Thang Bui

I'm a lecturer (equivalent to tenure-track Assistant Professor) in Machine Learning and Data Science at the University of Sydney, Australia.
I'm taking leave to spend time at Uber AI, working on probabilistic modelling, approximate inference, continual learning, model-based reinforcement learning, and bandits.

I completed my doctoral training at the Cambridge Machine Learning group, supervised by Richard Turner and advised by Carl Rasmussen. I worked on Gaussian process models and generic approximation methods.

Preprints

Connecting the Thermodynamic Variational Objective and Annealed Importance Sampling: Thang Bui
note
Variational autoregressive Gaussian processes for continual learning: Sanyam Kapoor, Theofanis Karaletsos, Thang Bui
https://arxiv.org/abs/2006.05468
Partitioned Variational Inference: A unified framework encompassing federated and continual learning: Thang Bui, Cuong Nguyen, Siddharth Swaroop, Rich Turner
https://arxiv.org/abs/1811.11206

Conference and journal papers

Hierarchical Gaussian process priors for Bayesian neural network weights: Theofanis Karaletsos and Thang Bui
NeurIPS 2020
arxiv
Variational Continual Learning: Cuong Nguyen, Yingzhen Li, Thang Bui, and Rich Turner,
ICLR 2018
arxiv
Neural graph learning: Training neural networks using graphs: Thang Bui, Sujith Ravi, and Vivek Ramavajjala
WSDM 2018
arxiv
A Unifying Framework for Sparse Gaussian Process Approximation using Power Expectation Propagation: Thang Bui, Josiah Yan, Rich Turner
JMLR 2017
arxiv code
Streaming Sparse Gaussian Process Approximations: Thang Bui, Cuong Nguyen, and Rich Turner
NIPS 2017
arxiv code
Deep Gaussian Processes for Regression using Approximate Expectation Propagation: Thang Bui, José Miguel Hernández-Lobato, Yingzhen Li, Daniel Hernández-Lobato, and Rich Turner
ICML 2016
paper code
Black-box alpha-divergence minimization: José Miguel Hernández-Lobato, Yingzhen Li, Mark Rowland, Daniel Hernández-Lobato, Thang Bui, and Rich Turner
ICML 2016
paper code
Learning stationary time series using Gaussian processes with nonparametric kernels: Felipe Tobar, Thang Bui, and Rich Turner
NIPS 2015 (Spotlight, acceptance rate = 3.6%)
paper
Tree-structured Gaussian process approximations: Thang Bui and Rich Turner
NIPS 2014 (Spotlight, acceptance rate = 3.6%)
paper code

Thesis

Efficient Deterministic Approximate Bayesian Inference for Gaussian Process Models: PhD thesis
University of Cambridge
pdf
typos in the print version

Workshop papers

Partitioned variational inference for federated Bayesian deep learning: Thang Bui, Cuong Nguyen, Siddharth Swaroop, and Rich Turner
NeurIPS Bayesian Deep Learning Workshop, 2018
Understanding and improving variational continual learning: Siddharth Swaroop, Cuong Nguyen, Thang Bui, and Rich Turner
NeurIPS Continual Learning Workshop, 2018
Natural variational continual learning: Hanna Tseran, Emtiyaz Khan, Tatsuya Harada and Thang Bui
NeurIPS Continual Learning Workshop, 2018
Variational continual learning for deep models: Cuong Nguyen, Yingzhen Li, Thang Bui, and Rich Turner
NIPS Bayesian Deep Learning Workshop, 2017
Online variational Bayesian inference: Algorithms for sparse Gaussian processes and theoretical bounds: Cuong Nguyen, Thang Bui, Yingzhen Li, and Rich Turner
ICML Time Series Workshop, 2017
Importance weighted autoencoders with random neural network parameters: Daniel Hernández-Lobato, Thang Bui, José Miguel Hernández-Lobato, Yingzhen Li, and Rich Turner
NIPS Workshop on Bayesian Deep Learning, 2016
Black-box alpha divergence for generative models: Thang Bui, Daniel Hernández-Lobato, José Miguel Hernández-Lobato, Yingzhen Li, and Rich Turner
NIPS Workshop on Advances in Approximate Bayesian Inference, 2016
Circular Pseudo-point approximations for scaling Gaussian processes: Will Tebbutt, Thang Bui and Rich Turner
NIPS Workshop on Advances in Approximate Bayesian Inference, 2016
Bayesian Gaussian process state space models via Power-EP: Thang Bui, Carl Rasmussen and Rich Turner
ICML Workshop on Data efficient Machine Learning, 2016
Training deep Gaussian processes using stochastic expectation propagation and probabilistic backpropagation: Thang Bui, José Miguel Hernández-Lobato, Yingzhen Li, Daniel Hernández-Lobato and Rich Turner
NIPS Workshop on Advances in Approximate Bayesian Inference, 2015
Stochastic variational inference for Gaussian process latent variable models using back constraints: Thang Bui and Rich Turner
NIPS Workshop on Black Box Learning and Inference, 2015
Black-box alpha-divergence minimisation: José Miguel Hernández-Lobato, Yingzhen Li, Daniel Hernández-Lobato, Thang Bui and Rich Turner
NIPS Workshops on Advances in Approximate Bayesian Inference and Black Box Learning and Inference, 2015.
Stochastic expectation propagation for large scale Gaussian process classification: Daniel Hernández-Lobato, José Miguel Hernández-Lobato, Yingzhen Li, Thang Bui and Rich Turner
NIPS Workshop on Advances in Approximate Bayesian Inference, 2015.
Design of covariance functions using inter-domain inducing variables: Felipe Tobar, Thang Bui and Rich Turner
NIPS Workshop on Time Series, 2015 Best paper prize

Misc

Sparse Approximations for Non-Conjugate Gaussian Process Regressions: Thang Bui and Rich Turner
report

Other things

Reviewer for JMLR (2016, 2017, 2018), NIPS (2016-20), ICLR (2017-21), ICML (2017-20), AISTATS (2018-21), UAI (2018) and various NIPS and ICML workshops

Contact

thang.buivn at at at at gmail.com