News archive
November 2023:
I will give a talk in the workshop day (Day 3) of the Probabilistic Numerics Spring School in Southampton, UK, which will take place in 2024. Thanks to Jon Cockayne for the invitation.
Together with Nada Cvetkovic (TU Eindhoven), I am co-organising a minisymposium at SIAM UQ 2024 in Trieste, Italy on well-posedness, stability, and approximation of Bayesian inverse problems.
I have been invited to give a talk in an ERC-funded workshop on `Statistical aspects of non-linear inverse problems’, at Cambridge University in 2024. Thanks to Richard Nickl for the invitation.
October 2023:
I have been invited to give a talk in the seminar on computational engineering at Lappeenranta University of Technology. Thanks to Lassi Roininen for the invitation.
September 2023:
I organised the talks on Day 2 of the Potsdam DA Days. Thanks to Stéphanie van der Pas (Amsterdam UMC), Dan Crisan (Imperial College London), Claudia Strauch (Aarhus University) and Chris Oates (Newcastle University) for joining us in Potsdam and speaking about their work.
I have been invited to give a talk in a minisymposium on advances in derivative-based algorithms in finite and infinite dimensions at SIAM UQ 2024. Thanks to Michael C. Brennan, Matthew T. Li, and Thomas O'Leary-Roseberry for the invitation.
August 2023:
I will give a talk in the Potsdam Data Assimilation Days workshop that will take place from 11 to 13 September and is jointly organised by the SFB 1294 in Potsdam and the SFB 1481 at Aachen. Thanks to the organising committee for the invitation.
May 2023:
Some news about a positive development in scientific publishing highlights an important issue, namely that publishing companies make large profits off the work that researchers do for free.
April 2023:
The paper on dimension-independent Markov chain Monte Carlo on the sphere (joint with Daniel Rudolf, Björn Sprungk, and T. J. Sullivan) was accepted for publication at the Scandinavian Journal of Statistics. More details to follow.
March 2023:
I gave an invited talk in the Dortmund Data Science Colloquium at TU Dortmund, Germany. Thanks to Alexander Munteanu for the invitation.
I gave an invited talk in the minisymposium MS 265: "Model-Based Optimal Experimental Design" at SIAM CSE 2023 in Amsterdam. Thanks to Xun Huan for the invitation.
February 2023:
I gave a talk in the online reading group on generalised Bayesian inference, organised by Jeremias Knoblauch. Thanks to Jeremias for the invitation and to Chris Oates for suggesting me as a speaker.
January 2023:
A preprint on choosing observation operators to mitigate model error in Bayesian inverse problems was uploaded to the arXiv. Joint work with Nada Cvetkovic, Harshit Bansal, and Karen Veroy-Grepl.
December 2022:
I will give an invited talk in the minisymposium MS 265: "Model-Based Optimal Experimental Design" at SIAM CSE 2023 in Amsterdam. Thanks to the organisers - Xun Huan, Peng Chen, Omar Ghattas, and Youssef Marzouk - for the invitation.
Harshit Bansal, Nada Cvetkovic (Eindhoven University of Technology) and I are organising the minisymposium "MS403: "Model Error-Aware Methods for Inverse Problems and Experimental Design" at SIAM CSE 2023 in Amsterdam.
The arXiv preprint on dimension-independent sphere Markov chain Monte Carlo, which is joint work with Daniel Rudolf, Björn Sprungk, and Tim Sullivan, was updated.
November 2022:
I spoke about randomised time integration for differential equations in the Cote d'Azur workshop on Computational Uncertainty Quantification. Thanks to Yuto Miyatake and Motonobu Kanagawa for the invitation.
I spoke about stochastic optimal control in the research seminar on "Numerical Analysis of Stochastic and Deterministic Partial Differential Equations" at the Freie Universität Berlin. Thanks to Claudia Schillings for the invitation.
September 2022:
I gave an invited talk in the workshop "Forecast Verification and Data Assimilation in intermediate and large scale models of geophysical fluid dynamics, with applications to medium range and seasonal forecasting", which was part of the Isaac Newton Institute Satellite Program on "Geophysical fluid dynamics; from mathematical theory to operational prediction". Thanks to the organisers (Tobias Kuna, Jeroen Wouters, Jochen Broecker, and Giulia Carigi) for the invitation.
I gave a talk in the workshop Symposium on Inverse Problems: From experimental models to data and back, jointly organised by the SFB 1294, the SFB 1456, and the Gesellschaft für Inverse Probleme e.V. (the German Society for Inverse Problems). Thanks to the organisers, including Sebastian Reich, for the opportunity.
I gave an invited lecture on randomised Bayesian inference in the 6th International Summer School on Resource-Aware Machine Learning, organised by the SFB 876 at TU Dortmund. Thanks to the organisers of the school and to Alex Munteanu from the SFB 876 for the invitation and for hosting me.
August 2022:
Takeru Matsuda, Yuto Miyatake and I will organise a minisymposium at ICIAM 2023 in Tokyo, Japan on the subject of probabilistic and stochastic approaches to numerical solutions of differential equations.
July 2022:
I gave an invited talk in the minisymposium "Informing physics-based models through data" in the section on numerical analysis and applications at the Equadiff '15 conference in Brno, the Czech Republic. Thanks to Karen Veroy-Grepl for the invitation.
June 2022:
I gave an invited talk in the Research Seminar on Stochastics. Thanks to Carsten Hartmann, Ralf Wunderlich, and Omar Kebiri for the invitation.
I completed a week-long research visit at the Computational Science group of the Eindhoven University of Technology, generously hosted by Karen Veroy-Grepl.
May 2022:
I attended a workshop on PDE-constrained Bayesian inverse problems in the Erwin Schrödinger Institute Thematic Programme on Computational Uncertainty Quantification .
March 2022:
I will give an invited talk in the Resource-aware Machine Learning - 6th International Summer School 2022 (REAML 2022) at the Technische Universität Dortmund, Germany. Thanks to Katja Ickstadt and Alexander Munteanu for the invitation.
I will participate in the workshop on "PDE-constrained Bayesian inverse problems: Interplay of spatial statistical models with Machine Learning in PDE discretizations" within the Thematic Programme on Computational Uncertainty Quantification at the Erwin Schrödinger Institute in Vienna, Austria. Thanks to Sebastian Reich for the invitation.
I will give an invited talk in the minisymposium "Informing physics-based models through data" at Equadiff 15 in Brno, the Czech Republic. Thanks to Karen Veroy-Grepl for the invitation.
I gave an invited talk about randomised time integration for operator differential equations at the 2022 workshop on Probabilistic Approaches to Numerical Computation (Prob Num 2022). Thanks to Chris Oates and T. J. Sullivan for the invitation.
I gave a talk about stochastic optimal control and Frechet derivatives at the 13th International Workshop on Stochastic Models and Control.
Giuseppe Carere joined the group as a doctoral researcher in Project A07 of the Collaborative Research Centre SFB 1294. He brings research experience in model order reduction for PDE-constrained optimal control problems with random inputs.
February 2022:
The open-access paper on randomised time integration for operator differential equations (joint with Martin Stahn and T. J. Sullivan) was published at Calcolo.
I gave a talk about random Bayesian inverse problems at the colloquium of CASA at the Eindhoven University of Technology. Thanks to Nada Cvetkovic for the invitation.
December 2021:
Two open-access papers on gamma-convergence of Onsager-Machlup functionals were published in the journal Inverse Problems. These papers are joint work with Birzhan Ayanbayev, Ilja Klebanov, and T. J. Sullivan.
The first paper concerns applications to maximum a posteriori (MAP) estimation in Bayesian inverse problems.
The second paper concerns MAP estimation for infinite product measures on Banach spaces.
A preprint on dimension-independent Markov chain Monte Carlo methods on the sphere was uploaded to the arXiv. This is joint work with Daniel Rudolf, Björn Sprungk, and T. J. Sullivan.
September 2021:
I gave a talk in the workshop 'Bayesian Modeling and Computing' organised by Daniel Rudolf and Michael Habeck within the CRC 1456 "Mathematics of Experiment".
August 2021:
Two preprints on gamma-convergence of Onsager-Machlup functionals were uploaded to the arXiv. These papers are joint work with Birzhan Ayanbayev, Ilja Klebanov, and T. J. Sullivan.
The first preprint concerns applications to maximum a posteriori (MAP) estimation in Bayesian inverse problems.
The second preprint concerns MAP estimation for infinite product measures on Banach spaces.
June 2021:
A preprint on Frechet derivatives of expected values of path functionals with respect to perturbations of the drift was uploaded to the arXiv.
May 2021:
The German Research Foundation has approved funding for another four years of the Collaborative Research Centre SFB 1294 "Data Assimilation". See the announcement on the SFB's website here.
I gave an online talk in the Research Seminar Stochastics at the TU Bergakademie Freiberg, organised by Björn Sprungk.
I gave an online talk in the Data Science and Computational Statistics Seminar of the University of Birmingham, organised by Xiaocheng Shang and others.
March 2021:
A preprint on randomised time integration for operator differential equations was uploaded to the arXiv. Joint work with Martin Stahn and T. J. Sullivan.
I spoke about randomised time integration at the GAMM-Workshop 2021 "Sion Digital", organised by Robert Altmann.
February 2021:
A paper co-authored with Nada Cvetkovic and Tim Conrad is published online in the SIAM journal Multiscale Modelling and Simulation. This paper concerns a numerical method for transition path theory.
November 2020:
A paper co-authored with Nada Cvetkovic and Tim Conrad is accepted for publication in the SIAM journal Multiscale Modelling and Simulation. This paper concerns a numerical method for transition path theory.
August 2020:
I will give a pre-recorded talk on joint work with T. J. Sullivan and Aretha Teckentrup in the minisymposium on Bayesian Statistics at the Bernoulli-IMS One World Symposium.
June 2020:
A paper co-authored with T. J. Sullivan and Aretha Teckentrup is accepted for publication in the conference proceedings of EnuMATH 2019. This paper reviews some of the main results of our work on random Bayesian inverse problems.
February 2020:
Martin Stahn joined the UQ group as a researcher. Together with Peter Koltai (FU Berlin), Martin and I collaborated on proving Frechet differentiability for certain transfer operators in dynamical systems. Martin and I will analyse certain stochastic methods for modelling dynamical systems in fluid dynamics.
I'm pleased to announce that Nada Cvetkovic joined the UQ group as a researcher. Nada obtained her PhD at the Freie Universität Berlin in 2019. Together with Tim Conrad (FU Berlin), Nada and I collaborated on a numerical method for transition path theory. Nada and I will analyse properties of convex polytopes that are relevant for numerical methods.
The UQ group is established at the Institut für Mathematik of Universität Potsdam, with the generous support of the SFB 1294 "Data Assimilation".
January 2020:
I gave a talk in the statistics seminar of Markus Reiß at HU Berlin on random posteriors for Bayesian inverse problems (joint work with T. J. Sullivan and Aretha Teckentrup).
November 2019:
I gave a talk in the probability and statistics seminar at Uni Potsdam organised by Sara Mazzonetto and Pierre Houdebert, on the topic of Frechet differentiability of Koopman and Perron-Frobenius operators (joint work with Peter Koltai and Martin Stahn).
I gave a talk in Wilhelm Stannat's seminar for stochastic partial differential equations at TU Berlin, on the topic of Frechet differentiability of Koopman and Perron-Frobenius operators (joint work with Peter Koltai and Martin Stahn).
The preprint "Error bounds for some approximate posterior measures in Bayesian inference", which is joint work with T. J. Sullivan and Aretha Teckentrup, is uploaded to the arXiv. This preprint reviews the stability results that we proved in our paper on random posteriors for Bayesian inverse problems.
October 2019:
The joint paper with Andrew Stuart and T. J. Sullivan on strong convergence of probabilistic integrators is published in Statistics and Computing.
I received an invitation to attend the Hausdorff School on Markov Chain Monte Carlo in Bonn, organised by Nawaf Bou-Rabee and Andreas Eberle.
I gave a contributed talk on joint work on a method for time series analysis based on transition path theory (joint work with Nada Cvetkovic and Tim Conrad) at the workshop on "Nonlinear and stochastic methods in climate and geophysical fluid dynamics" at the Institut Henri Poincaré, within the trimester program "The Mathematics of Climate and Environment". Thanks to the organisers and the scientific committee for the opportunity.
I gave a talk at EnuMATH 2019 about Bayesian inference with random or deterministic approximate posteriors (joint work with Aretha Teckentrup and T. J. Sullivan). Thanks to Kathrin Smetana (U. Twente) and Olivier Zahm (INRIA Bordeaux) for the invitation to speak at their minisymposium on randomised methods.
September 2019:
The joint paper with Peter Koltai and Martin Stahn (formerly Martin Plonka) on Frechet differentiability of Koopman and Perron-Frobenius operators is published online in Nonlinearity.
I updated the Ethos section of the webpage to include links to some petitions that I have signed.
July 2019:
I gave a 30-minute talk on strong convergence of probabilistic integrators for ODEs (joint work with Andrew Stuart and Tim Sullivan) in a miniysmposium on probabilistic numerics at the SciCADE conference in Innsbruck. Thanks to Takeru Matsuda from the University of Tokyo and Yuto Miyatake from Osaka University for organising the minisymposium and for the invitation.
A preprint is uploaded to the arXiv: "Convergent discretisation schemes for transition path theory for diffusion processes". Joint work with Nada Cvetkovic and Tim Conrad of FU Berlin.
A short note is uploaded to the arXiv: "On a degeneracy ratio for bounded convex polytopes" (joint work with Nada Cvetkovic of FU Berlin).
May 2019:
I gave an invited 50-minute talk in a symposium on machine learning and scientific computing at the IWR in the University of Heidelberg, organised by Rob Scheichl. Topics of the talk: graphical models and molecular dynamics, stochastic optimal control and rare events, and randomised numerical integration.
A paper I wrote with Peter Koltai and Martin Plonka from FU Berlin is accepted for publication in the journal Nonlinearity. The subject: Frechet differentiability of Koopman and Perron-Frobenius operators for stochastic dynamical systems.
April 2019:
I started an open topic postdoctoral fellowship at the University of Potsdam, funded by the University of Potsdam and hosted by Sebastian Reich in the Institute of Mathematics.
March 2019
I gave a 30-minute talk at the Oberwolfach Workshop on Uncertainty Quantification on joint work with T. J. Sullivan on the equivalence of maximum a posteriori (MAP) estimators.
January 2019
A paper that I co-authored with Andrew Stuart from Caltech and T. J. Sullivan from FU Berlin is accepted for publication in a special issue of Statistics and Computing. The topic: convergence analysis of randomised numerical integration schemes for ordinary differential equations.
December 2018
I gave an invited 30-minute talk at a symposium on uncertainty quantification at the University of Potsdam, organised by Sebastian Reich. The topic was my paper with T. J. Sullivan on the equivalence of maximum a posteriori (MAP) estimators.