Dr. Jack Jewson
I am currently a Juan de la Cierva Research fellow in the Department of Economics and Business at Universitat Pompeu Fabra (UPF) working with Dr. David Rossell and Dr. Piotr Zweirnik.
Sept 2019 - Mar 2022: Postdoctoral Research Assistant, Barcelona School of Economics (BSE), Universitat Pompeu Fabra (UPF) working with Dr. David Rossell and Dr. Piotr Zweirnik.
I obtained my PhD titled ''Bayesian Inference in the M-open world'' from the University of Warwick (in collaboration with the University of Oxford) as part of the Oxford-Warwick Statistics Programme (OxWaSP) under the supervision of Prof. Jim Q. Smith (Warwick) and Prof. Chris Holmes (Oxford) (Submitted 2019, Awarded 2020).
My research interests surround the methodological and philosophical challenges encountered when conducting Bayesian analyses for modern, high dimensional inference problems. In such scenarios a 'full' Bayesian analysis is inevitably impossible and one must resort to approximations, either at a modelling or a computational level.
I am particularly interested in general Bayesian updating and the use of loss functions to make principled Bayesian inference and improved decision making.
Model misspecification, robust statistics and the M-open world
Divergences, loss functions and scoring rules
Bayesian Sparse Vector Autoregressive Switching Models with Application to Human Gesture Phase Segmentation: Hadj-Amar, B., Jewson, J., Vannucci, M., arXiv preprint arXiv:2302.05347 (2023)
On the Stability of General Bayesian Inference: Jewson, J., Smith, J. Q., Holmes, C., arXiv preprint arXiv:2301.13701 (2023)
Graphical model inference with external network data: Jewson, J., Li, L., Battaglia, L., Hansen, S., Rossell, D., Zwiernik, P., arXiv preprint arXiv:2210.11107 (2022)
General Bayesian Loss Function Selection and the use of Improper Models: Jewson, J., Rossell, D., Journal of the Royal Statistical Society: Series B (Statistical Methodology), 84 ( 5), 1640– 1665. 2022.
Bayesian Approximations to Hidden Semi-Markov Models for Telemetric Monitoring of Physical Activity: Hadj-Amar, B., Jewson, J., Fiecas, M., Bayesian Analysis, Bayesian Anal. Advance Publication, 1-31, 2022.
An Optimization-centric View on Bayes' Rule: Reviewing and Generalizing Variational Inference: Knoblauch, J., Jewson, J., Damoulas, T.. Journal of Machine Learning Research, 23 (132), 1-109, 2022.
Mitigating statistical bias within differentially private synthetic data: Ghalebikesabi, S., Wilde, H., Jewson, J., Doucet, A., Vollmer, S., Holmes, C.,. In The 38th Conference on Uncertainty in Artificial Intelligence, 2022 (UAI 2022).
Foundation of Bayesian Learning from Synthetic Data: Wilde, H., Jewson, J., Vollmer, S., Holmes, C.,. In International Conference on Artificial Intelligence and Statistics (pp. 541-549). PMLR, 2021 (AISTATS, 2021).
Doubly Robust Bayesian Inference for Non-Stationary Streaming Data with β-Divergences: Knoblauch, J., Jewson, J., Damoulas, T., Advances in Neural Information Processing Systems (NeurIPS). 2018. (poster) (video)
Principles of Bayesian Inference Using General Divergence Criteria: Jewson, J.; Smith, J.Q.; Holmes, C. Entropy 2018, 20, 442
A comment on the Duckworth–Lewis–Stern method: Comments on "The Duckworth-Lewis-Stern method: extending the Duckworth-Lewis methodology to deal with modern scoring rates" by S.E. Stern, appearing as a viewpoint in JORS (2017). S.E. Stern's response can be found here
Learning Summary Statistic Hyperparameters: Discussion (p60) on the Read Paper ''Bayesian Restricted Likelihood Methods: Conditioning on Insufficient Statistics in Bayesian Regression'' by J. Lewis, S. MacEachern, and Y. Lee, appearing in Bayesian Analysis (2021).
Subjective Bayesian Updating: My discussion on the Read Paper "Beyond subjective and objective in statistics" by A. Gelman and C. Hennig, a condensed version appears in JRSSA (2017).
My PhD Thesis titled ''Bayesian Inference in the M-open world'' is available here
My MMorse Master's Thesis examining the Duckworth-Lewis method in the context of English county cricket is here