Find Academic Supervisors

Ingo SteinwartView profile

Professor

• Statistical Learning Theory
• Kernel-based learning algorithms
• Cluster Analysis
• +6 more

Ingo Steinwart is a Full Professor and Head of the Institute for Stochastics and Applications at the University of Stuttgart, within the Faculty of Mathematics and Physics. He holds a Chair for Stochastics and has held academic positions since 2010, including roles at Los Alamos National Laboratory and the University of California, Santa Cruz. His research focuses on statistical learning theory, kernel-based methods, cluster analysis, and neural networks, with significant contributions to the theoretical foundations of machine learning. Education: He earned a Doctorate (Dr. rer. nat.) in Mathematics from Friedrich-Schiller University, Jena (2000) and a Diploma in Mathematics from Carl-von-Ossietzky University, Oldenburg (1997). Research Interests: Steinwart's work emphasizes rigorous mathematical analysis of learning algorithms, including kernel methods, support vector machines (SVMs), and density estimation. His research explores topics like the capacity of function classes, generalization bounds, and applications to large-scale data. Notable contributions include advancements in SVM theory, reproducing kernel Hilbert spaces, and cluster analysis techniques. Publications: His work spans foundational papers on SVMs, kernel methods, and learning theory, with a focus on theoretical guarantees and algorithmic efficiency. Recent contributions address topics like adaptive learning rates, neural network initialization, and the theoretical limits of kernel-based methods. Grants & Editorial Roles: Steinwart serves as Associate Editor for journals such as the Journal of Complexity and the Annals of Statistics. He has organized conferences like COLT and contributed to software tools like liquidSVM for SVM implementation. Labs & Software: He leads development of software packages like liquidSVM (for SVMs) and liquidCluster (for cluster analysis), emphasizing computational efficiency and automated hyperparameter selection.

Thorsten HohageView profile

Professor

• Inverse Problems
• Partial Differential Equations
• Regularization Theory
• +4 more

Prof. Thorsten Hohage is a Full Professor at the Institute for Numerical and Applied Mathematics, Georg-August University of Göttingen since 2009. He leads the Inverse Problems Group, focusing on computational and theoretical aspects of inverse problems in partial differential equations, regularization theory, and numerical algorithms. His research spans applications in optics, medical imaging (MRI), helioseismology, and X-ray phase contrast imaging. **Career**: Full Professor (W3) since 2009 Associate Professor (W2) 2007–2009 Junior Professor 2002–2007 Postdoc at Zuse Institute Berlin (2000–2002) Research Assistant at Johannes-Kepler University Linz (1996–2000) **Research Interests**: Inverse scattering and PDE-based inverse problems Regularization theory for statistical and deterministic inverse problems Efficient numerical methods (e.g., learned infinite elements) Applications in helioseismology, phase retrieval, and medical imaging **Notable Contributions**: Developed methodologies for transparent boundary conditions, variational source conditions, and regularization techniques in Besov spaces. His work bridges theory and practice in computational inverse problems. **Affiliations**: Member of the Max Planck Institute for Solar System Research (since 2017) and part of the GGNB Physics of Biological and Complex Systems program.

Navid AnsariView profile

Researcher

• Deep Learning
• Generative Models
• Optimization
• +4 more

Navid Ansari is a doctoral researcher at the Max Planck Institute for Informatics (MPI-INF) and Saarland University, affiliated with the Artificial Intelligence Aided Design and Manufacturing group under the Computer Graphics department. His work bridges academia and industry, with internships at Amazon AWS and the Max Planck Institute for Brain Research. PhD in Computer Science (2021–Present), Saarland University & MPI-INF MSc in Visual Computing (2018–2021), Saarland University BSc in Electrical Engineering (2013–2017), Shiraz University Navid's research focuses on deep learning, generative models, and optimization techniques. His work includes mixed-integer optimization for neural networks, Bayesian design optimization, uncertainty quantification in AI systems, and applications in computational manufacturing and molecular design. He has contributed to top-tier venues like AAAI, NeurIPS (Spotlight), and SIGGRAPH. His publications address trends in AI-aided design optimization, uncertainty-aware modeling, sparsity in large language models, and inverse molecular design aligned with molecular dynamics. This work spans applications in manufacturing, computational chemistry, and natural language processing. Navid has collaborated with institutions such as Amazon AWS (LLM sparsification) and the Max Planck Institute for Brain Research (behavioral neuroscience). He is part of the Saarland Informatics Campus, a hub for visual computing and AI research.

• Machine Learning
• Physics-Informed Neural Networks
• Computational Fluid Dynamics
• +12 more

Dr. Michael Rom is a researcher at the Institute for Geometry and Practical Mathematics (IGPM), part of the Faculty of Mathematics, Computer Science, and Natural Sciences at RWTH Aachen University, Germany. He is actively engaged in applied mathematics and computational engineering, with a strong focus on fluid dynamics, tribology, and machine learning applications in industrial processes. Research Interests: Physics-Informed Neural Networks Transpiration Cooling and Porous Media Flow Hydrodynamic Lubrication and Cavitation Modeling Reduced Basis and Homogenization Methods Grid Generation and B-Spline Surfaces Grinding Wheel and Surface Modeling His recent publications (2021–2025) show a clear trend toward integrating deep learning and data-driven methods into classical engineering problems, particularly in manufacturing, aerospace thermal protection, and tribological systems. He frequently collaborates with experts in mechanical engineering and materials science, bridging disciplines to solve complex multiphysics problems. Scientific Awards: No awards mentioned in the provided text. Dr. Rom has supervised or co-supervised several students, as evidenced by co-authored works with individuals like M. Schultz, F. Brumand-Poor, and V. König. He has also contributed to teaching at RWTH Aachen, offering courses in numerical computation and finite element methods. His research is supported by collaborations with industrial and academic partners, particularly in digital production and thermal spray technologies. Laboratories and Research Teams: Institute for Geometry and Practical Mathematics (IGPM), RWTH Aachen Collaboration with the Chair of Production Metrology and Quality Management (F. Klocke) Part of interdisciplinary projects such as Internet of Production and Monetization of Technical Data

Nils ThuereyView profile

Professor

• Computer Graphics
• Fluid Dynamics
• Machine Learning
• +3 more

Prof. Nils Thuerey is a Professor of Physically Based Simulation at the Technical University of Munich (TUM), within the TUM School of Computation, Information and Technology. He specializes in computer graphics, fluid dynamics, and machine learning applications in physics-based simulations. His work focuses on developing algorithms for fluid simulations, turbulence modeling, and differentiable physics solvers. He earned his doctorate in fluid simulations from the University of Erlangen-Nuremberg (2006) and held positions at ETH Zurich, ScanlineVFX, before joining TUM in 2013. Research Interests: Fluid Dynamics and Turbulence Modeling Machine Learning for Physics Simulations Deep Learning in Computer Graphics Differentiable Simulations and Inverse Problems Key Awards: Technical Oscar (2012) for contributions to fluid effects in films/games ERC Starting Grant (2014) for 'realFlow' research University of Erlangen Staedtler Graduation Award (2008) His recent work emphasizes integrating deep learning with physical models, such as differentiable solvers for fluid dynamics and probabilistic methods for weather forecasting. He leads projects like WeatherBench and develops tools like φ-ML and phiflow for scientific computing.

Elisabeth UllmannView profile

Associate Professor

• Uncertainty Quantification
• Scientific Computing
• Numerical Analysis
• +3 more

Prof. Elisabeth Ullmann is an Associate Professor for Scientific Computing and Uncertainty Quantification at the Technical University of Munich (TUM), affiliated with the TUM School of Computation, Information, and Technology and the Department of Mathematics. She holds a PhD from TU Bergakademie Freiberg (2008) and has held academic roles including postdoctoral positions at the University of Bath, University of Hamburg, and University of Maryland. Her research focuses on developing efficient algorithms for uncertainty quantification in partial differential equations with random coefficients, Bayesian inverse problems, and rare event simulation. She is an Associate Editor for the SIAM Journal on Scientific Computing and SIAM/ASA Journal on Uncertainty Quantification , and teaches courses on numerical methods for uncertainty quantification and partial differential equations. Her work emphasizes multilevel Monte Carlo methods, stochastic Galerkin discretizations, and probabilistic numerical methods. Notable contributions include error analysis for rare event probabilities and multilevel estimators for high-dimensional problems. She collaborates internationally and maintains an active research group in Scientific Computing & Uncertainty Quantification at TUM.

Michael EngelView profile

Researcher

• Bayesian Inverse Problems
• Machine Learning
• Optimization Methods
• +3 more

Michael Engel is a Researcher at the Chair of Remote Sensing Technology, Technical University of Munich since March 2021. He holds a Master's in Materials Science and Engineering (2020) and a Bachelor's in Engineering Science (2018), both from TU Munich, with a focus on Uncertainty Quantification and Bayesian Inverse Problems. His research interests revolve around optimizing objective functions in inverse problems, leveraging methods like Curriculum Learning and Multi-Task Systems via Multiresolution Analysis. He combines classical numerical techniques with machine learning to address challenges in Bayesian problems and regression. Education: Bachelor's in Engineering Science, TU Munich (2018) Master's in Materials Science and Engineering, TU Munich (2020) Research interests include: - Bayesian Inverse Problems - Machine Learning Integration in Numerical Methods - Optimization of Multi-Task Systems - Uncertainty Quantification in Environmental Models Key projects include the Global Earth Monitor (GEM) , aiming to enhance Copernicus data utilization for large-scale environmental monitoring. His publications span remote sensing optimization, wavelet-based sensitivity analysis, and deforestation assessment using multi-sensor data. He contributes to the TU Munich's Engineering Risk Analysis Group and collaborates with the Chair of Numerical Mathematics and Hydrology. His work bridges physical modeling principles with computational efficiency in earth observation systems.

Martin KolbView profile

Professor

• Stochastic Processes
• Probability Theory
• Mathematical Physics
• +3 more

Roles & Affiliations: Martin Kolb is a Professor (W3) at Paderborn University's Institute of Mathematics (Applied Mathematics and Stochastics group leader). He previously held roles including Professor (W2) at Paderborn (2014–2017), Lecturer at University of Reading (2012–2014), and Harrison Early Career Professor at Warwick University (2011–2012). Education: PhD in Mathematics from TU Kaiserslautern (2009), Diploma in Mathematics from Ludwig-Maximilians-University Munich (2005). Studied at University of Copenhagen (2002–2003). Research Interests: Focuses on stochastic processes, probability theory, and mathematical physics. Key topics include quasi-stationary distributions, Fleming-Viot models, spectral asymptotics, and applications in quantum computing (e.g., Photonic Quantum Computing project). His work bridges theoretical mathematics with computational methods and physical systems. Scientific Engagement: Organized major events like the Weierstrass-Lecture series (since 2019) featuring Fields medalists and co-organized workshops such as the Oberwolfach Workshop on Stochastic Processes (2020). Active in conference committees for topics like 'Stochastic Processes under Constraints' (2016). Labs/Teams: Leads research projects including 'PhoQC: Photonic Quantum Computing' and 'Super-Brownian Motion with single point source'.

Dirk BlömkerView profile

Professor

• Stochastic partial differential equations
• Surface growth
• Cahn-Hilliard equation
• +11 more

Dirk Blömker is a Professor in the Institute of Mathematics at the University of Augsburg (since 2006). His research focuses on Stochastic Partial Differential Equations (SPDEs) with applications to surface growth models , phase separation (Cahn-Hilliard, Allen-Cahn), stochastic bifurcation , and multiscale analysis (amplitude/modulation equations, fast diffusion). He also works on filter theory (Ensemble Kalman) and numerics of SPDEs . Habilitation : RWTH Aachen (2006) Doctorate : University of Augsburg (2000) Diploma : University of Münster (1997) Intermediate Diploma : University of Münster (1993) Abitur : Gymnasium Lengerich (1989) His work spans rigorous mathematical analysis of SPDEs (existence/uniqueness, blow-up, regularity) and numerical methods (convergence analysis, a-posteriori bounds). Recent articles examine fractional noise, Lévy processes, and ensemble Kalman inversion. Collaborators include Luigi Bianchi, Claudia Schillings, and Philipp Wacker. He has held academic positions at RWTH Aachen (Assistant/Acting Lecturer), University of Bonn (Acting Lecturer), and University of Warwick (DFG Research Fellowship).

• Statistical Modeling
• Adaptive Smoothing
• Dimension Reduction
• +4 more

Dr. Joerg Polzehl is a senior researcher at the Weierstrass Institute for Applied Analysis and Stochastics (WIAS) in Berlin, Germany, where he leads research in the Stochastic Algorithms and Nonparametric Statistics group. His work focuses on developing statistical methodologies for analyzing complex imaging data, particularly in the field of magnetic resonance imaging. Polzehl maintains active collaborations with neuroscientists and medical researchers to advance quantitative analysis techniques for brain imaging. Polzehl's research interests center on statistical modeling, adaptive smoothing techniques, and dimension reduction methods with primary applications in image processing and neuroscience. He has pioneered structural adaptive smoothing approaches that preserve edges and boundaries in medical images while reducing noise. His work bridges theoretical statistics with practical applications in neuroimaging, where he develops methods that account for the specific challenges of MRI data including low signal-to-noise ratios and complex spatial dependencies. Analysis of Polzehl's recent publications reveals a strong focus on advancing MRI data analysis techniques across multiple modalities including diffusion MRI, functional MRI, and quantitative MRI. His work consistently emphasizes the development of statistical methods that are both theoretically sound and practically implementable, with a particular emphasis on reproducibility and open science. The integration of nonparametric statistical approaches with domain-specific knowledge of MRI physics represents a distinctive feature of his research program. Polzehl has made significant contributions through the development of several R software packages that implement his methodological innovations, including aws (Adaptive Weights Smoothing), adimpro (Adaptive Smoothing of Digital Images), dti (dMRI Analysis), qMRI (Analysis of quantitative MRI Experiments), and fmri (Analysis of fMRI Experiments). These packages have become important tools in the neuroimaging community, enabling researchers to apply advanced statistical methods to their imaging data. As co-author of the book "Magnetic Resonance Brain Imaging: Modeling and Data Analysis using R" (2019) with Karsten Tabelow, Polzehl has created a comprehensive resource that bridges the gap between statisticians and neuroimaging researchers. The book provides practical guidance on analyzing various types of MRI data using R, with fully reproducible examples that demonstrate best practices in neuroimaging data analysis.

• Algebraic Geometry
• Combinatorial Algebraic Geometry
• Syzygies
• +11 more

Daniele Agostini is a Junior Professor (Tenure Track) in the Department of Mathematics at Eberhard Karls University of Tübingen, affiliated with the Faculty of Mathematics and Natural Sciences. He is actively involved in the research section on Combinatorial Algebraic Geometry and serves as the international coordinator for the Mathematics Department. University: Eberhard Karls University of Tübingen School: Faculty of Mathematics and Natural Sciences Department: Department of Mathematics Academic Rank: Assistant Professor (Junior Professor, Tenure Track) Email: daniele.agostini@uni-tuebingen.de His primary research lies in algebraic geometry, with a focus on syzygies, curves, surfaces, abelian varieties, and their moduli spaces. He also explores interactions with probability, statistics, and mathematical physics. His recent publications reflect deep work in theta functions, KP solitons, and likelihood geometry, often in collaboration with institutions like MPI-MiS Leipzig and researchers such as Bernd Sturmfels. The 15 most recent publications demonstrate a consistent trend in combinatorial algebraic geometry, particularly in the study of moduli spaces, integrable systems, and computational aspects of algebraic varieties. Key themes include the geometry of curves and their Jacobians, tropical methods, and applications to mathematical physics and statistics. His work bridges theoretical depth with computational experimentation. Otto Hahn Medal (awarded to co-supervised PhD student Claudia Fevola, 2024) Agostini advises several PhD students, including Kees Heesterbeek, Clemens Nollau, Javier Sendra, and Claudia Fevola. He has been involved in significant research projects such as the ANR-DFG POK0 project on positivity of K-trivial varieties and the SFB-TRR 195 'Symbolic Tools'. He has taught courses such as Algebraic Curves, Algebraic Geometry, and Number Theory at Tübingen, Leipzig, and Berlin. He leads a research group in combinatorial algebraic geometry, with current members including postdoc Felix Röhrle and PhD students. His research group includes current and former members such as Demir Eken, Claudia Fevola, Leonid Monin, and Angel Rios Ortiz. He is also part of the URTAG seminar and contributes to the CRC TRR 195 collaboration, emphasizing interdisciplinary and international research engagement.

• Biomedical Engineering
• Biosignals
• Signal Processing
• +7 more

Thomas Schanze serves as Professor at Central Hesse University of Applied Sciences within the Faculty of Life Science Engineering, holding multiple leadership roles including Dean of the Faculty, Head of the Institute of Biomedical Engineering, and Senate Member. His office is located at Ringallee 37, D14.2.05. His research spans Biomedical Engineering , with specialization in Biosignals, Neuroprosthetics, and Computational Physics . Key focus areas include neural signal processing, machine learning applications in medical diagnostics, and control systems for biomedical devices. His work bridges theoretical physics with clinical applications, particularly in ECG analysis and neuroprosthetic development. Recent publications demonstrate consistent contributions to biomedical signal processing, with emphasis on noise reduction techniques (2017), ECG classification methods (2017), and advanced particle tracking algorithms (2016-2017). His research shows strong interdisciplinary connections between physics, engineering, and clinical medicine. Schanze actively promotes young talent through student competitions, mentoring award-winning participants in DGBMT contests (2014-2020). He serves on doctoral committees including the Doctoral Center for Engineering Sciences at Research Campus Central Hesse. Administratively, he organizes major conferences including AUTOMED (2013-2018) and BMT Annual Conferences (2012-2018), and previously led the Master's program in Biomedical Engineering until March 2024.

• Statistics
• Optimization
• Inverse Problems
• +3 more

Mina Saee Bostanabad is a Research Fellow at the University of Göttingen , affiliated with the Institute of Numerical and Applied Mathematics and the RTG 2088 Structure Recognition in Complex Data research training group. Her work intersects Statistics, Optimization, Inverse Problems, Machine Learning , and Data Science , focusing on extracting mathematical structures from complex datasets. Email: m.saee@math.uni-goettingen.de Phone: +49 551 39 24515 Affiliation: University of Göttingen, Germany She serves as the responsible lecturer for the Machine Learning in Data Science course (ID 503236), contributing to postgraduate education in data science methodologies.

Roland PotthastView profile

Professor

• Numerical Weather Prediction (NWP)
• Data Assimilation
• Inverse Problems
• +9 more

Prof. Dr. Roland Potthast is a Full Professor for Applied Mathematics at the University of Reading, UK, and Director of the Meteorological Analysis and Modeling Department (FE1) at the German Weather Service (DWD) since 2020. He leads ~110-120 researchers focused on operational numerical weather prediction (NWP), Earth system modeling, and data assimilation for the ICON model. Key Institutions: University of Reading, German Weather Service (DWD), Max-Planck-Institute for Meteorology (MPI-M), Karlsruhe Institute of Technology (KIT), ETH Zurich. Research: Data assimilation, inverse problems, ensemble/particle filtering, high-resolution NWP systems (ICON-global, ICON-EU, ICON-D2), and integration of machine learning with weather modeling. Scientific Contributions: Developed hybrid ensemble-variational data assimilation (EnVar) schemes, 4D-LETKF for convective-scale systems, and ultra-rapid data assimilation methods. His work impacts air traffic control, severe weather warnings, and energy supply applications. Awards/Grants: ESM Project (2022-2030), BeCoM (2022-2025), CONTRAILS (2022-2025). DFG grants, EPSRC funding, EU projects, and international fellowships. Leadership: Steering Committee member for ICON Consortium, COSMO Consortium, and EUMETNET AI program. He also contributes to supercomputing initiatives at HLRS and Gauss Centre.

• Numerical Analysis
• Sparse Grids
• Machine Learning
• +3 more

Prof. Dr. Michael Griebel is a leading academic at the Institute for Numerical Simulation (INS) at the University of Bonn , Germany. His work focuses on numerical methods for partial differential equations (PDEs), sparse grid techniques, and applications in computational science, machine learning, and materials modeling. Research Interests : Numerical analysis, sparse grids, high-dimensional approximation, tensor product methods, stochastic processes, and multiscale simulations. Collaborations : Co-author of over 50 publications with researchers in computational mathematics, fluid dynamics, and quantum chemistry. Prominent research trends include advancements in sparse grid methods for uncertainty quantification, kernel-based regression, and efficient solvers for ill-posed problems. His work on space-filling curves and domain decomposition enables scalable parallel algorithms for complex simulations. Scientific Contributions : Editor of 10+ volumes in Lecture Notes in Computational Science and Engineering . Co-developer of the EXAHD exascale sparse grid framework for plasma physics simulations. Foundational work on tensor product spaces and their applications to electronic structure theory. Innovations in adaptive wavelet solvers and parallel multigrid methods. Software and Projects : Key developer of sparse grid solvers and adaptive algorithms. His INS Preprint Series archives over 200 technical reports, reflecting his group’s impact on computational methods.