Uncertainty, Risk and Related Concepts in Machine Learning. Machine learning is the science of making sense of data. It does not and cannot remove all risk and uncertainty. This project proposes to study the foundations of how machine learning uses, represents and communicates risk and uncertainty. It aims to do so by finding new theoretical connections between diverse notions that have arisen in allied disciplines. These include risk, uncertainty, scoring rules and loss functions, divergences, ....Uncertainty, Risk and Related Concepts in Machine Learning. Machine learning is the science of making sense of data. It does not and cannot remove all risk and uncertainty. This project proposes to study the foundations of how machine learning uses, represents and communicates risk and uncertainty. It aims to do so by finding new theoretical connections between diverse notions that have arisen in allied disciplines. These include risk, uncertainty, scoring rules and loss functions, divergences, statistics and different ways of aggregating information. By building a more complete theoretical map it is expected that new machine learning methods will be developed, but more importantly that machine learning will be able to be better integrated into larger socio-technical systems.Read moreRead less
Large dynamic time-varying models for structural macroeconomic inference. This project aims to broaden the range of macroeconomic models that have an integrated capacity for both greater realism and efficiency in analysis. This approach will be applied to two contexts at the forefront of current macroeconomic research, the effects of noisy productivity signals on business cycles and the effects of fiscal policy shocks. Flexible macro-econometric models underpin accurate inference by economists ....Large dynamic time-varying models for structural macroeconomic inference. This project aims to broaden the range of macroeconomic models that have an integrated capacity for both greater realism and efficiency in analysis. This approach will be applied to two contexts at the forefront of current macroeconomic research, the effects of noisy productivity signals on business cycles and the effects of fiscal policy shocks. Flexible macro-econometric models underpin accurate inference by economists and policymakers and the project outputs should provide widespread and significant benefits by improving policy and boosting Australia’s comparative advantage.Read moreRead less
Bayesian inference for psychological theories with intractable likelihood. This project pursues breakthroughs which allow important questions of basic and applied science to be addressed using mathematical theories from cognitive psychology. Advances are made through an interdisciplinary effort, combining recent developments in econometric and statistical methods and cognitive science. The outcomes will advance knowledge and open up new avenues for applied research in important aspects of psych .... Bayesian inference for psychological theories with intractable likelihood. This project pursues breakthroughs which allow important questions of basic and applied science to be addressed using mathematical theories from cognitive psychology. Advances are made through an interdisciplinary effort, combining recent developments in econometric and statistical methods and cognitive science. The outcomes will advance knowledge and open up new avenues for applied research in important aspects of psychology. This research will result in new methods available to the wider scientific community which open up new horizons for understanding basic cognition, and human behavior in many domains. Read moreRead less
Diversification failures and improved measures of uncertainty. The project aims to develop new statistical tools, applicable when the conventional paradigm that diversification reduces risk fails and when textbook approaches to risk quantification severely under-report risk. The new tools enhance our capacity to build and manage natural, social and human-made systems in uncertain environments. Our effective response to many threats including financial crises and natural events, depends on this c ....Diversification failures and improved measures of uncertainty. The project aims to develop new statistical tools, applicable when the conventional paradigm that diversification reduces risk fails and when textbook approaches to risk quantification severely under-report risk. The new tools enhance our capacity to build and manage natural, social and human-made systems in uncertain environments. Our effective response to many threats including financial crises and natural events, depends on this capacity. Thus, the expected benefits in the form of more reliable and robust risk analytics will accrue when they are most needed.Read moreRead less
Understanding macroeconomic fluctuations with unobserved networks. Whilst empirical evidence suggests that firm-level shocks can have large aggregate effects, via network connections, macroeconomic policies have mostly an aggregate nature. This project aims to build a new framework to disentangle aggregate shocks from shocks to individual units. The major innovations are i) to infer the network from the data and ii) to jointly estimate aggregate factors and network effects. Expected outcomes are ....Understanding macroeconomic fluctuations with unobserved networks. Whilst empirical evidence suggests that firm-level shocks can have large aggregate effects, via network connections, macroeconomic policies have mostly an aggregate nature. This project aims to build a new framework to disentangle aggregate shocks from shocks to individual units. The major innovations are i) to infer the network from the data and ii) to jointly estimate aggregate factors and network effects. Expected outcomes are i) measures of systemic risk and ii) a theoretical framework to study the optimality of aggregate versus sectoral stabilization policies. Benefits include a better understanding of macroeconomic fluctuations in Australia and proposed economic policies to mitigate large and persistent declines in employment and GDP.Read moreRead less
Statistical Analysis of State-Dependent Government Spending Multipliers. This project aims to provide a new statistical analysis of the government spending multiplier by acknowledging that government spending is the sum of sectoral spending which has heterogeneous effects on the economy. An added complication is that the multiplier can also be state-dependent, meaning that its magnitude can differ across recessions and expansions. Expected outcomes of this project include a better understanding ....Statistical Analysis of State-Dependent Government Spending Multipliers. This project aims to provide a new statistical analysis of the government spending multiplier by acknowledging that government spending is the sum of sectoral spending which has heterogeneous effects on the economy. An added complication is that the multiplier can also be state-dependent, meaning that its magnitude can differ across recessions and expansions. Expected outcomes of this project include a better understanding of the components of the multiplier by novel decomposition and the development of a new statistical test for the state-dependency of the multiplier. This should provide significant benefits to researchers by bringing in new tools and insights and to policymakers by providing timely guidance on fiscal policies.Read moreRead less
Flexible models and methods for cognitive model-based decision-making. This project aims to apply mathematical decision models to important questions of basic and applied science. Advances will be pursued through an interdisciplinary effort, combining recent developments in econometric and statistical methods, cognitive science and computing. The expected outcomes will bring a proven and powerful approach to a new range of questions investigating psychological aspects of choices about health c ....Flexible models and methods for cognitive model-based decision-making. This project aims to apply mathematical decision models to important questions of basic and applied science. Advances will be pursued through an interdisciplinary effort, combining recent developments in econometric and statistical methods, cognitive science and computing. The expected outcomes will bring a proven and powerful approach to a new range of questions investigating psychological aspects of choices about health care and consumer purchases. This project will provide significant benefits to the wider scientific community to understand basic cognition, and human behaviour in many domains.Read moreRead less
Deep learning based time series modeling and financial forecasting. This project pursues breakthroughs in time series modelling and develops novel statistical models and inference techniques, with a focus on modelling of financial time series data. The advances will be achieved through interdisciplinary research, combining recent advances in machine learning, Bayesian computation, financial econometrics and the increasing availability of Big Data. The outcomes will provide a new range of proven ....Deep learning based time series modeling and financial forecasting. This project pursues breakthroughs in time series modelling and develops novel statistical models and inference techniques, with a focus on modelling of financial time series data. The advances will be achieved through interdisciplinary research, combining recent advances in machine learning, Bayesian computation, financial econometrics and the increasing availability of Big Data. The outcomes will provide a new range of proven and powerful approaches for analysing time series and understanding time effects. The methodologies developed will lead to a greater accuracy in financial forecasting and risk management, and open up new horizons for the wider scientific community to analyse time series data.Read moreRead less