Broadly, my research interests centre around building and evaluating uncertainty-aware (also known as probabilistic) models for decision making that learn from data. Thus Statistics, Machine Learning, and Probabilistic modelling are the fields that I naturally find myself working in. These days, I mostly focus on uncertainty modelling in machine learning, especially Deep Learning. Making machine learning models output probability distributions that capture the uncertainty of predictions and developing coherent calibration evaluation methodology for probabilistic models are my main current research undertakings. However, the largest part of my research so far has dealt with learning and decision strategies under uncertainty in probabilistic temporal models. In more detail, I have engaged in the study of

  • sequential testing/estimation and quickest change-point detection problems in Bayesian Statistics,
  • trading strategies and valuation of financial derivatives in Financial Mathematics,
  • stochastic differential equations and stochastic filtering techniques.

Research papers

Article’s subject area is provided in square brackets. Authors are listed in alphabetical order (convention in mathematical disciplines) unless stated otherwise.

Evaluation of model calibration in classification [Machine Learning, Statistics]
First author, written in collaboration with D. Widmann, C. Andersson, F. Lindsten, J. Roll, T. Schön.
Submitted, 2018 (preprint to appear soon).

Optimal stopping of a Brownian bridge with an unknown pinning point [Probabilistic Modelling, Statistics]
Written in collaboration with E. Ekström.
Stochastic Processes and Their Applications, conditionally accepted, 2018 (preprint on arxiv, local pdf).

Monotonicity and robustness in Wiener disorder detection [Probabilistic Modelling, Statistics] Written in collaboration with E. Ekström.
To appear in Sequential Analysis, 2018 (preprint on arxiv, local pdf).

Asset liquidation under drift uncertainty and regime-switching volatility [Financial Mathematics, Probabilistic Modelling]
Single author.
Applied Mathematics & Optimization, 2018 (final version, preprint on arxiv, local pdf).

Optimal liquidation of an asset under drift uncertainty [Financial Mathematics, Probabilistic Modelling]
Written in collaboration with E. Ekström.
SIAM Journal on Financial Mathematics, vol. 7, no. 1, 2016, pp. 357-381 (final version, preprint on arxiv, local pdf).

Bayesian sequential testing of the drift of a Brownian motion [Statistics, Probabilistic Modelling] Written in collaboration with E. Ekström.
ESAIM: Probability and Statistics, vol. 19, 2015, pp. 626-648 (final version, preprint on arxiv, local pdf).

The 3/2-model as a stochastic volatility approximation for a large-basket price-weighted index [Financial Mathematics, Probabilistic Modelling] Written in collaboration with B. Hambly.
International Journal of Theoretical and Applied Finance,  vol. 18, 2015, ID 1550041 (final version, preprint as local pdf).

Theses

Optimal Sequential Decisions in Hidden-State Models
PhD thesis (Introduction with a summary of the included papers).
Defended on June 9, 2017, at Ångström Laboratory, Uppsala University..
Opponent: Prof. Huyên Pham, University Paris Diderot (Paris 7).
In the thesis, I found and investigated optimal statistical decision procedures for problems from Finance, Engineering, and Statistics.

Optimal Stopping under Drift Uncertainty
Licentiate thesis (Half-PhD thesis), Uppsala University, 2015 (final version, local pdf).
Reviewer: Dr Pavel Gapeev (London School of Economics).

Approximation of a large-basket index with an application to the pricing of variance options
Master’s thesis, Oxford University, 2012.
Supervisor: Prof. Ben Hambly.

 

Google Scholar page