How is information organised in the mind? Learning structured mental representations from data. One of the biggest questions in psychology is to understand the principles that the mind uses to organise information. This project is both a search for these underlying psychological laws, and an attempt to develop new statistical technologies and mathematical tools that can be used to organise information in applied settings.
Nonparametric Machine Learning for Modern Data Analytics. This project intends to develop next-generation machine-learning methods to cope with the growing data deluge. Modern data analytics tasks need to interpret and derive values from complex, growing data. Intended outcomes of the project include new Bayesian nonparametric methods that can express arbitrary dependency amongst multiple, heterogeneous data sources with infinite model complexity, together with algorithms to perform inference an ....Nonparametric Machine Learning for Modern Data Analytics. This project intends to develop next-generation machine-learning methods to cope with the growing data deluge. Modern data analytics tasks need to interpret and derive values from complex, growing data. Intended outcomes of the project include new Bayesian nonparametric methods that can express arbitrary dependency amongst multiple, heterogeneous data sources with infinite model complexity, together with algorithms to perform inference and deduce knowledge from them; new Bayesian statistical inference for set-valued random variables that moves beyond vectors and matrices to enrich our analytics toolbox to deal with sets; and a new deterministic fast inference to meet with real-world demand.Read moreRead less
Subband centroids and deep neural networks for robust speech recognition. This project aims to improve the robustness and accuracy of automatic speech and speaker recognition systems. Though these systems work reasonably well in noise-free environments, their performance deteriorates drastically even in the presence of a small amount of noise. To overcome this problem, this project proposes a missing-feature approach for robust speech and speaker recognition. This approach is expected to make th ....Subband centroids and deep neural networks for robust speech recognition. This project aims to improve the robustness and accuracy of automatic speech and speaker recognition systems. Though these systems work reasonably well in noise-free environments, their performance deteriorates drastically even in the presence of a small amount of noise. To overcome this problem, this project proposes a missing-feature approach for robust speech and speaker recognition. This approach is expected to make the speech and speaker recognition systems less sensitive to additive background noise and make them more useful in telecommunications and business.Read moreRead less
Reconstructing proteins to explain and engineer biological diversity. The aim of this project is to develop computational methods to construct entirely new proteins. Computational reconstruction of enzymes that have been extinct for over 400 million years has revealed remarkable opportunities for biotechnological innovation. The intended outcomes are to develop bioinformatics methods to broaden the scope of ancestral protein reconstruction to include protein super-families, to establish what spe ....Reconstructing proteins to explain and engineer biological diversity. The aim of this project is to develop computational methods to construct entirely new proteins. Computational reconstruction of enzymes that have been extinct for over 400 million years has revealed remarkable opportunities for biotechnological innovation. The intended outcomes are to develop bioinformatics methods to broaden the scope of ancestral protein reconstruction to include protein super-families, to establish what specific changes led to the evolutionary success of a protein, and to re-run evolution to generate proteins that perform in conditions suitable for industrial and agricultural applications, in particular the production of hydroxylated fatty acids for bioplastics. By examining proteins from many life forms, the project plans to develop a novel bioinformatics strategy to understand their evolution and engineer new proteins for use in production of chemical commodities.Read moreRead less
Topological data analysis for enhanced modelling of the physical properties of complex micro-structured materials. The way water flows through sandstone depends on the connectivity of its pores, the balance of forces in a grain silo on the contacts between individual grains, and the impact resistance of metal foam in a car door on the arrangement of its cells. These structural properties are described mathematically by topology. Advanced three-dimensional X-ray imaging can now reveal the interna ....Topological data analysis for enhanced modelling of the physical properties of complex micro-structured materials. The way water flows through sandstone depends on the connectivity of its pores, the balance of forces in a grain silo on the contacts between individual grains, and the impact resistance of metal foam in a car door on the arrangement of its cells. These structural properties are described mathematically by topology. Advanced three-dimensional X-ray imaging can now reveal the internal detail of micro-structured materials. Recent developments in image analysis mean it is possible to compute accurate topological information from such images. This project aims to investigate how fundamental measures of shape influence the physical properties of complex materials and clarifies the mathematics that underpins these relationships.Read moreRead less
Next-generation Protein Structural comparison using Information Theory. Progress in protein structural biology relies heavily on key computational technologies, structural alignment being an indispensable one. Despite its importance the structural alignment problem has not been formulated, much less solved, in a consistent and reliable way. This project aims to rectify this by combining novel information-theoretic inference with advances in constraint optimisation and visualisation. State-of-the ....Next-generation Protein Structural comparison using Information Theory. Progress in protein structural biology relies heavily on key computational technologies, structural alignment being an indispensable one. Despite its importance the structural alignment problem has not been formulated, much less solved, in a consistent and reliable way. This project aims to rectify this by combining novel information-theoretic inference with advances in constraint optimisation and visualisation. State-of-the-art alignment methods aim to be produced for biologists to generate statistically-rigorous and biologically-trustworthy alignments, and allow them to visualise structural relationships in unprecedented ways. This project is expected to provide direct payoffs to the fields of protein science, crystallography and bioinformatics.Read moreRead less
Discovery Early Career Researcher Award - Grant ID: DE170101134
Funder
Australian Research Council
Funding Amount
$360,000.00
Summary
Feasible algorithms for big inference. This project aims to develop algorithms for computationally-intensive statistical tools to analyse Big Data. Big Data is ubiquitous in science, engineering, industry and finance, but needs special machine learning to conduct correct inferential analysis. Computational bottlenecks make many tried-and-true tools of statistical inference inadequate. This project will develop tools including false discovery rate control, heteroscedastic and robust regression an ....Feasible algorithms for big inference. This project aims to develop algorithms for computationally-intensive statistical tools to analyse Big Data. Big Data is ubiquitous in science, engineering, industry and finance, but needs special machine learning to conduct correct inferential analysis. Computational bottlenecks make many tried-and-true tools of statistical inference inadequate. This project will develop tools including false discovery rate control, heteroscedastic and robust regression and mixture models, via Big Data-appropriate optimisation and composite-likelihood estimation. It will make open, well-documented, and accessible software available for the scalable and distributable analysis of Big Data. The expected outcome is a suite of scalable algorithms to analyse Big Data.Read moreRead less
Novel dissimilarity techniques for characterising noisy spatial networks. This project will invent new and widely applicable ways of summarising fundamental characteristics of noisy spatial networks that change slightly in space or time. The techniques developed will be applied to solve important problems in two diverse applications - predicting disease spread in wildlife and protecting human biometric information.
Discovery Early Career Researcher Award - Grant ID: DE130101605
Funder
Australian Research Council
Funding Amount
$289,000.00
Summary
Composing machine learning via market mechanisms. This project aims to better understand connections between learning algorithms and markets as aggregators of information and develop new, principled techniques for combining predictions. This will improve our ability to construct systems that make predictions based on multiple, complex and structured sources of data.
Discovery Early Career Researcher Award - Grant ID: DE190100045
Funder
Australian Research Council
Funding Amount
$377,829.00
Summary
Efficient and effective analytics for real-world time series forecasting. This project aims to create efficient, effective techniques that provide accurate forecasts for heterogeneous sets of time series of varying sizes. Exploiting similarities between time series means using many related series, not larger series when building forecasts. The expected outcomes should be innovative methods that improve accuracy and allow forecasting with shorter time series. The project addresses the need to exp ....Efficient and effective analytics for real-world time series forecasting. This project aims to create efficient, effective techniques that provide accurate forecasts for heterogeneous sets of time series of varying sizes. Exploiting similarities between time series means using many related series, not larger series when building forecasts. The expected outcomes should be innovative methods that improve accuracy and allow forecasting with shorter time series. The project addresses the need to exploit properties of big data accurately in a short time frame, which is transforming many industries. This should enable more accurate and reliable forecasts across industries as diverse as retail, food manufacturing, transport, mining, tourism, energy, and technology.Read moreRead less