Sergios Agapiou

DEPARTMENT OF MATHEMATICS AND STATISTICS, UNIVERSITY OF CYPRUS
1 University Avenue 2109 Nicosia, Cyprus · [email protected]

I am an Assistant Professor at the Department of Mathematics and Statistics, University of Cyprus.

My work is at the interface of Differential Equations and Probability/Statistics. More concretely, my research focuses on Bayesian nonparametric Statistics (including but not limited to inverse problems), and I am interested in both theoretical questions, such as the asymptotic performance of posterior distributions in the infinitely-informative data limit, as well as computational questions, such as the design and analysis of sampling algorithms in high dimensions.

I am also broadly interested in Data Science related questions, and I am one of the three directors of University of Cyprus' interdepartmental MSc in Data Science.

I am currently serving as chair of the Program Council and co-chair of the Prize Committee of The International Society for Bayesian Analysis.

You can find my CV here.

Publications

[13] Heavy-tailed Bayesian nonparametric adaptation

S. Agapiou and I. Castillo

Submitted

arXiv, code

[12] Adaptive inference over Besov spaces in the white noise model using p-exponential priors

S. Agapiou and A. Savva

To appear in Bernoulli

arXiv

[11] Laplace priors and spatial inhomogeneity in Bayesian inverse problems

S. Agapiou and S. Wang

Bernoulli, volume 30, number 2 (2024), pages 878-910

Online publication, arXiv

[10] Designing truncated priors for direct and inverse Bayesian problems

S. Agapiou and P. Mathé

Electronic Journal of Statistics, volume 16, number 1 (2022), pages 158-200

Online publication, arXiv

[9] Rates of contraction of posterior distributions based on p-exponential priors

S. Agapiou, M. Dashti and T. Helin

Bernoulli, volume 27, number 3 (2021), pages 1616-1642

Online publication (supplement), arXiv

[8] Modeling the first wave of Covid-19 pandemic in the Republic of Cyprus

S. Agapiou, A. Anastasiou, A. Baxevani, T. Christofides, E. Constantinou, G. Hadjigeorgiou, C. Nicolaides, G. Nikolopoulos and K. Fokianos

Scientific Reports, 11, Article number: 7342 (2021)

Online publication, arXiv

[7] Sparsity promoting and edge-preserving maximum a posteriori estimators in non-parametric Bayesian inverse problems

S. Agapiou, M. Burger, M. Dashti and T. Helin

Inverse Problems, volume 34, number 4 (2018)

Online publication, arXiv

[6] Posterior contraction in Bayesian inverse problems under Gaussian priors

S. Agapiou and P. Mathé

New trends in parameter identification for mathematical models, Springer series Trends in Mathematics (2018), pages 1-19

Online publication, arXiv

[5] Unbiased Monte Carlo: posterior estimation for intractable/infinite-dimensional models

S. Agapiou, G. O. Roberts and S. J. Vollmer

Bernoulli, volume 24, number 3 (2018), pages 1726-1786

Online publication (supplement), arXiv

[4] Importance Sampling: Intrinsic Dimension and Computational Cost

S. Agapiou, D. Sanz-Alonso, Omiros Papaspiliopoulos and A. M. Stuart

Statistical Science, volume 32, number 3 (2017), pages 405-431

Online publication (supplement), arXiv

[3] Analysis of the Gibbs sampler for hierarchical inverse problems

S. Agapiou, J. M. Bardsley, O. Papaspiliopoulos and A. M. Stuart

SIAM/ASA Journal on Uncertainty Quantification, volume 2, issue 1 (2014), pages 511-544

Online publication, arXiv

[2] Bayesian posterior contraction rates for linear severely ill-posed inverse problems

S. Agapiou, A. M. Stuart and Yuan-Xiang Zhang

Journal of Inverse and Ill-Posed Problems, volume 22, issue 3 (2014), pages 297-321

Online publication, arXiv

[1] Posterior contraction rates for the Bayesian approach to linear ill-posed inverse problems

S. Agapiou, S. Larsson and A. M. Stuart

Stochastic Processes and their Applications, volume 123, issue 10 (2013), pages 3828-3860

Online publication, arXiv

Theses

Aspects of Bayesian inverse problems

PhD thesis supervised by Andrew Stuart

pdf

A classical and a Bayesian approach to linear ill-posed inverse problems

MSc thesis supervised by Andrew Stuart

pdf

Stable and Infinitely Divisible distributions

Diploma thesis supervised by Spiros Argyros

pdf (in greek)

Teaching

MAS 451 - Final Exam 2023 final2023.zip

DSC 532 - Final Exam lab files and question 4 dataset

UCY's MSc in Data Science presentation

ΜΑΣ 458 - Στατιστική Ανάλυση Δεδομένων

Sergios Agapiou

Publications

[13] Heavy-tailed Bayesian nonparametric adaptation

[12] Adaptive inference over Besov spaces in the white noise model using p-exponential priors

[11] Laplace priors and spatial inhomogeneity in Bayesian inverse problems

[10] Designing truncated priors for direct and inverse Bayesian problems

[9] Rates of contraction of posterior distributions based on p-exponential priors

[8] Modeling the first wave of Covid-19 pandemic in the Republic of Cyprus

[7] Sparsity promoting and edge-preserving maximum a posteriori estimators in non-parametric Bayesian inverse problems

[6] Posterior contraction in Bayesian inverse problems under Gaussian priors

[5] Unbiased Monte Carlo: posterior estimation for intractable/infinite-dimensional models

[4] Importance Sampling: Intrinsic Dimension and Computational Cost

[3] Analysis of the Gibbs sampler for hierarchical inverse problems

[2] Bayesian posterior contraction rates for linear severely ill-posed inverse problems

[1] Posterior contraction rates for the Bayesian approach to linear ill-posed inverse problems

Theses

Aspects of Bayesian inverse problems

A classical and a Bayesian approach to linear ill-posed inverse problems

Stable and Infinitely Divisible distributions

Selected talks

- A new way of achieving adaptation (slides, video recording)

- Adaptive rates of contraction for spatially inhomogeneous unknowns

- Gauss vs Laplace rates of contraction under Besov regularity

- Statistical modelling of the effective reproduction number of COVID-19 in Cyprus (in greek)

- Edge-preserving Bayesian inversion

- The intrinsic dimension of Importance Sampling

- Bayesian posterior contraction rates via classical regularization techniques

- Practical unbiased Monte Carlo for intractable models

- Analysis of the Gibbs Sampler for hierarchical inverse problems

- Posterior contraction rates for Bayesian inverse problems

Teaching