Frank Werner
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Frank Werner has a total of 42 co-authors. Here are the top co-authors by number of shared publications:
- Axel Munk
Institute for Mathematical Stochastics University of Göttingen Goldschmidtstrasse 7 37077 Göttingen Germany
8 shared publications - Bernd Hofmann
Faculty of Mathematics , Chemnitz University of Technology , 09107 Chemnitz , Germany
7 shared publications - Thorsten Hohage
Institute of Numerical and Applied Mathematics, University of Göttingen, Göttingen, Germany
6 shared publications Axel Munk
Institute for Mathematical Stochastics, University of Göttingen , Göttingen
5 shared publications- Housen Li
Institute for Mathematical Stochastics University of Göttingen Göttingen Germany
3 shared publications - Shuai Lu
School of Mathematical Sciences, Fudan University, Shanghai 200433, China
3 shared publications
Latest Publications
- Open Access
Analyzing the Error of Generative Diffusion Models: From Euler-Maruyama to Higher-Order Schemes
Although generative diffusion models (GDMs) are widely used in practice, their theoretical foundations remain limited, especially concerning the impact of different discretization schemes applied to the underlying stochastic differential equation (SDE). Existing convergence analysis largely focuses on Euler-Maruyama (EM)-like methods and does not extend to higher-order schemes, which are naturally expected to provide improved discretization accuracy. In this paper, we establish asymptotic 2-Wasserstein convergence results for SDE-based discretization methods employed in sampling for GDMs. We provide an all-at-once error bound analysis of the EM method that accounts for all error sources and establish convergence under all prevalent score-matching error assumptions in the literature, assuming a strongly log-concave data distribution. Moreover, we present the first error bound result for arbitrary higher-order SDE-discretization methods with known strong L_2 convergence in dependence on the discretization grid and the score-matching error. Finally, we complement our theoretical findings with an extensive numerical study, providing comprehensive experimental evidence and showing that, contrary to popular believe, higher order discretization methods can in fact retain their theoretical advantage over EM for sampling GDMs.
Open AccessAnalyzing the Error of Generative Diffusion Models: From Euler-Maruyama to Higher-Order Schemes
Although generative diffusion models (GDMs) are widely used in practice, their theoretical foundations remain limited, especially concerning the impact of different discretization schemes applied to the underlying stochastic differential equation (SDE). Existing convergence analysis largely focuses on Euler-Maruyama (EM)-like methods and does not extend to higher-order schemes, which are naturally expected to provide improved discretization accuracy. In this paper, we establish asymptotic 2-Wasserstein convergence results for SDE-based discretization methods employed in sampling for GDMs. We provide an all-at-once error bound analysis of the EM method that accounts for all error sources and establish convergence under all prevalent score-matching error assumptions in the literature, assuming a strongly log-concave data distribution. Moreover, we present the first error bound result for arbitrary higher-order SDE-discretization methods with known strong L_2 convergence in dependence on the discretization grid and the score-matching error. Finally, we complement our theoretical findings with an extensive numerical study, providing comprehensive experimental evidence and showing that, contrary to popular believe, higher order discretization methods can in fact retain their theoretical advantage over EM for sampling GDMs.
- 12 September 2025
- Published by Springer Nature in The European Physical Journal Special Topics
- Vol. 235 (21) , 1-23
A continuous and robust version of dynamical component analysis
- 12 September 2025
- Published by Springer Nature in The European Physical Journal Special Topics
- Vol. 235 (21) , 1-23
A continuous and robust version of dynamical component analysis
- Open Access
- 3 February 2025
- Published by Taylor & Francis in Numerical Functional Analysis and Optimization
- Vol. 46 (4-5) , 322-347
A Unified Concept of the Degree of Ill-posedness for Compact and Non-Compact Linear Operator Equations in Hilbert Spaces Under the Auspices of the Spectral Theorem
- 3 February 2025
- Published by Taylor & Francis in Numerical Functional Analysis and Optimization
- Vol. 46 (4-5) , 322-347
Open AccessA Unified Concept of the Degree of Ill-posedness for Compact and Non-Compact Linear Operator Equations in Hilbert Spaces Under the Auspices of the Spectral Theorem
- 1 January 2025
- Published by American Institute of Mathematical Sciences (AIMS) in Inverse Problems and Imaging
- Vol. 19 (6) , 1268-1301
Maximum a posteriori testing in statistical inverse problems
This paper is concerned with a Bayesian approach to testing hypotheses in statistical inverse problems. Based on the posterior distribution $ \Pi \left(\cdot |Y = y\right) $, we want to infer whether a feature $ \langle\varphi, u^\dagger\rangle $ of the unknown quantity of interest $ u^\dagger $ is positive. This can be done by the so-called maximum a posteriori test. We provide a frequentistic analysis of this test's properties such as level and power, and prove that it is a regularized test in the sense of Kretschmann et al. (2024). Furthermore we provide lower bounds for its power under classical spectral source conditions in case of Gaussian priors. Numerical simulations illustrate its superior performance both in moderately and severely ill-posed situations.- 1 January 2025
- Published by American Institute of Mathematical Sciences (AIMS) in Inverse Problems and Imaging
- Vol. 19 (6) , 1268-1301
Maximum a posteriori testing in statistical inverse problems
This paper is concerned with a Bayesian approach to testing hypotheses in statistical inverse problems. Based on the posterior distribution $ \Pi \left(\cdot |Y = y\right) $, we want to infer whether a feature $ \langle\varphi, u^\dagger\rangle $ of the unknown quantity of interest $ u^\dagger $ is positive. This can be done by the so-called maximum a posteriori test. We provide a frequentistic analysis of this test's properties such as level and power, and prove that it is a regularized test in the sense of Kretschmann et al. (2024). Furthermore we provide lower bounds for its power under classical spectral source conditions in case of Gaussian priors. Numerical simulations illustrate its superior performance both in moderately and severely ill-posed situations.- Open Access
- 1 January 2025
- Published by Osterreichische Akademie der Wissenschaften, Verlag in ETNA - Electronic Transactions on Numerical Analysis
- Vol. 63, 33-62
Variational Poisson denoising via augmented Lagrangian methods
Electronic Transactions on Numerical Analysis (ETNA) is an electronic journal for the publication of significant new developments in numerical analysis and scientific computing. Papers of the highest quality that deal with the analysis of algorithms for the solution of continuous models and numerical linear algebra are appropriate for ETNA, as are papers of similar quality that discuss implementation and performance of such algorithms. New algorithms for current or new computer architectures are appropriate provided that they are numerically sound. However, the focus of the publication should be on the algorithm rather than on the architecture. The journal is published by the Kent State University Library in conjunction with the Institute of Computational Mathematics at Kent State University, and in cooperation with the Johann Radon Institute for Computational and Applied Mathematics of the Austrian Academy of Sciences (RICAM). Reviews of all ETNA papers appear in Mathematical Reviews and Zentralblatt für Mathematik. Reference information for ETNA papers also appears in the expanded Science Citation Index. ETNA is registered with the Library of Congress and has ISSN 1068-9613.
- 1 January 2025
- Published by Osterreichische Akademie der Wissenschaften, Verlag in ETNA - Electronic Transactions on Numerical Analysis
- Vol. 63, 33-62
Open AccessVariational Poisson denoising via augmented Lagrangian methods
Electronic Transactions on Numerical Analysis (ETNA) is an electronic journal for the publication of significant new developments in numerical analysis and scientific computing. Papers of the highest quality that deal with the analysis of algorithms for the solution of continuous models and numerical linear algebra are appropriate for ETNA, as are papers of similar quality that discuss implementation and performance of such algorithms. New algorithms for current or new computer architectures are appropriate provided that they are numerically sound. However, the focus of the publication should be on the algorithm rather than on the architecture. The journal is published by the Kent State University Library in conjunction with the Institute of Computational Mathematics at Kent State University, and in cooperation with the Johann Radon Institute for Computational and Applied Mathematics of the Austrian Academy of Sciences (RICAM). Reviews of all ETNA papers appear in Mathematical Reviews and Zentralblatt für Mathematik. Reference information for ETNA papers also appears in the expanded Science Citation Index. ETNA is registered with the Library of Congress and has ISSN 1068-9613.
- 21 October 2024
- Published by Oxford University Press (OUP) in Journal of the Royal Statistical Society Series B: Statistical Methodology
- Vol. 87 (2) , 510-528
Multiscale scanning with nuisance parameters
We develop a multiscale scanning method to find anomalies in a d-dimensional random field in the presence of nuisance parameters. This covers the common situation that either the baseline-level or additional parameters such as the variance are unknown and have to be estimated from the data. We argue that state of the art approaches to determine asymptotically correct critical values for multiscale scanning statistics will in general fail when such parameters are naively replaced by plug-in estimators. Instead, we suggest to estimate the nuisance parameters on the largest scale and to use (only) smaller scales for multiscale scanning. We prove a uniform invariance principle for the resulting adjusted multiscale statistic, which is widely applicable and provides a computationally feasible way to simulate asymptotically correct critical values. We illustrate the implications of our theoretical results in a simulation study and in a real data example from super-resolution STED microscopy. This allows us to identify interesting regions inside a specimen in a pre-scan with controlled family-wise error rate.
- 21 October 2024
- Published by Oxford University Press (OUP) in Journal of the Royal Statistical Society Series B: Statistical Methodology
- Vol. 87 (2) , 510-528
Multiscale scanning with nuisance parameters
We develop a multiscale scanning method to find anomalies in a d-dimensional random field in the presence of nuisance parameters. This covers the common situation that either the baseline-level or additional parameters such as the variance are unknown and have to be estimated from the data. We argue that state of the art approaches to determine asymptotically correct critical values for multiscale scanning statistics will in general fail when such parameters are naively replaced by plug-in estimators. Instead, we suggest to estimate the nuisance parameters on the largest scale and to use (only) smaller scales for multiscale scanning. We prove a uniform invariance principle for the resulting adjusted multiscale statistic, which is widely applicable and provides a computationally feasible way to simulate asymptotically correct critical values. We illustrate the implications of our theoretical results in a simulation study and in a real data example from super-resolution STED microscopy. This allows us to identify interesting regions inside a specimen in a pre-scan with controlled family-wise error rate.
- Open Access
- 27 December 2023
- Published by IOP Publishing in Inverse Problems
- Vol. 40 (2) , 025005
Adaptive minimax optimality in statistical inverse problems via SOLIT—Sharp Optimal Lepskiĭ-Inspired Tuning
We consider statistical linear inverse problems in separable Hilbert spaces and filter-based reconstruction methods of the form f ˆ α = q α T ∗ T T ∗ Y , where Y is the available data, T the forward operator, q α α ∈ A an ordered filter, and α > 0 a regularization parameter. Whenever such a method is used in practice, α has to be appropriately chosen. Typically, the aim is to find or at least approximate the best possible α in the sense that mean squared error (MSE) E [ ∥ f ˆ α − f † ∥ 2 ] w.r.t. the true solution f † is minimized. In this paper, we introduce the Sharp Optimal Lepskiĭ-Inspired Tuning (SOLIT) method, which yields an a posteriori parameter choice rule ensuring adaptive minimax rates of convergence. It depends only on Y and the noise level σ as well as the operator T and the filter q α α ∈ A and does not require any problem-dependent tuning of further parameters. We prove an oracle inequality for the corresponding MSE in a general setting and derive the rates of convergence in different scenarios. By a careful analysis we show that no other a posteriori parameter choice rule can yield a better performance in terms of the order of the convergence rate of the MSE. In particular, our results reveal that the typical understanding of Lepskiĭ-type methods in inverse problems leading to a loss of a log factor is wrong. In addition, the empirical performance of SOLIT is examined in simulations.
- 27 December 2023
- Published by IOP Publishing in Inverse Problems
- Vol. 40 (2) , 025005
Open AccessAdaptive minimax optimality in statistical inverse problems via SOLIT—Sharp Optimal Lepskiĭ-Inspired Tuning
We consider statistical linear inverse problems in separable Hilbert spaces and filter-based reconstruction methods of the form f ˆ α = q α T ∗ T T ∗ Y , where Y is the available data, T the forward operator, q α α ∈ A an ordered filter, and α > 0 a regularization parameter. Whenever such a method is used in practice, α has to be appropriately chosen. Typically, the aim is to find or at least approximate the best possible α in the sense that mean squared error (MSE) E [ ∥ f ˆ α − f † ∥ 2 ] w.r.t. the true solution f † is minimized. In this paper, we introduce the Sharp Optimal Lepskiĭ-Inspired Tuning (SOLIT) method, which yields an a posteriori parameter choice rule ensuring adaptive minimax rates of convergence. It depends only on Y and the noise level σ as well as the operator T and the filter q α α ∈ A and does not require any problem-dependent tuning of further parameters. We prove an oracle inequality for the corresponding MSE in a general setting and derive the rates of convergence in different scenarios. By a careful analysis we show that no other a posteriori parameter choice rule can yield a better performance in terms of the order of the convergence rate of the MSE. In particular, our results reveal that the typical understanding of Lepskiĭ-type methods in inverse problems leading to a loss of a log factor is wrong. In addition, the empirical performance of SOLIT is examined in simulations.
- 11 December 2023
- Published by IOP Publishing in Inverse Problems
- Vol. 40 (1) , 015013
Optimal regularized hypothesis testing in statistical inverse problems
Testing of hypotheses is a well studied topic in mathematical statistics. Recently, this issue has also been addressed in the context of inverse problems, where the quantity of interest is not directly accessible but only after the inversion of a (potentially) ill-posed operator. In this study, we propose a regularized approach to hypothesis testing in inverse problems in the sense that the underlying estimators (or test statistics) are allowed to be biased. Under mild source-condition type assumptions, we derive a family of tests with prescribed level α and subsequently analyze how to choose the test with maximal power out of this family. As one major result we prove that regularized testing is always at least as good as (classical) unregularized testing. Furthermore, using tools from convex optimization, we provide an adaptive test by maximizing the power functional, which then outperforms previous unregularized tests in numerical simulations by several orders of magnitude.
- 11 December 2023
- Published by IOP Publishing in Inverse Problems
- Vol. 40 (1) , 015013
Optimal regularized hypothesis testing in statistical inverse problems
Testing of hypotheses is a well studied topic in mathematical statistics. Recently, this issue has also been addressed in the context of inverse problems, where the quantity of interest is not directly accessible but only after the inversion of a (potentially) ill-posed operator. In this study, we propose a regularized approach to hypothesis testing in inverse problems in the sense that the underlying estimators (or test statistics) are allowed to be biased. Under mild source-condition type assumptions, we derive a family of tests with prescribed level α and subsequently analyze how to choose the test with maximal power out of this family. As one major result we prove that regularized testing is always at least as good as (classical) unregularized testing. Furthermore, using tools from convex optimization, we provide an adaptive test by maximizing the power functional, which then outperforms previous unregularized tests in numerical simulations by several orders of magnitude.
- 4 December 2023
- Published by Society for Industrial & Applied Mathematics (SIAM) in SIAM Journal on Scientific Computing
- Vol. 45 (6) , A3020-A3046
On a Dynamic Variant of the Iteratively Regularized Gauss–Newton Method with Sequential Data
For numerous parameter and state estimation problems, assimilating new data as they become available can help produce accurate and fast inference of unknown quantities. While most existing algorithms for solving those kinds of ill-posed inverse problems can only be used with a single instance of the observed data, in this work we propose a new framework that enables existing algorithms to invert multiple instances of data in a sequential fashion. Specifically we will work with the well-known iteratively regularized Gauss–Newton method (IRGNM), a variational methodology for solving nonlinear inverse problems. We develop a theory of convergence analysis for a proposed dynamic IRGNM algorithm in the presence of Gaussian white noise. We combine this algorithm with the classical IRGNM to deliver a practical (blended) algorithm that can invert data sequentially while producing fast estimates. Our work includes the proof of well-definedness of the proposed iterative scheme, as well as various error bounds that rely on standard assumptions for nonlinear inverse problems. We use several numerical experiments to verify our theoretical findings and to highlight the benefits of incorporating sequential data. The context of the numerical experiments comprises various parameter identification problems, including a toy elliptic PDE example and one of electrical impedance tomography.
- 4 December 2023
- Published by Society for Industrial & Applied Mathematics (SIAM) in SIAM Journal on Scientific Computing
- Vol. 45 (6) , A3020-A3046
On a Dynamic Variant of the Iteratively Regularized Gauss–Newton Method with Sequential Data
For numerous parameter and state estimation problems, assimilating new data as they become available can help produce accurate and fast inference of unknown quantities. While most existing algorithms for solving those kinds of ill-posed inverse problems can only be used with a single instance of the observed data, in this work we propose a new framework that enables existing algorithms to invert multiple instances of data in a sequential fashion. Specifically we will work with the well-known iteratively regularized Gauss–Newton method (IRGNM), a variational methodology for solving nonlinear inverse problems. We develop a theory of convergence analysis for a proposed dynamic IRGNM algorithm in the presence of Gaussian white noise. We combine this algorithm with the classical IRGNM to deliver a practical (blended) algorithm that can invert data sequentially while producing fast estimates. Our work includes the proof of well-definedness of the proposed iterative scheme, as well as various error bounds that rely on standard assumptions for nonlinear inverse problems. We use several numerical experiments to verify our theoretical findings and to highlight the benefits of incorporating sequential data. The context of the numerical experiments comprises various parameter identification problems, including a toy elliptic PDE example and one of electrical impedance tomography.
- 21 November 2023
- Published by Oxford University Press (OUP) in Microscopy
- Vol. 73 (3) , 287-300
Toward quantitative super-resolution microscopy: molecular maps with statistical guarantees
Quantifying the number of molecules from fluorescence microscopy measurements is an important topic in cell biology and medical research. In this work, we present a consecutive algorithm for super-resolution (stimulated emission depletion (STED)) scanning microscopy that provides molecule counts in automatically generated image segments and offers statistical guarantees in form of asymptotic confidence intervals. To this end, we first apply a multiscale scanning procedure on STED microscopy measurements of the sample to obtain a system of significant regions, each of which contains at least one molecule with prescribed uniform probability. This system of regions will typically be highly redundant and consists of rectangular building blocks. To choose an informative but non-redundant subset of more naturally shaped regions, we hybridize our system with the result of a generic segmentation algorithm. The diameter of the segments can be of the order of the resolution of the microscope. Using multiple photon coincidence measurements of the same sample in confocal mode, we are then able to estimate the brightness and number of molecules and give uniform confidence intervals on the molecule counts for each previously constructed segment. In other words, we establish a so-called molecular map with uniform error control. The performance of the algorithm is investigated on simulated and real data.
- 21 November 2023
- Published by Oxford University Press (OUP) in Microscopy
- Vol. 73 (3) , 287-300
Toward quantitative super-resolution microscopy: molecular maps with statistical guarantees
Quantifying the number of molecules from fluorescence microscopy measurements is an important topic in cell biology and medical research. In this work, we present a consecutive algorithm for super-resolution (stimulated emission depletion (STED)) scanning microscopy that provides molecule counts in automatically generated image segments and offers statistical guarantees in form of asymptotic confidence intervals. To this end, we first apply a multiscale scanning procedure on STED microscopy measurements of the sample to obtain a system of significant regions, each of which contains at least one molecule with prescribed uniform probability. This system of regions will typically be highly redundant and consists of rectangular building blocks. To choose an informative but non-redundant subset of more naturally shaped regions, we hybridize our system with the result of a generic segmentation algorithm. The diameter of the segments can be of the order of the resolution of the microscope. Using multiple photon coincidence measurements of the same sample in confocal mode, we are then able to estimate the brightness and number of molecules and give uniform confidence intervals on the molecule counts for each previously constructed segment. In other words, we establish a so-called molecular map with uniform error control. The performance of the algorithm is investigated on simulated and real data.