The fortnightly BRAG meeting will be held this Thursday 06/10 at 1 pm via Zoom/GP-Y801. This week we will have presentations by @Nushrath Najimuddin and @Adam Bretherton.
Zoom Link: https://qut.zoom.us/j/85315873893?pwd=bWlWci9lM2Z3RmFCdjc5bmppWG1qUT09
Password: brag@QUT (if prompted)
Nushrath’s Talk
Title: Improving supply chain operations through experimental design and predictive models
Abstract: The overall research problem that this thesis will address is how to use supply chain data to inform operations within the Australian avocado industry to improve the quality of avocados being exported internationally. In addition, we seek to determine how such models can be used to inform future experiments and data collections to improve our understanding of avocado production to benefit the industry more generally. This will include on-farm practices and appropriate practices for exporting this fruit.
Adam’s Talk
Title: Transfer Sequential Monte Carlo.
Abstract: The transferability problem arises in many different areas, where one wants to use the outcomes of a statistical model applied at some source domain to help provide additional information about
a related statistical model at some target domain. Updating prior beliefs based on data is a core tenet of Bayesian inference. In the Bayesian context, model transfer extends Bayesian updating by incorporating information from a well-known source domain into a target domain. Consider the scenario where a target domain has insufficient data to enable useful inference. Model transfer allows us to borrow information from a source domain with sufficient data to improve inference. The transferability problem then is a question of when to transfer information, which information to transfer, and how to transfer this information. This problem appears across several domains, with some solutions exploiting the underlying properties of the source model, while others create informative priors with the source information.
We introduce a novel transferability method Transfer Sequential Monte Carlo (TSMC), which is a principled statistical approach that is flexible enough to accommodate a variety of different modelling problems, provide useful inference to the underlying systems and account for negative transfer. Our method harnesses Sequential Monte Carlo (SMC) methods to incorporate tempered information from the source domain. While using widely applicable information criteria (WAIC) to choose the optimal amount of tempering to reduce negative transfer. Results from a simulation study will be presented to showcase our methods ability to transfer suitable information from source domains with varying levels of similarity to the target domain.