research Network (PHRN) - an NCRIS funded initiative to establish data linkage capabilities across Australia. Currently Anna is the Deputy Head of the new Research and Data Analytics Hub located within the Faculty of Health Sciences at Curtin.
CIC Seminar May 2018
The Fast Multipole Boundary Element Method for Large-Scale Modelling in Computational Acoustics
Dr Daniel Wilkes
The Fast Multipole Boundary Element Method (FMBEM) is a numerical method which allows for the computational modelling of large-scale wave scattering or radiation problems (acoustics, elastodynamics, and electromagnetics) with significantly reduced computational resources.
This presentation will provide a general overview of the FMBEM algorithm and present numerical results for large-scale modelling problems in underwater acoustics, elastodynamics, and acoustic coupled fluid-structure interaction problems.
A simple algorithm for small-scale parallelisation (<64 cores) of the FMBEM will also be discussed
Daniel Wilkes is a research fellow at the Centre of Marine Science and Technology (CMST) at Curtin University's Department of Applied Physics. Daniel's research work focuses on the development of fast algorithms for computational modelling in underwater acoustics for a range of applications including large-scale or high frequency acoustic scattering/sound radiation, target strength modelling, sound radiation from pile driving and modal analysis of vibrating structures.
CIC Special Guest Seminar March 2018
Caches all the way down: Infrastructure for Data Science
Prof David Abramson
The rise of big data science has created new demands for modern computer systems. While floating performance has driven computer architecture and system design for the past few decades, there is renewed interest in the speed at which data can be ingested and processed. Early exemplars such as Gordon, the NSF funded system at the San Diego Supercomputing Centre, shifted the focus from pure floating-point performance to memory and IO rates.
At the University of Queensland we have continued this trend with the design of FlashLite, a parallel cluster equipped with large amounts of main memory, flash disk, and a distributed shared memory system (ScaleMP’s vSMP). This allows applications to place data “close” to the processor, enhancing processing speeds. Further, we have built a geographically distributed multi-tier hierarchical data fabric called MeDiCI, which provides an abstraction of very large data stores across the metropolitan area. MeDiCI leverages industry solutions such as IBM’s Spectrum Scale and SGI’s DMF platforms.
Caching underpins both FlashLite and MeDiCI. In this presentation I will describe the design decisions and illustrate some early application studies that benefit from the approach. I will also highlight some of the challenges that need to be solved for this approach to become mainstream.
David Abramsonhas been involved in computer architecture and high performance computing research since 1979. He has held appointments at Griffith University, CSIRO, RMIT and Monash University and prior to joining University of Queensland, he was the Director of the Monash e-Education Centre, Science Director of the Monash e-Research Centre, and a Professor of Computer Science in the Faculty of Information Technology at Monash. From 2007 to 2011 he was an Australian Research Council Professorial Fellow.
David has expertise in High Performance Computing, distributed and parallel computing, computer architecture and software engineering. He has produced in excess of 200 research publications, and some of his work has also been integrated in commercial products. One of these, Nimrod, has been used widely in research and academia globally, and is also available as a commercial product, called EnFuzion, from Axceleon. His world-leading work in parallel debugging is sold and marketed by Cray Inc, one of the world’s leading supercomputing vendors, as a product called ccdb. David is a Fellow of the Association for Computing Machinery (ACM), the Institute of Electrical and Electronic Engineers (IEEE), the Australian Academy of Technology and Engineering (ATSE), and the Australian Computer Society (ACS). He is currently a visiting Professor in the Oxford e-Research Centre at the University of Oxford.
CIC Seminar March 2018
Recent Splitting Schemes for the Incompressible Navier-Stokes Equations
Professor Peter Minev
Two different approaches to the time discretization of the incompressible Navier-Stokes equation will be discussed in this talk. The first approach relies on a particular perturbation of the continuity equation and results in a technique for incompressible flow that requires the solution of one dimensional problems only. These problems can be solved with a tridiagonal direct solver on a massive parallel cluster with a Schur complement technique. The accuracy of this class of schemes is fully comparable to the accuracy of the classical projection methods for incompressible flow.
In the second approach the artificial compressibility method for approximation of the incompressible Navier-Stokes equations is generalized. It allows for the construction of schemes of any order in time that require the solution of a fixed number of vectorial parabolic problems, depending only on the desired order of the scheme. This approach has several advantages in comparison to the traditional projection schemes widely used for the unsteady Navier-Stokes equations. The accuracy and stability of the resulting schemes will be demonstrated on examples with manufactured solutions.
Peter Minev received his PhD degree in applied mathematics from the University of Sofia in 1991. Since 2004 he has been a full professor in applied mathematics at the University of Alberta, Canada.
He has authored or co-authored more than seventy papers in refereed journals and conference proceedings and is the advisor of twenty six graduate students and postdoctoral fellows. His general areas of interest include numerical analysis of PDEs, computational fluid dynamics and MHD, fluid mechanics and multiscale methods. Peter is a member of the Advisory Board of International Journal for Numerical Methods in Fluids, and a member of the Editorial Board of International Journal for Numerical Analysis and Modelling.
When: 14 March 2018, 9am
Where: B211:230, Curtin University
RSVP: Registration will be via Eventbrite
CIC Seminar November 2017
Professor Maciej Paszynski presents
Fast and Smooth simulations of time-dependent problems
We present fast and smooth simulations of time dependent-problems, in one, two and three dimensions. We discretize the time axis, by introducing several time steps. Theus the time-dependent problem is transformed into a sequence of the stationary problems to be solved at particular time steps.
What makes the simulations smooth is the fact that we use higher order B-splines basis functions for spatial discretization at every time step. We approximate the solution at every time step with a linear combination of tensor product B-spline basis functions that span over the regular computational mesh. What makes the simulations fast, is the fact that we have the so-called alternating directions algorithm which solves these problems in every time step with the highest possible linear computational cost.
Our simulations can be performed on a multi-core laptop, and they will run for around 30 minutes. In our simulations, we can use both so-called explicit and implicit schemes. This means that we can increase the accuracy of a simulation by increasing arbitrarily the number of B-spline basis functions used.
We show how to apply our method for time-dependent simulations of several physical phenomena, including the heat transfer, non-linear flow in heterogeneous media, elastic wave propagation problem, tumor growth simulations, and the simulation of the pollutant from a chimney. We present movies generated from the performed simulations.
Maciej Paszynski is a Full Professor of Computer Science at Department of Computer Science at AGH University, Krakow, Poland. He did his PhD in Mathematics with Applications to Computer Science from Jagiellonian University, Krakow, Poland from 2009 to 2013. He is a former postdoc of Prof. Leszek Demkowicz from the Institute for Computational and Engineering Sciences (ICES), The University of Texas at Austin (UT).
Maciej collaborates with Prof Victor Calo from Curtin University, Prof Leszek from ICES, UT, Prof David Pardo from Basque Center for Applied Mathematics (BCAM), Bilbao, Spain, Prof Keshav Pingali from ICES, UT, Prof Ignacio Muga from The Catholic University of Valparaiso, Chile, Pro. Frederick Valentine from LNCC, Petropolis, Brasil and Prof Rafael Montenegro-Armass from The University of Las Palmas de Gran Canaria.
His research interest includes fast solvers for mesh-based computations. He co-authored over 40 papers in indexed journals and gave over 100 presentations.
CIC Seminar October 2017
Professor Dirk Ifenthaler presents Can Data and Analytics Support Learning and Teaching?
Remarkable repertoires of computer-based applications and systems have been developed for supporting learning and teaching. Currently, promising learning analytics applications are being developed which utilise data produced through these computer-based applications and systems. Such learner generated data and other relevant information may be further used to personalise and continuously adapt these learning environments. Accordingly, learning analytics emphasise insights and responses to real-time learning processes based on educational information from digital learning environments, administrative systems, and social platforms. Learners may benefit from learning analytics through optimised learning pathways, personalised interventions, and real-time scaffolds. Further, learning analytics provide teachers detailed analysis and monitoring on the individual student level, allowing to identify particularly instable factors, like motivation or attention losses, before they occur.
Along theoretical foundations and empirical evidence from cognitive psychology as well as the learning sciences, this presentation will explore the readiness and promising opportunities of the emerging learning analytics field.
Professor Dirk Ifenthaler’s research focuses on the intersection of cognitive psychology, educational technology, data analytics, and organisational learning – see Dirk’s website for a full list of scholarly outcomes at www.ifenthaler.info. Professor Ifenthaler who currently works at the University of Mannheim is the Editor-in-Chief of the Springer journal Technology, Knowledge and Learning (http://www.springer.com/10758).
The presentation will be followed by afternoon tea and networking opportunity.
CIC Visitor Seminar – September 2017
Professor Rainer Schnell presents
Recent Developments in Bloom Filter-based Methods for Privacy Preserving Record Linkage
Combining data of the same unit using multiple databases is becoming increasingly popular in official statistics, the social sciences, medical research, criminology and other quantitative research fields. If the data entities are natural persons, linking different data sources is often restricted to encrypted personal identifiers when no unique personal linkage key (PID) is available. In this setting, techniques of the very active research field of Privacy Preserving Record Linkage (PPRL) have to be used.
Recently, Bloom Filters for PPRL applications have become increasingly popular, due to their similarity-preserving properties which enable the use of similarity threshold-based linkage techniques. Since these properties could lead to security vulnerabilities, the cryptographic properties of Bloom Filters are of special interest.
Previous research has shown that the straight application of Bloom Filters has a non-zero re-identification risk. Some recently developed techniques for attacking Bloom Filter encodings will be described shortly. We will present new results on new methods to make attacks on Bloom Filters increasingly difficult. These computationally simple algorithms modify the identifiers by different cryptographic diffusion techniques, among these are applications of Cellular Automata and Markov Chains. The presentation will demonstrate these new algorithms and show their performance concerning precision and recall in large databases.
Rainer Schnell holds the chair for Research Methodology in the Social Sciences at the University of Duisburg-Essen, Germany. He is former Director of the Centre for Comparative Surveys at City University London and graduated with a postdoctoral degree in Research Methodology at the University of Mannheim in 1996. His research focuses on non-sampling errors, applied sampling, census operations and privacy preserving record linkage. Professor Schnell founded the German Record Linkage Centre and was founding editor of the journal "Survey Research Methods". He is the author of books on Statistical Graphics (1994), Nonresponse (1997), Survey Methodology (2012), Research Methodology (10th ed. 2013) and of more than 30 papers on Privacy Preserving Record Linkage.
The presentation will be followed by afternoon tea and networking.
CIC Seminar September 2017
Professor Vishnu Pareek presents Modelling of Pyrolysis Process
Pyrolysis is an important thermochemical conversion process for the conversion of biomass into useful commodities. In this presentation, Professor Pareek will discuss the fundamentals of the pyrolysis process with special emphasis on the multi-scale aspects of its modelling. A number of case-studies for different scales of modelling will also be discussed.
Vishnu Pareek is a Professor of Chemical Engineering at Curtin University. He has varied research interests encompassing both fundamental and applied nature research but his focus is in the area of computational fluid dynamics (CFD) and multiphase modelling.
He is currently also serving as the Head of School for the School of Chemical and Petroleum Engineering.
CIC Seminar July 2017
Speaker: Dr. Andrea Bedini
Title: The Blockchain - A gentle introduction.
RSVP: Please register via Eventbrite by COB Monday 10 July 2017
In his presentation, Dr Bedini will offer a gentle introduction to the concepts behind the blockchain, answering the three big questions that arise with any new technology: What is it? How does it work? And why do we need it?
The blockchain is the disruptive technology at the base of Bitcoin, the world's first decentralised cryptocurrency. It has received media attention and the public discussion has some enthusiasts claiming that the blockchain is the biggest invention since the emergence of the Internet.
The blockchain's ground-breaking innovation lies in allowing people to safely conduct transactions electronically without the requirement of a trusted third-party. From its first application in the financial industry, this technology is now finding applications in many other industries such as real-estate, law, energy, and even music.
Andrea Bedini obtained his PhD in theoretical physics from Milan University and has been a Research Fellow at the School of Mathematics and Statistics of the University of Melbourne. His research included mathematical combinatorics, statistical physics, genomics and traffic flow modelling, and was supported by the ARC Centre of Excellence for Mathematics and Statistics of Complex (MASCOS) and the ARC Centre of Excellence for Mathematical and Statistical Frontiers Systems (ACEMS). Andrea is also an experienced software developer and has published scientific software packages to complement his academic publications.
He is one of the maintainers of PyTables, a Python library for managing large datasets. After 5 years in research, Andrea has decided that developing tools for research might be more exciting than research itself, and someone else should do the grant-writing. He has now joined the Curtin Institute for Computation as a data scientist/computational specialist assisting the Curtin Business School.
CIC Seminar June 2017
Weapons of Maths Construction: how a combination of maths theory and algorithms conquers inequalities and constructs optimal solutions.
Prof. Song Wang
RSVP: Please register via Eventbrite at this link by COB Monday 12 June 2017
Professor Song Wang received a B.Sc. from Wuhan University of Hydraulic and Electric Engineering, China, in 1982 and a Ph.D. in Numerical Analysis from Trinity College Dublin, Ireland, in 1989. He worked with a Dublin-based hi-tech company -Tritech Ltd., The University of New South Wales, Curtin University of Technology and The University of Western Australia, before returning to Curtin University in July 2014 as Professor and Head of Department of Mathematics & Statistics. His research interests include numerical solution of PDEs, numerical optimization and optimal control, optimum design and computational finance. He has published many research papers in these areas and his early work on numerical solution of semiconductor device equations of the drift-diffusion model has been featured in a review paper and a book chapter published respectively in a top journal Reports on Progress in Physics and a volume of the prestigious book series Handbook of Numerical Analysis. He is currently on the editorial boards of several international journals.