New Directions in Bayesian Statistics: formulation, computation and application to exemplar challenges. Bayesian statistics is a fundamental statistical and machine learning approach for density estimation, data analysis and inference. However, there remain open questions regarding the formulation of the model, the likelihood and priors, and efficient computation. This project proposes new approaches that address these issues, and applies them to two exemplar challenges: the impact of climate ch ....New Directions in Bayesian Statistics: formulation, computation and application to exemplar challenges. Bayesian statistics is a fundamental statistical and machine learning approach for density estimation, data analysis and inference. However, there remain open questions regarding the formulation of the model, the likelihood and priors, and efficient computation. This project proposes new approaches that address these issues, and applies them to two exemplar challenges: the impact of climate change on the Great Barrier Reef and better understanding neurological diseases related aging, in particular Parkinson's Disease. Read moreRead less
Random Discrete Structures: Approximations and Applications. The behaviour of many real world systems can be modelled by random discrete structures evolving over time. For example, the sizes of populations of frogs in some close patches of forests can be modelled as interacting random processes. The aim of the project is to investigate large discrete random structures that arise from real world application in areas such as biology, complex networks and insurance. The proposed project is at the i ....Random Discrete Structures: Approximations and Applications. The behaviour of many real world systems can be modelled by random discrete structures evolving over time. For example, the sizes of populations of frogs in some close patches of forests can be modelled as interacting random processes. The aim of the project is to investigate large discrete random structures that arise from real world application in areas such as biology, complex networks and insurance. The proposed project is at the interface of mathematics and 'big data' applications and so the work of the project aims to provide theoretical and heuristic underpinnings useful in the algorithms and techniques of practitioners. Understanding the applications in the project requires new, broadly applicable methods and developing such is a complementary aim.Read moreRead less
Statistical problems involving measurement errors and sparsity. The project tackles research in complex problems where the information contained in the data is sparse and corrupted by measurement errors. With the aid of modern computing methods, the project will develop new, sophisticated techniques that have applications in areas such as genomics, national security, environmental pollution, public health and nutrition.
Measure-valued analysis of stochastic populations. The project aims to develop new mathematical models and tools for the rigorous analysis of very general stochastic populations that are subject to internal competition and feedback. The proposed mathematical framework is that of measure-valued processes, a setting needed to encompass the complexity and random structure inherent in such systems. Models of this kind have real-world applications in evolutionary biology, cell kinetics and cancer res ....Measure-valued analysis of stochastic populations. The project aims to develop new mathematical models and tools for the rigorous analysis of very general stochastic populations that are subject to internal competition and feedback. The proposed mathematical framework is that of measure-valued processes, a setting needed to encompass the complexity and random structure inherent in such systems. Models of this kind have real-world applications in evolutionary biology, cell kinetics and cancer research, and are essential to our understanding of the persistence of endemic disease and of the preservation of endangered species. The results of this project are expected to provide insight into the behaviour and (in-)stabilities of complex stochastic populations, and offer guidance for their management.Read moreRead less
Discovery Early Career Researcher Award - Grant ID: DE150101044
Funder
Australian Research Council
Funding Amount
$315,000.00
Summary
New computational approaches for branching processes in population biology. Branching processes are powerful modelling tools in population biology. They describe how individuals live and reproduce according to specific probability laws, and can be used to answer a wide range of population-related questions. This project aims to develop new algorithmic methods for a tractable class of branching processes called Markovian binary trees. Following a matrix analytic approach, it will deliver new resu ....New computational approaches for branching processes in population biology. Branching processes are powerful modelling tools in population biology. They describe how individuals live and reproduce according to specific probability laws, and can be used to answer a wide range of population-related questions. This project aims to develop new algorithmic methods for a tractable class of branching processes called Markovian binary trees. Following a matrix analytic approach, it will deliver new results on the efficient estimation of model parameters, and on the effects of random environments on population dynamics. These results will be used to study significant problems in evolutionary and conservation biology, thereby establishing the relevance of the developed techniques.Read moreRead less
Stochastic majorization--minimization algorithms for data science. The changing nature of acquisition and storage data has made the process of drawing inference infeasible with traditional statistical and machine learning methods. Modern data are often acquired in real time, in an incremental nature, and are often available in too large a volume to process on conventional machinery. The project proposes to study the family of stochastic majorisation-minimisation algorithms for computation of inf ....Stochastic majorization--minimization algorithms for data science. The changing nature of acquisition and storage data has made the process of drawing inference infeasible with traditional statistical and machine learning methods. Modern data are often acquired in real time, in an incremental nature, and are often available in too large a volume to process on conventional machinery. The project proposes to study the family of stochastic majorisation-minimisation algorithms for computation of inferential quantities in an incremental manner. The proposed stochastic algorithms encompass and extend upon a wide variety of current algorithmic frameworks for fitting statistical and machine learning models, and can be used to produce feasible and practical algorithms for complex models, both current and future.
Read moreRead less
Australian Laureate Fellowships - Grant ID: FL110100003
Funder
Australian Research Council
Funding Amount
$1,814,346.00
Summary
New directions, new problems and new data types in statistical science. Statistically challenging problems today involve answering many more questions than we have data. Solving them will elucidate the causes of diseases such as cancer, and provide better security for the community. The project will develop new methods for tackling these challenging problems, taking statistical science in new directions.
Computational methods for population-size-dependent branching processes. Branching processes are the primary mathematical tool used to model populations that evolve randomly in time. Most key results in the theory are derived under the simplifying assumption that individuals reproduce and die independently of each other. However, this assumption fails in most real-life situations, in particular when the environment has limited resources or when the habitat has a restricted capacity. This project ....Computational methods for population-size-dependent branching processes. Branching processes are the primary mathematical tool used to model populations that evolve randomly in time. Most key results in the theory are derived under the simplifying assumption that individuals reproduce and die independently of each other. However, this assumption fails in most real-life situations, in particular when the environment has limited resources or when the habitat has a restricted capacity. This project aims to develop novel and effective algorithmic techniques and statistical methods for a class of branching processes with dependences. We will use these results to study significant problems in the conservation of endangered island bird populations in Oceania, and to help inform their conservation management.Read moreRead less
New universality in stochastic systems. This project aims to uncover new analyses and effects in the complex behaviour of non-linear systems with random noise. Many systems originate near an unstable equilibrium. This project will develop a new mathematical theory that establishes a universality in the way the long term effect of noise expresses itself as random initial conditions in the dynamics. It will fill gaps in Mathematics and make refinements to existing fundamental scientific laws by in ....New universality in stochastic systems. This project aims to uncover new analyses and effects in the complex behaviour of non-linear systems with random noise. Many systems originate near an unstable equilibrium. This project will develop a new mathematical theory that establishes a universality in the way the long term effect of noise expresses itself as random initial conditions in the dynamics. It will fill gaps in Mathematics and make refinements to existing fundamental scientific laws by including random initial conditions as predicted by our theory. This will advance our understanding of complex systems subjected to noise and will provide significant benefits in the scientific discoveries in Biology, Ecology, Physics and other Sciences where such systems are frequently met.Read moreRead less
Increasing the efficiency and interpretability of stepped wedge trials. Stepped wedge cluster randomised trials are increasingly being used to test interventions, across many disciplines. This project aims to develop highly efficient trial designs and new methods for the estimation of causally interpretable effects when adherence to interventions is not perfect. This project expects to generate new design types that reduce resources required to test interventions, and methods to understand how t ....Increasing the efficiency and interpretability of stepped wedge trials. Stepped wedge cluster randomised trials are increasingly being used to test interventions, across many disciplines. This project aims to develop highly efficient trial designs and new methods for the estimation of causally interpretable effects when adherence to interventions is not perfect. This project expects to generate new design types that reduce resources required to test interventions, and methods to understand how these interventions work. Expected outcomes include tools to help researchers develop cheaper and more appealing trials, tools to estimate causal effects, the methodology underpinning these tools, and new collaborations. This should provide significant benefits by allowing more interventions to be tested and understood.Read moreRead less