Mitigating bias in statistical analyses of data collected over time. This project aims to develop innovative nonparametric distribution and regression curve estimation techniques from data collected over time. These curves are key statistical tools for describing populations, but often, their estimators are inefficient when the data are massive, growing and change over time, or too restrictive when the data exhibit measurement errors and a fraction of them are equal to zero. The project expects ....Mitigating bias in statistical analyses of data collected over time. This project aims to develop innovative nonparametric distribution and regression curve estimation techniques from data collected over time. These curves are key statistical tools for describing populations, but often, their estimators are inefficient when the data are massive, growing and change over time, or too restrictive when the data exhibit measurement errors and a fraction of them are equal to zero. The project expects to develop novel, less restrictive and more realistic nonparametric curve estimation methods in these complex settings. Outcomes include new practical statistical methods and software to benefit experts in diverse fields from nutrition and epidemiology, to environmental science and digital platforms, amongst others.Read moreRead less
Discovery Early Career Researcher Award - Grant ID: DE240100144
Funder
Australian Research Council
Funding Amount
$444,447.00
Summary
Universal Model Selection Criteria for Scientific Machine Learning. This project aims to develop provably reliable universal model selection criteria to facilitate trustworthy scientific machine learning. Combining stochastic methods with an innovative geometric approach to basic statistical principles, this project expects to characterise, combine, and refine the most successful heuristics for designing and training huge models, such as deep neural networks, into a cohesive theoretical framewor ....Universal Model Selection Criteria for Scientific Machine Learning. This project aims to develop provably reliable universal model selection criteria to facilitate trustworthy scientific machine learning. Combining stochastic methods with an innovative geometric approach to basic statistical principles, this project expects to characterise, combine, and refine the most successful heuristics for designing and training huge models, such as deep neural networks, into a cohesive theoretical framework. The expected outcomes include a general toolkit for assisting neural network design at the forefront of scientific applications. This should significantly improve the quality of scientific predictions by facilitating confident adoption of deep learning methods into the pantheon of trustworthy modeling techniques. Read moreRead less