A5003419029- Stochastic processes and financial applications
- Statistical Methods and Inference
- Financial Risk and Volatility Modeling
- Sparse and Compressive Sensing Techniques
- Advanced Harmonic Analysis Research
- Statistical Methods and Bayesian Inference
- Markov Chains and Monte Carlo Methods
- Numerical methods in inverse problems
- advanced mathematical theories
- Gaussian Processes and Bayesian Inference
- Stochastic processes and statistical mechanics
- Complex Systems and Time Series Analysis
- Physics and Engineering Research Articles
- Capital Investment and Risk Analysis
- Traffic and Road Safety
- Advanced Statistical Methods and Models
- Computational Physics and Python Applications
- Point processes and geometric inequalities
- Nonlinear Differential Equations Analysis
- Direction-of-Arrival Estimation Techniques
- Advanced Neuroimaging Techniques and Applications
- Economic and Environmental Valuation
- Particle physics theoretical and experimental studies
- Hydrology and Drought Analysis
- Distributed Sensor Networks and Detection Algorithms
Karlsruhe Institute of Technology
2022-2025
Universität Hamburg
2016-2023
Université Paris Dauphine-PSL
2015-2016
Humboldt-Universität zu Berlin
2012-2015
Abstract Motivated by the high computational costs of classical simulations, machine-learned generative models can be extremely useful in particle physics and elsewhere. They become especially attractive when surrogate efficiently learn underlying distribution, such that a generated sample outperforms training limited size. This kind GANplification has been observed for simple Gaussian models. We show same effect simulation, specifically photon showers an electromagnetic calorimeter.
Abstract Besov spaces with dominating mixed smoothness, on the product of real line and torus as well bounded domains, are studied. A characterization these function in terms differences is provided. Applications to random fields, like Gaussian fields stochastic heat equation, discussed, based a Kolmogorov criterion for regularity smoothness.
Quantile estimation in deconvolution problems is studied comprehensively. In particular, the more realistic setup of unknown error distributions covered. Our plug-in method based on a density estimator and minimax optimal under minimal natural conditions. This closes an important gap literature. Optimal adaptive obtained by data-driven bandwidth choice. As side result, we obtain rates for distribution functions with distributions. The applied to real data example.
Parameter estimation for a parabolic linear stochastic partial differential equation in one space dimension is studied observing the solution field on discrete grid fixed bounded domain. Considering an infill asymptotic regime both coordinates, we prove central limit theorems realized quadratic variations based temporal and spatial increments as well double time space. Resulting method of moments estimators diffusivity volatility parameter inherit normality can be constructed robustly with...
We study the nonparametric calibration of exponential Lévy models with infinite jump activity. In particular our analysis applies to self-decomposable processes whose density can be characterized by $k$-function, which is typically nonsmooth at zero. On one hand estimation drift, activity measure $α:=k(0+)+k(0-)$ and analogous parameters for derivatives $k$-function are considered on other we estimate nonparametrically $k$-function. Minimax convergence rates derived. Since depend $α$,...
Abstract In recent years, neural network-based classification has been used to improve data analysis at collider experiments. While this strategy proves be hugely successful, the underlying models are not commonly shared with public and rely on experiment-internal as well full detector simulations. We show a concrete implementation of newly proposed strategy, so-called Classifier Surrogates, trained inside experiments, that only utilise publicly accessible features truth information. These...
As a starting point we prove functional central limit theorem for estimators of the invariant measure geometrically ergodic Harris-recurrent Markov chain in multi-scale space. This allows to construct confidence bands density with optimal (up undersmoothing) $L^{\infty}$-diameter by using wavelet projection estimators. In addition our setting applies drift estimation diffusions observed discretely fixed observation distance. We function and finally adaptive completely data-driven estimator.
We estimate linear functionals in the classical deconvolution problem by kernel estimators. obtain a uniform central limit theorem with $\sqrt{n}$–rate on assumption that smoothness of is larger than ill–posedness problem, which given polynomial decay rate characteristic function error. The distribution generalized Brownian bridge covariance structure depends error and functionals. proposed estimators are optimal sense semiparametric efficiency. class wide enough to incorporate estimation...
Observing prices of European put and call options, we calibrate exponential Lévy models nonparametrically. We discuss the efficient implementation spectral estimation procedures for finite jump activity as well self-decomposable models. Based on sample variances, confidence intervals are constructed volatility, drift and, pointwise, density. As demonstrated by simulations, these perform in terms size coverage probabilities. compare performance infinite based options German DAX index find...
We study the estimation of covariance matrix $\Sigma$ a $p$-dimensional normal random vector based on $n$ independent observations corrupted by additive noise. Only general nonparametric assumption is imposed distribution noise without any sparsity constraint its matrix. In this high-dimensional semiparametric deconvolution problem, we propose spectral thresholding estimators that are adaptive to $\Sigma$. establish an oracle inequality for these under model miss-specification and derive...
Si une fonctionnelle dans un problème inverse non-paramétrique peut être estimée à vitesse paramétrique, alors la minimax ne donne aucune information sur le caractère mal posé du problème. Pour avoir borne inférieure plus précise, nous étudions l’efficacité semi-paramétrique sens de Hájek–Le Cam pour l’estimation des modèles indirects réguliers. Ces derniers sont caractérisés comme que l’on approcher localement par modèle linéaire bruit blanc décrit l’opérateur score généralisé. Un théorème...