Haowei Zhao

Neural Posterior Learning for Bayesian Model Choice

This project focuses on using neural network as an approach to address the limitations in Bayesian Model Choice for models with intractable likelihoods. Traditional approaches usually rely on summaries of data that can lose important information and suffers from curse of dimensionality. The core component is to investigate deep learning architectures to understand which ones perform the best when models are overlapping and use the findings to develop an ABC methodology using neural networks embeddings as summary statistics to perform model choice via ABC. This would provide uncertainty quantification for the candidate models available for model choice.

Haowei Zhao

The University of Sydney

Haowei is completing a Bachelor of Science degree at the University of Sydney, majoring in Mathematics and Statistics. He started his career in the property industry but found his true interest in mathematics, prompting a career shift toward research. He plans to take Honours degree and PhD. Outside his studies, he enjoys exploring the nature with his family.

You may be interested in

Yichen Jiang

Cell type deconvolution methods for spatial isoform-resolution expression data

THI HAI ANH LUYEN

Clustering Time Series Using Spectral Density–Based Distances

Huateng Zhu

Limit theorems for the Curie- Weiss-Potts model

Brenton Horne

Investigating spontaneous symmetry breaking of spatial Kerr solitons in fractional spatial dimensions using Fourier spectral methods.