Open Research Repository
The Open Research Repository is the University’s online open access repository for collecting, maintaining and disseminating the scholarly output of the University.
Contribute to the Open Research Repository
Communities in DSpace
Select a community to browse its collections.
Recent Submissions
Machine Learning Assisted Signal Enhancement
(2025) Yan, Longfei Felix
In an era of abundant signals, the ability to obtain desired signals while rejecting undesired ones has become increasingly crucial. Often, the desired signals are mixed with interference or contaminated by noise. Signal enhancement techniques play a vital role in performing tasks such as signal separation, extraction, and suppression. This thesis addresses critical challenges in signal enhancement tasks by harnessing the power of machine learning techniques. Firstly, we propose a novel independence criterion called the Finite Basis Independence Criterion (FBIC). This criterion estimates the Hirschfeld-Gebelein-Rényi maximum correlation coefficient between tested variables and is based on mapping functions from a subspace of finite basis. FBIC detects dependence between variables in linear time and outperforms more computationally expensive kernel-based counterparts. Extensive testing in Independent Component Analysis benchmarks demonstrates its potential for various signal separation applications. Secondly, we conduct a comprehensive robustness analysis of a popular signal enhancement approach: fixed beamforming based on first-order linear Differential Microphone Arrays (DMAs). We demonstrate that both bounded and unbounded phase errors of microphones can affect the mainlobe orientation of the beamformer. Analytically derived white noise gain thresholds indicate when mainlobe misorientation occurs. Through rigorous mathematical derivations, we prove that a higher number of microphones and increased spacing between microphones contribute to the robustness of the beamformer. This work provides practical guidelines for designing robust first-order linear DMAs. Thirdly, we propose a neural network model to optimize both the geometry and spatial filter of linear DMAs. The model consists of two feed forward neural networks and is trained end-to-end. The signals enhanced by this model exhibit superior quality compared to those obtained from conventional DMA approaches. Furthermore, the model offers flexibility in controlling the tradeoff between different performance metrics, allowing for customized optimization. Lastly, we extend the neural network model to a general framework that allows optimization of microphone arrays of any geometry, along with their spatial filters. This model employs ResNets and augmented Lagrangian techniques to achieve state-of-the-art frequency-invariant fixed beamforming performance. We showcase our performance in linear, circular, and concentric circular microphone arrays. Moreover, our findings challenge the conventional belief that concentric circular arrays require multiple rings, as we demonstrate that good performance can be achieved with only one ring. Overall, this thesis contributes novel techniques and insights to the field of signal enhancement, leveraging machine learning approaches to address key challenges. The proposed criteria, guidelines and models have the potential to advance various signal separation applications and enhance the overall quality of processed signals.
Socioeconomic factors, medication subsidisation and the use of preventativecardiovascular disease medications in Australia.
(Swansea University, UK, 2022) Paige, Ellie; Banks, Emily; Agostino, Jason; Brieger, David; Page, Karen; Joshy, Grace; Barrett, Eden; Korda, Rosemary
Objectives
Cardiovascular disease (CVD) events are highly preventable through appropriate treatment and disproportionally affect socioeconomically disadvantaged individuals. This study quantified the relationship of socioeconomic factors to dispensing and persistent use of lipid- and blood pressure-lowering medication following hospital admission for a major CVD event (myocardial infarction, ischaemic stroke/transient ischaemic attack).
Approach
Data from 8,285 people with major CVD events aged ≥45 years from the Australian 45 and Up Study with linked medication dispensing data were used. Modified Poisson regression was used to estimate relative risks (RRs) for combined lipid- and blood pressure-lowering dispensing at three-months following hospital discharge and for 12-month persistent use, in relation to education, income, and level of medication subsidisation.
Results
Overall, 56% were dispensed guideline-recommended medications at three months and 37% persistently used them across 12 months. After adjusting for demographic factors, type of CVD and history of CVD hospitalisation, RRs for lowest (no educational qualifications) compared to highest education level (university degree) were 1.14 (95% CI: 1.06, 1.22) for medication dispensing and 1.15 (1.02, 1.29) for persistent medication use; 1.14 (1.06, 1.22) and 1.17 (1.04, 1.32) respectively for lowest (<$20,000) versus highest (≥$70,000) household pre-tax income; and 1.25 (1.17, 1.33) and 1.28 (1.15, 1.43) respectively for those receiving highest versus lowest subsidisation. There was little to no evidence of a relationship of income and education to medication use after adjustment for medication subsidisation.
Conclusions
While preventive medication use is sub-optimal, subsidisation is substantially associated with increased use and accounts for most of the relationship with socioeconomic position, suggesting medication subsidy schemes are working in the intended direction
Faultlines and contact zones: A new forum for Migration Studies
(Oxford University Press, 2018) Gamlen, Alan; Betts, Alexander; Delano, Alexandra; Lacroix, Thomas; Paoletti, Somerville; Sigona, Nando; Vargas-Silva, Carlos
Migration studies is entering a new era. The intellectual roots of the field stretch back at least to the nineteenth century, since which time it has focused on the drivers of human mobility and the processes of adaptation that follow. But never before has it drawn such sustained attention from so many researchers across such a broad range of backgrounds. Migration is becoming an increasingly visible and important element of human experience, and more than ever before, migration studies is becoming a distinctive and integrated field of scholarship, with its own approaches and institutions. This journal is being established to help galvanize the field still further. It aims to provide an interdisciplinary forum for leading research that develops the core concepts, data, and methods needed by migration scholars in the twenty-first century and beyond. It has several specific priorities.
Incidence and outcomes of neonatal acute kidney injury (AWAKEN): a multicentre, multinational, observational cohort study
(Elsevier B.V, 2017) Jetton, Jennifer; Boohaker, L; Sethi, Sidharth; Wazir, Sanjay; Rohatgi, S; Soranno, D; Chishti, A; Woroniecki, R; Mammen, C; Swanson, J; Kent, Alison; Fletcher, Jeffery
Background Findings from single-centre studies suggest that neonatal acute kidney injury (AKI) is associated with poor outcomes. However, because of the small sample size of those studies, few inferences can been made regarding the independent associations between AKI, mortality, and hospital length of stay. We aimed to establish whether neonatal AKI is independently associated with increased mortality and length of hospital stay. Methods We did this multicentre, multinational, retrospective cohort study of critically ill neonates admitted to 24 participating neonatal intensive care units (NICUs) in four countries (Australia, Canada, India, USA) between Jan 1 and March 31, 2014. We included infants born or admitted to a level 2 or 3 NICU and those who received intravenous fluids for at least 48 h. Exclusion criteria were admission at age 14 days or older, congenital heart disease requiring surgical repair within 7 days of life, lethal chromosomal anomaly, death within 48 h of admission, inability to determine AKI status, or severe congenital kidney abnormalities. We defined AKI as an increase in serum creatinine of 0·3 mg/dL or more (≥26·5 μmol/L) or 50% or more from the previous lowest value, or a urinary output of less than 1 mL/kg per h on postnatal days 2–7. We used logistic regression to calculate crude odds ratios (ORs) and associated 95% CIs for the association between AKI and likelihood of death. We used linear regression to calculate the crude parameter estimates and associated 95% CIs for the association between AKI and length of hospital stay. Multivariable logistic and linear regression models were run to account for potential confounding variables. We additionally created regression models stratified by gestational age groups (22 weeks to <29 weeks, 29 weeks to <36 weeks, and ≥36 weeks). This study is registered with ClinicalTrials.gov, number NCT02443389. Findings We enrolled 2162 infants, of whom 2022 (94%) had data to ascertain AKI status. 605 (30%) infants had AKI. Incidence of AKI varied by gestational age group, occurring in 131 (48%) of 273 of patients born at 22 weeks to less than 29 weeks, 168 (18%) of 916 patients born at 29 weeks to less than 36 weeks, and 306 (37%) of 833 patients born at 36 weeks or older. Infants with AKI had higher mortality than those without AKI (59 [10%] of 605 vs 20 [1%] of 1417 infants; p<0·0001), and longer length of hospital stay (median 23 days [IQR 10–61] vs 19 days [9–36]; p<0·0001). These findings were confirmed in both crude analysis of mortality (OR 7·5, 95% CI 4·5–12·7; p<0·0001 for AKI vs no AKI) and length of stay (parameter estimate 14·9 days, 95% CI 11·6–18·1; p<0·0001) and analysis adjusted for multiple confounding factors (adjusted OR 4·6, 95% CI 2·5–8·3; p<0·0001 and adjusted parameter estimate 8·8 days, 95% CI 6·1–11·5; p<0·0001, respectively). Interpretation Neonatal AKI is a common and independent risk factor for mortality and increased length of hospital stay. These data suggest that AKI might have a similar effect in neonates as in paediatric and adult patients. Strategies designed to prevent AKI and treatments to reduce the burden of AKI, including renal support devices designed for neonates, are greatly needed to improve the outcomes of these vulnerable infants.
The epidemiology of tuberculosis in the Australia Capital Territory, 2006-2015
(National Centre for Disease Control, 2017) Jones, Belinda; Johnston, Vanessa; Appuhamy, Ranil; Kaczmarek , Marlena; Hurwitz, Mark
Aim
To review the epidemiology of tuberculosis (TB) in the Australian Capital Territory (ACT) over a 10 year period. Methods: A retrospective analysis of the ACT TB notif ication data from 1 January 2006 to 31 December 2015 was conducted. Results
Over the 10 year study period there were 171 TB notifications in the ACT, with an increasing trend in the number of notifications over time. The median age of cases was 36 years (range 14 to 91 years) and 53.8% of cases were male. Most TB cases (84.2%) were born overseas. Among Australian-born cases the most common risk factor for acquiring TB was close/household contact with a known case of TB (30.8%). The most common risk factor in the overseas-born population was past travel or residence in a high-risk country (86.9%). Of all the TB cases notif ied, 82.4% successfully completed treatment.
Conclusion
There was an increasing trend in the number of TB notifications in the ACT over the study period. The highest rate of TB notifications remained in the overseas-born population; with other studies suggesting this is commonly due to reactivation of latent tuberculosis infection (LTBI). As Australia starts working towards TB elimination, options for the screening and management of LTBI, especially in high risk populations, need to be explored.