Research
Research interests
In general, my research interests include
Statistical Theory and Methods:
Developing efficient estimation methods (MCMC, SMC and Variational Bayes) for structured and/or high-dimensional data
Bayesian analysis and statistical machine learning, deep learning
Quantum computation for statistics
Applied:
Cognitive science/behavioral psychology
Business analytics, financial econometrics
Some current projects:
Statistical inference with intractable likelihood: The likelihood function is the cornerstone in statistical inference. However, in many modern statistical applications, the likelihood function is either analytically or computationally intractable. This makes it challenging to carry out likelihood-based inference using popular methodologies such as Variational Bayes, Importance Sampling, Markov chain Monte Carlo and Sequential Monte Carlo methods, which require exact evaluations of the likelihood function at each parameter value. This research project focuses on developing efficient methods for doing statistical inference with intractable likelihood.
Enabling Bayesian inference for big and high-dimensional data: Recent advances in technology have produced increasingly large volumes of data. Data are big in terms of both the number of observations (tall data) and the number of observed variables (high-dimensional data). This leads to many research opportunities as well as challenges in statistical inference, in particular simulation-based Bayesian inference. This research project attempts to enable Bayesian inference for Big Data and Big Models. In particular, it focuses on subsampling-based Markov chain Monte Carlo and Hamiltonian Monte Carlo for tall data, and Variational Bayes estimation methods for extremely high-dimensional data.
Deep learning based time series modeling and financial forecasting: This ARC-funded project pursues breakthroughs in modelling time effects which help reveal the hidden underlying structure in time series data, with a focus on flexible modelling of financial time series data. 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 their time series data
Cognitive science: flexible models and methods for cognitive model-based decision-making. This ARC-funded 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.
Research grants
ARC Training Centre in Data Analytics for Resources and Environment, 2021-, $3.8m, Chief Investigator.
Exploring quantum computation for data analysis. ACEMS research support scheme for 2021, $19,600. Usyd Business School Pilot Research Scheme for 2021-22, $45,000.
Bayesian inference for psychological theories with intractable likelihood. ARC Discovery Project, 2021-2023, $378,500, Chief Investigator.
Deep learning based time series modeling and financial forecasting. ARC Discovery Project, 2020-2022, $280,000, Chief Investigator.
Flexible models and methods for cognitive model-based decision-making; ARC Discovery Project, 2018-2020, $348,912, Chief Investigator.
Deep learning based time series modeling and financial forecasting, Usyd Business School Pilot Research Scheme, 2018-2019, $40,000, Chief Investigator.
Australian Research Council's Centre of Excellence for Mathematical and Statistical Frontiers (ACEMS), 2018-2021, Chief Investigator of a share $70,000.
Data Analysis for intractable models with applications to big data and psychology; Usyd Business School Pilot Research Scheme, 2017, $16,000, Chief Investigator.
Publications
A full list of my publications can be found on ResearchGate or Google Scholar. Most of the computer code used in my papers can be found at https://github.com/VBayesLab
Books
Data Analytics and Insights, textbook for Chartered Accountants Australia and New Zealand (published in Oct 2021 by Wiley)
Bayesian Computation with Variational Bayes, research monograph (in writing).
Statistical methodologies and theory
Godichon-Baggioni, Nguyen and Tran (2024) Natural Gradient Variational Bayes without Fisher Matrix Analytic Calculation and Its Inversion. Journal of the American Statistical Association.
Lopatnikova, Tran & Sisson (2024). An Introduction to Quantum Computing for Statisticians and Data Scientists. Foundations of Data Science.
Nguyen, N., Tran, M., Chandra, R. (2024). Sequential reversible jump MCMC for dynamic Bayesian neural networks. Neurocomputing, 564, 126960.
Lopatnikova & Tran (2023). Quantum variational Bayes on manifolds, IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2023).
Gunawan, D., Kohn, R., Tran, M. (2023). Flexible and Robust Particle Density Tempering for State Space Models [Preprint]. Econometrics and Statistics, in press.
Jie, R., Gao, J., Vasnev, A., Tran, M. (2022). Adaptive hierarchical hyper-gradient descent. International Journal of Machine Learning and Cybernetics, 13(12), 3785-3805.
Tran, Nguyen, Nguyen (2021). Variational Bayes on Manifolds. Statistics and Computing [Preprint] [Code] 31(6), 1-17.
Matias Quiroz, Minh-Ngoc Tran, Mattias Villani, Robert Kohn and Khue Dung Dang (2021). The block-Poisson estimator for optimally tuned exact subsampling MCMC. Journal of Computational and Graphical Statistics. 30(4), 877-888.
X. Yu, D. Nott, M.-N. Tran and N. Klein (2021). Assessment and adjustment of approximate inference algorithms using the law of total variance. Journal of Computational and Graphical Statistics. 30(4), 977-990.
D. Gunawan, K.-D. Dang, M. Quiroz, R. Kohn, M.-N. Tran (2020). Subsampling Sequential Monte Carlo for Static Bayesian Models. Statistics and Computing. 30(6), 1741-1758. [Preprint]
Salomone, Quiroz, Kohn , Villani and Tran (2020). Spectral Subsampling MCMC for Stationary Time Series. ICML 2020.
Jie, Gao, Vasnev and M.N. Tran (2020). HyperTube: A Framework for Online Hyperparameter Optimization with Resource Constraints. IEEE Access.
Jie, Gao, Vasnev and M.N. Tran (2020). Regularized flexible activation function combination for deep neural networks. ICPR 2020.
Dao Tung, M.N. Tran (2020). Flexible multivariate regression density estimation. Communications in Statistics: Theory and Methods.
Tran, Nguyen, Nott and Kohn (2020) Bayesian Deep Net GLM and GLMM. Journal of Computational and Graphical Statistics. 29(1), 97-113 .Code is available here.
Khue-Dung Dang, Matias Quiroz, Robert Kohn, Minh-Ngoc Tran, Mattias Villani (2019). Hamiltonian Monte Carlo with Energy Conserving Subsampling. Journal of Machine Learning Research, 20(100), pp. 1-31.
Matias Quiroz, Mattias Villani, Robert Kohn, Minh-Ngoc Tran (2019). Speeding Up MCMC by Efficient Data Subsampling. Journal of the American Statistical Association, 114 (526), 831-843. [Preprint]
David Gunawan, Minh-Ngoc Tran, Kosuke Suzuki, Josef Dick and Robert Kohn (2019). Computationally Efficient Bayesian Estimation of High Dimensional Archimedian Copulas with Discrete and Mixed Margins , Statistics and Computing, 29, 933–946. [Preprint]
Victor Ong, David Nott, Minh-Ngoc Tran, Scott Sisson and Christopher Drovandi (2018). Likelihood-free inference in high dimensions with synthetic likelihood. Computational Statistics and Data Analysis, Vol 128, pages 271-291.
Villani, M., Quiroz, M., Kohn, R., Tran, M.N. and Dang, K.D. (2018). Subsampling MCMC - A review for the survey statistician. Sankhya A (invited paper, accepted)
Victor Ong, David Nott, Minh-Ngoc Tran, Scott Sisson and Christopher Drovandi (2018). Variational Bayes with Synthetic Likelihood. Statistics and Computing, vol.28:4, pp. 971-988 [Paper]
Matias Quiroz, Minh-Ngoc Tran, Mattias Villani, Robert Kohn (2018). Speeding up MCMC by Delayed Acceptance and Data Subsampling. Journal of Computational and Graphical Statistics, vol.27:1, pp. 12-22
Christopher Drovandi and Minh-Ngoc Tran (2018). Improving the Efficiency of Fully Bayesian Optimal Design of Experiments using Randomised Quasi-Monte Carlo. Bayesian Analysis, vol.13:1, pp. 139-162 [paper]
D.T. Tung, M-N Tran, T.M. Cuong (2018). Bayesian Adaptive Lasso with Variational Bayes for Variable Selection in High-dimensional Generalized Linear Mixed Models. Communications in Statistics: Simulation and Computation (to appear) [Preprint] [Code]
Minh-Ngoc Tran, David Nott & Robert Kohn (2017). Variational Bayes with Intractable Likelihood. Journal of Computational and Graphical Statistics, vol.26:4, pp. 873-882 [Preprint]
Minh-Ngoc Tran, David Nott, Anthony Kuk & Robert Kohn (2016). Parallel variational Bayes for large datasets with an application to generalized linear mixed models. Journal of Computational and Graphical Statistics, 25(2), pp. 626-46 [Preprint]
Minh-Ngoc Tran, Micheal Pitt & Robert Kohn (2016). Adaptive Metropolis-Hastings sampling using reversible dependent mixture proposals, Statistics and Computing, 26 (1), pp. 361-81 [Preprint] [Journal version]
Minh-Ngoc Tran, Paolo Giordani, Xiuyan Mun, Robert Kohn and Micheal Pitt (2014). Copula-type estimators for flexible multivariate density modeling using mixtures. Journal of Computational and Graphical Statistics, 23(4), 1163-1178, [Preprint] [Journal version]
Chenlei Leng, Minh-Ngoc Tran & David J. Nott (2014). Bayesian adaptive Lasso. Annals of the Institute of Statistical Mathematics, 66:221-244. [Preprint] [Journal version] [Code]
Paolo Giordani, Xiuyan Mun, Minh-Ngoc Tran and Robert Kohn (2013). Flexible multivariate density estimation with marginal adaptation. Journal of Computational and Graphical Statistics, 22(4), 814-829. [Preprint].
Minh-Ngoc Tran, David J. Nott and Robert Kohn (2012). Simultaneous variable selection and component selection for regression density estimation with mixtures of heteroscedastic experts. Electronic Journal of Statistics, 6, 1170–1199. [Code] [Journal version]
David J. Nott, Minh-Ngoc Tran & Chenlei Leng (2012). Variational approximation for heteroscedastic linear models and matching pursuit algorithms. Statistics and Computing, 22(2), 497-512. [Preprint]
Minh-Ngoc Tran, David J. Nott & Chenlei Leng (2012). The predictive Lasso. Statistics and Computing, 22(5), 1069-1084. [Preprint]
David J. Nott, Lucy Marshall & Minh-Ngoc Tran (2012). The ensemble Kalman filter is an ABC algorithm. Statistics and Computing, special issue on ABC algorithms, 22(6), 1273-1276. [Preprint]
Weiyu Zhang, Xiaoxia Cao and Minh-Ngoc Tran (2012). The structural features and the deliberative quality of online discussions. Telematics and Informatics, 30(2), 74-86.
Minh-Ngoc Tran, Paolo Giordani and Robert Kohn (2012). Discussion on “Fast sparse regression and classification” by Jerome Friedman, International Journal of Forecasting, 28(3), 749-750.
Minh-Ngoc Tran (2011). A criterion for optimal predictive model selection, Communications in Statistics: Theory and Methods, 40(5), 893-906. [Preprint]
Minh-Ngoc Tran (2011). The loss rank criterion for variable selection in linear regression analysis, Scandinavian Journal of Statistics, 38(3), 466–479. [Preprint]
Marcus Hutter & Minh-Ngoc Tran (2010). Model selection with the loss rank principle, Computational Statistics and Data Analysis, 54(5), 1288-1306. [Preprint]
Minh-Ngoc Tran (2009). Penalized Maximum Likelihood Principle for Choosing Ridge Parameter, Communications in Statistics: Simulation and Computation, 38(8), 1610-1624. [Preprint]
D. G. Hung, V. Q. Hoang, N. V. Huu, T. M. Ngoc, L. H. Phuong (2006). A statistical method for development of credit scoring system, Vietnam Journal of Mathematical Applications, 4(2), 1-16 (in Vietnamese)
N. V. Huu, V. Q. Hoang, T. M. Ngoc (2006). Central limit theorems for functional of jump Markov processes, Vietnam Journal of Mathematics, 33(4), 443-461.
Cognitive science, experimental psychology, consumer behaviour
Dao, Gunawan, Tran, Kohn, Hawkins and Brown (2024) Bayesian Inference for Evidence Accumulation Models with Regressors Psychological Methods (to appear)
Dao, Gunawan, Tran, Kohn, Hawkins and Brown (2022) Efficient Selection Between Hierarchical Cognitive Models: Cross-validation With Variational Bayes. Psychological Methods.
D. Gunawan, G. Hawkins, M.-N. Tran, R. Kohn, S. Brown (2022). Time-Evolving Psychological Processes Over Repeated Decisions. Psychological Review, 129(3), 438-456 [Preprint]
Tran, Scharth, Gunawan, Kohn, Brown and Hawkins (2021). Robustly estimating the marginal likelihood for cognitive models via importance sampling. Behavior Research Methods, 53(3), 1148–1165, [Preprint] [Previous version]
Laura Wall, David Gunawan, Scott D Brown, Minh-Ngoc Tran, Robert Kohn, Guy E Hawkins (2021). Identifying relationships between cognitive processes across tasks, contexts, and time. Behavior Research Methods, 53(1), 78-95.
D. Gunawan, G. Hawkins, M.-N. Tran, R. Kohn, S. Brown (2019). New Estimation Approaches for the Linear Ballistic Accumulator Model. Journal of Mathematical Psychology, 90, 1023-68. [Preprint]
Financial econometrics, economic statistics
Liu, Tran, Wang, Gerlach and Kohn (2023). DeepVol: A Pre-Trained Universal Asset Volatility Model.
N. Nguyen, M.-N. Tran, D. Gunawan, R. Kohn (2023). A Statistical Recurrent Stochastic Volatility Model for Stock Markets. Journal of Business and Economics Statistics, 41(2), 414-428. Previous version: A long short-term memory stochastic volatility model. [Code]
A. Virbickaitė, H. Nguyen, MN Tran (2023). Bayesian predictive distributions of oil returns using mixed data sampling volatility models. Resources Policy, 86, 1041-67.
H Nguyen, N Nguyen, MN Tran (2023). A dynamic leverage stochastic volatility model. Applied Economic Letters, 30(1), 97-102.
N. Nguyen, M.-N. Tran, R. Kohn (2022). Recurrent Conditional Heteroskedasticity. Journal of Applied Econometrics [Preprint], [Code] 37(5), 1031-1054.
Z. Li, M.-N. Tran, C. Wang, R. Gerlach and J. Gao (2020). A Bayesian Long Short-Term Memory Model for Value at Risk and Expected Shortfall Joint Forecasting. [Preprint]
Unpublished papers
Matias Quiroz, Minh-Ngoc Tran, Mattias Villani, Robert Kohn (2017). Exact subsampling MCMC [Preprint]
David Gunawan, Minh-Ngoc Tran and Robert Kohn (2016). Fast Inference for Intractable Likelihood Problems using Variational Bayes [Preprint]
Minh-Ngoc Tran, Christopher Strickland, Micheal Pitt & Robert Kohn (2014). Annealed importance sampling for models with latent variables [Preprint]
Micheal Pitt, Minh-Ngoc Tran, Marcel Scharth & Robert Kohn (2013). On the existence of moments for high dimensional importance sampling [Preprint]
Minh-Ngoc Tran & Marcus Hutter (2011). Model selection by loss rank for classification and unsupervised learning [Preprint]
PhD thesis
Some perspectives on the problem of model selection, National University of Singapore, defended on November 9, 2011. [pdf] [slides]
My group
Anna Lopatnikova (Senior research fellow), Jessi Le (Research associate), Hung Dao (postdoc)
PhD students: Former: Khue-Dung Dang (now lecturer@Uni of Melbourne), Nghia Nguyen (now business analyst@Commonwealth Bank), Dao Thanh Tung (Head of math department @Vietnam), Hung Dao (postdoc@UNSW). Current: Paco Tseng (statistical methodology), Eldrin Hermoso (consumer behaviour), Chen Liu (financial econometrics), Megan Nguyen (statistical methodology), Rangika Peiris (financial econometrics), Dario Draca (statistical methodology).
Master/Interns/Honours Students
Anne Dao (Master student, now analyst @ Macquarie Group), Erin Zhang (honours, now data analyst@China), Maggie Zhong (Uni medal honours, now postgrad student@Columbia Uni, New York), Monica Castillo (Intern), Rohan Dickeson (honours, now business analyst@CBA), Joe Tan (honours, now postgrad@UCL, London), Silvia Du (2020 honours student), Tiffany Li (2020 honours student), Hugh Dawson (2021)
PhD thesis examiner for
Cathy Lee (UTS, now @ Google, Zurich), Nathaniel Tomasetti (Monash Uni)