Peter Richtarik is a Professor at the King Abdullah University of Science and Technology (KAUST), specializing in Machine Learning, Optimization, and Federated Learning. He actively teaches courses such as Stochastic Gradient Descent Methods and mentors PhD and MS students like Konstantin Burlachenko, Kai Yi, and Lukang Sun. His research spans distributed optimization, parameter-efficient fine-tuning, and theoretical frameworks for non-convex and non-smooth problems. Recent work includes 2025 contributions to Bernoulli-LoRA (theoretical PEFT) and Gluon (LMO-based optimizers). He co-developed Thanos (block-wise pruning) and BurTorch (CPU-optimized training framework). His publications focus on communication efficiency ( ATA , LoCoDL ), differential privacy ( DP-RBCD ), and stochastic proximal methods. Richtarik received the Charles Broyden Prize for his work on quasi-Newton methods. He frequently presents at workshops like MLSS in Senegal and FLOW seminars.
Michael Knaus is a Junior Professor (Assistant Professor) in the Department of Economics within the Faculty of Economics and Social Sciences at the University of Tübingen, Germany. His office is located at Mohlstraße 36, 4th floor, room 415. He teaches graduate-level courses on causal inference and causal machine learning. Dr. Knaus specializes in the intersection of causal inference and machine learning, with particular expertise in Double Machine Learning methods. His research focuses on developing advanced statistical techniques to estimate treatment effects across various economic contexts including labor markets, finance, education, and health economics. His work bridges theoretical econometrics with practical applications, emphasizing methodological rigor and real-world relevance. His recent publications demonstrate a clear progression toward increasingly sophisticated methods for handling heterogeneous treatment effects and complex causal structures. His research shows strong integration of machine learning algorithms with causal inference frameworks to address challenging policy questions across multiple domains. Double Machine Learning based Program Evaluation under Unconfoundedness (The Econometrics Journal, 2022) Heterogeneous Employment Effects of Job Search Programmes: A Machine Learning Approach (Journal of Human Resources, 2022) How Does Post-Earnings Announcement Sentiment Affect Firms' Dynamics? (Journal of Financial Econometrics, 2024) Effect or Treatment Heterogeneity? Policy Evaluation with Aggregated and Disaggregated Treatments (2021) Dr. Knaus has made significant methodological contributions through his development of the causalDML R package, which implements Double Machine Learning methods for binary and multiple treatment effect estimation. His work has been published in top econometrics and economics journals and has gained recognition in the research community, with his GitHub repository accumulating 36 stars. He frequently collaborates with Michael Lechner, a leading researcher in causal inference and program evaluation. His teaching includes E464 Causal Inference and E463 Causal Machine Learning, both graduate courses that combine theoretical foundations with practical implementation using R. These courses prepare students for advanced research and data science roles requiring sophisticated causal reasoning skills, emphasizing hands-on application of methods to real-world problems.
László Kozma is an Assistant Professor at the Theoretical Computer Science group of the Institute of Computer Science (Freie Universität Berlin). He obtained his PhD from Saarland University under Raimund Seidel, followed by postdoctoral positions at Tel Aviv University and TU Eindhoven. His research focuses on self-adjusting data structures , adaptive algorithms , and combinatorial optimization with applications to problems like the Traveling Salesman Problem, binary search trees, and geometric data structures. Academic Affiliation: Freie Universität Berlin (since 2018) Education: PhD in Computer Science (Saarland University, 2016); postdoc at Tel Aviv University and TU Eindhoven. His work explores the intersection of data structures, combinatorial algorithms, and geometric methods. He has made significant contributions to problems involving pattern-avoidance in inputs, saddlepoint detection , and self-adjusting heaps . Key areas include: Adaptive algorithms for pattern-avoiding inputs Optimal tree and heap structures Geometric and stochastic approaches to optimization Complexity analysis of classical algorithms Recent publications highlight efficient solutions for exponential cut problems (ESA 2025), balanced TSP partitioning (EuroCG 2025), and randomized saddlepoint algorithms (ESA 2024). His research often bridges theory and practice, exemplified by the smooth heap implementation and fun projects like Recursi and Cuckoo Hashing visualization.
Wiebke Kirleis is Professor for Environmental Archaeology and Archaeobotany at the Institute for Pre- and Protohistory, Faculty of Arts and Humanities, Christian-Albrechts-Universität zu Kiel (CAU). She has held this position since May 2014 and was previously a junior professor at the same university from 2008 to 2014. She is actively involved in major interdisciplinary research projects, including the Cluster of Excellence ROOTS and the Collaborative Research Centre 1266 'Scales of Transformation', where she serves as co-speaker and technical platform speaker. PhD in Natural Sciences, University of Göttingen (2002) Diploma in Biology, University of Göttingen (1998) Studies in Botany, Anthropology, and Environmental History, University of Göttingen (1990–1998) Wiebke Kirleis' research centers on the interaction between humans and plants during the late Pleistocene and Holocene, with a focus on plant economy, domestication, agricultural development, and landscape transformation. Her work explores how wild and cultivated plants were used for nutrition, ritual, and technical purposes, and how agricultural innovations such as ploughing and manuring impacted Central European landscapes. She has a strong geographical focus on Northern and Central Germany, Southeastern Europe, the Near and Middle East, and Southeast and East Asia, particularly Indonesia. Her methodological expertise includes archaeobotany (analysis of charred and waterlogged plant remains), palynology (pollen and spore analysis from sediment cores), and anthracology (study of charcoal remains). Her recent research includes studies on broomcorn millet diffusion, early farming in Scandinavia and Russia, and plant use in Bronze Age settlements across Europe and Turkey. She leads and collaborates on numerous interdisciplinary projects, emphasizing the integration of botanical data into broader archaeological narratives. Her publications reflect a consistent trend in environmental archaeology, focusing on plant domestication, agricultural transitions, and human-induced landscape change. Keywords across her recent work include archaeobotany, palynology, Neolithic and Bronze Age subsistence, and environmental reconstruction. Her research spans subfields such as food security, cultural transmission of plant use, early monumentality, and socio-economic aspects of prehistoric agriculture. Board member, Cluster of Excellence ROOTS Co-speaker, CRC 1266 'Scales of Transformation' Speaker, ROOTS Technical Platform Board member, Johanna-Mestorf-Akademie Chair, Scientific Advisory Board, Dithmarschen Stone Age Park (since 2023) Full member, German Archaeological Institute (DAI) Elected member, Commission for Archaeology of Non-European Cultures (KAAK) Wiebke Kirleis has supervised numerous research projects funded by the DFG and other agencies, including the Graduate School 'Human Development in Landscapes'. She has advised students and collaborated internationally with institutions in Italy, Norway, Indonesia, and Australia. Her leadership roles in academic self-administration include serving on planning, budget, and equality committees at CAU Kiel. She is committed to science communication, having launched the digital exhibition 'Alles bleibt anders' and produced educational films on archaeobotany. She is also involved in mentoring female postdocs through the via:mento program. She leads the archaeobotanical component in the 'Superfood' module of the SFB 1266 digital exhibition and is actively involved in the ROOTS Cluster’s Subcluster 'Dietary ROOTS'. Her lab and research team operate within the Institute for Pre- and Protohistory, contributing to the university’s strong focus on interdisciplinary landscape archaeology.
Markus Lange-Hegermann serves as Professor of Mathematics and Data Science at Ostwestfalen-Lippe University of Applied Sciences (TH OWL) since 2018 and holds a board position at the Institute for Industrial Information Technology (inIT). His career bridges academic research and industrial applications, with expertise in translating machine learning theory into practical engineering solutions for automation and manufacturing sectors. His educational foundation includes a Diplom (Master equivalent) in Computer Mathematics from RWTH Aachen University (2004-2008) followed by a Dr. rer. nat. (PhD equivalent) in algorithmic differential algebra (2008-2014). Prior to academia, he gained industry experience at FEV GmbH as an R&D engineer (2014-2017) and P3 automotive GmbH as a Data Science Consultant (2017-2018). Lange-Hegermann's research centers on probabilistic machine learning with distinctive emphasis on physics-informed approaches. He develops Gaussian process methodologies that incorporate differential equations to model time dependencies, uncertainties, and physical constraints in industrial systems. His work enables robust data-based modeling and optimization for cyber-physical systems, with applications spanning predictive maintenance, process control, and quality assurance in manufacturing. Analysis of his 15 most recent publications (2024-2025) reveals consistent innovation in physics-integrated machine learning, particularly using Gaussian processes to solve partial differential equations and optimal control problems. The research demonstrates strong industrial applicability across domains including medical imaging, material science, automotive engineering, and brewing processes, with recurring themes of anomaly detection in time-series data and uncertainty-aware decision making. His scientific contributions have earned significant recognition: Forschungspreis TH OWL (2024) Top reviewer award at NeurIPS (2023) Outstanding reviewer award at NeurIPS (2021) Best poster award at Bosch AI CON (2019) Borchers Plakette for outstanding dissertation (2014) Springorum Denkmünze for outstanding diploma (2009) As chairman of the Data Science study program and vice chairman of undergraduate examination boards, Lange-Hegermann actively shapes academic curricula while supervising graduate theses. His governance roles include serving on professorship search committees at multiple institutions and contributing to examination regulations. He maintains active research funding through collaborations with industrial partners and reviews proposals for initiatives like It's OWL and 3IA Côte d’Azur. Lange-Hegermann leads the Mathematics and Data Sciences research group within inIT, fostering collaboration between theoretical machine learning and industrial automation. He co-founded AICOmmunityOWL and the Informatics Europe working group on Data Analysis and Reporting, while organizing machine learning reading groups and data science hackathons to bridge academic research with industrial problem-solving.
Professor Yann Disser is a faculty member in the Department of Mathematics at TU Darmstadt since 2021. He previously held an Assistant Professor (tenure-track) position at TU Darmstadt (2016-2021), a PostDoc position at TU Berlin (2012-2016), and a Visiting Professor role at Augsburg University (2015). His research spans Combinatorial Optimization , Online Algorithms , Graph Exploration , Computational Complexity , and Robust Optimization . Current position: Professor (W2), TU Darmstadt Previous: Assistant Professor (2016-2021), TU Darmstadt PostDoc & Habilitation: TU Berlin (2012-2016) His research focuses on algorithmic approaches to optimization problems, including: Combinatorial Optimization for problems like Steiner trees and knapsack variants Online Algorithms with applications to scheduling and transportation Graph Exploration by mobile agents and complexity bounds Computational Complexity of linear programming pivot rules Network Flows and geometric reconstruction Recent publications analyze incremental maximization with greedy methods, lower bounds for active-set methods , and universal circuit designs . His work appears in top venues like IPCO , ESA , and SODA . He advises a team of researchers including David Weckbecker, Farehe Soheil, and Alexander Birx. Current and past advisees often focus on algorithmic theory, with placements at institutions like HPI Potsdam and Merck.
Prof. Dr. Wolfgang Hillert is a leading physicist at the University of Hamburg , serving as the Bjørn-Wiik Professor for Accelerator Physics since 2016. Affiliated with the Institute of Experimental Physics under the Faculty of Mathematics, Informatics and Natural Sciences, he specializes in Accelerator Physics , Superconducting Accelerator Technology , and Free-Electron Lasers (FEL) . His work focuses on polarized electron beams, SRF cavity optimization, and gravitational wave detection methods. Education: Physics degree from University of Bonn (1987), Promotion in Atmospheric Physics (1992), Habilitation in Physics (2001) Leadership Roles: Head of Accelerator Physics Group (2016–present), Managing Director of Institute of Experimental Physics (2019–2021) Research Trends: His recent work spans superconducting RF cavities for gravitational wave detectors ( 2025 ), resonant slow extraction in electron boosters, and atomic layer deposition of superconducting thin films. Publications highlight advancements in beam dynamics , cryogenic systems , and terahertz generation . Teaching & Outreach: He has lectured on Accelerator Physics since 2002 and engaged in public science communication, including talks on Physics of Music (2005–2021) and teacher training programs at DESY. Labs & Collaborations: Leads the Accelerator Physics Group at DESY, collaborates on projects like XFELO and BGO-OD beamline , and contributes to international schools (CAS) and symposia.
Prof. Dr.-Ing. Richard Membarth is a Research Professor for System-on-a-Chip and AI at the Edge Computing at Technische Hochschule Ingolstadt (THI). He is affiliated with the Hardware-Software Co-Design group and holds a secondary position at the German Research Center for Artificial Intelligence (DFKI) Saarbrücken. Co-creator of DSL frameworks like AnyDSL and Hipacc Key contributor to MetaDL (AI metaprogramming) and PRIME (predictive rendering) His research bridges GPU computing , domain-specific languages , and compiler technology , with recent work on Vulkan SPIR-V compilation and device-driven SpMV algorithms . Notable awards include the HiPEAC Paper Award (2018) and multiple Best Paper Awards for his compiler frameworks.
Prof. Dr. Lubomir Banas is a full-time Professor at the Faculty of Mathematics , University of Bielefeld. His research focuses on numerical analysis of stochastic partial differential equations (SPDEs) , particularly in micromagnetism, phase field models, and stochastic games. He leads Subproject B03 in the SFB 1283 project 'Taming Uncertainty and Profiting from Randomness and Low Regularity in Analysis, Stochastics and Their Applications.' Research Interests: Numerical methods for SPDEs and singular-degenerate PDEs Adaptive finite element techniques and a posteriori estimates Phase field models (Cahn-Hilliard, obstacle potentials) Stochastic games with asymmetric information Computational micromagnetism and magnetostriction Self-organized criticality and nonlinear stochastic flows Recent work includes: 2025: Numerical approximation of biharmonic wave maps and stochastic games 2024: Sharp interface limits for stochastic Cahn-Hilliard equations 2023: Singular-degenerate SPDEs and a posteriori estimates 2022: Stochastic total variation flow and Hamilton-Jacobi-Bellman equations 2021: Nematic electrolytes and homogenization of two-phase flows He serves on examination boards for Bachelor's and Master's programs in Mathematics and Mathematical Physics, and supervises graduate students within the Bielefeld Graduate School in Theoretical Sciences . His publications (over 30) address convergence analysis, error estimation, and computational modeling in applied mathematics.
Daniel Kasper is a researcher at the University of Hamburg's Faculty of Education, specializing in Educational Science with a focus on International Educational Monitoring and Reporting. He currently serves as the National Project Manager for TIMSS 2027, continuing his leadership role from previous TIMSS cycles (2023, 2019). His work centers on large-scale educational assessments, statistical methodology, and primary education research. Dr. Kasper completed his Habilitation in Educational Science with special consideration of empirical educational research at TU Dortmund (2012-2020), followed by his PhD in Educational Science at the same institution (2010-2012). He earned his Diploma in Educational Science from the University of Münster (2001-2007). His research interests span evaluation of education systems, primary school research, educational disparities, and advanced statistical methods in education. He has developed significant expertise in TIMSS methodology, multilevel modeling, and analysis of large-scale assessment data, with numerous publications addressing methodological challenges in international comparative studies. Analysis of his 15 most recent publications reveals a strong focus on TIMSS methodology and results, particularly regarding mathematics and science competencies in primary education. His work demonstrates expertise in statistical methodology for educational assessment, with several publications developing and refining analytical techniques for large-scale data. A recurring theme is examining educational disparities related to student composition, socioeconomic factors, and gender differences. Dr. Kasper has served as National Project Manager for multiple TIMSS cycles (2019, 2023, 2027) and has been involved in PIRLS studies, demonstrating sustained leadership in major international educational assessments. His work bridges methodological innovation with practical application in educational monitoring. He teaches courses across all academic levels, including 'Introduction to Empirical Research Methods' for Bachelor students, 'Methods of Empirical Educational Research' for Master students, and advanced workshops on longitudinal scaling and the Rasch model for doctoral candidates. His teaching reflects his dual expertise in educational research methodology and statistical analysis.
Cédric Soutil is a Researcher at the Conservatoire National des Arts et Métiers (CNAM) , affiliated with the CEDRIC Laboratory. His work spans combinatorial optimization , integer programming , quadratic programming , and algorithm design , with a focus on solving complex optimization problems in scheduling and graph theory. Recent publications highlight his expertise in non-separable and non-convex quadratic integer programming , knapsack problems , and online computation . His research trends emphasize mathematical reformulations , upper bound algorithms , and optimization models for real-world applications like horse race scheduling and hydrogen production . He has collaborated extensively with researchers such as A. Houdayer , D. Quadri , and P. Tolla , contributing to over two decades of academic output in operations research and combinatorial optimization .
Tony Stillfjord is an Associate Professor at the Centre for Mathematical Sciences, Lund University, Sweden. He previously held postdoctoral positions at the Max Planck Institute for Dynamics of Complex Technical Systems and Chalmers/University of Gothenburg. Ph.D. and MSc in Numerical Analysis from Lund University Funded by WASP (Wallenberg AI, Autonomous Systems and Software Program) Research Interests : Numerical methods for partial differential equations, splitting schemes, stochastic optimization, and large-scale differential Riccati equations. His work focuses on developing low-rank approximations and robust optimization algorithms. Recent Publications highlight advancements in Lie/Strang splitting for operator-valued Riccati equations, stochastic descent methods, and GPU-accelerated splitting schemes. Software Contributions : Developed DREsplit (MATLAB package for differential Riccati equations) and BST20_CODE for stochastic optimization experiments. Contact : Office at MH:562E, Lund University. Email: tony.stillfjord@math.lth.se . URL: tonystillfjord.net
Prof. Dr. Thomas Eckert holds a professorship at the Faculty of Psychology and Education, Ludwig-Maximilians-Universität München (LMU Munich), specializing in historical socialization research and qualitative methods since 2003. His career spans from 1987-2002 as an employee at Freiburg University's Institute of Educational Sciences, followed by habilitation (1999) and promotion to full professor. Research Focus: Professionalization in pedagogical fields School quality research Educational equity and inclusion Quantitative/qualitative social research methods Lifelong learning systems Recent Publications: Address inclusion monitoring (2021-2025), educational equity frameworks (2022), and methodological innovations in statistical modeling (2019). His work bridges empirical pedagogy with policy analysis, examining transitions in educational systems and regional learning structures. Teaching Contributions: He has developed curriculum materials for distance learning courses and contributed to handbooks on adult education, educational statistics, and teaching practice organization.
Professor J. Debus is a distinguished academic in Radiation Oncology at Heidelberg University's Medical Faculty, with extensive research focused on particle therapy, medical physics, and cancer treatment optimization. His work spans clinical trials, radiobiology, and technical innovations in radiation delivery systems. Primary Affiliation: German Cancer Research Center (DKFZ), Heidelberg Research Focus: Particle therapy, radiation oncology, medical physics Key Collaborations: Mein S., Liew H., Tessonnier T., and other leading researchers in radiation oncology Professor Debus' research interests center on advancing particle therapy techniques including proton, carbon ion, and emerging modalities like helium and oxygen ion therapy. His work addresses critical challenges in radiation oncology such as normal tissue sparing, hypoxia-induced radioresistance, and precision treatment delivery. He has made significant contributions to understanding the biological effects of different radiation types and optimizing treatment protocols for various cancer types including head and neck cancers, brain metastases, and prostate cancer. His recent publications demonstrate leadership in clinical trials (GUARD, ESTRON, PROBASE) and technical innovations in radiation delivery systems, particularly in the emerging field of FLASH radiotherapy and ultra-high dose rate treatments. Professor Debus has published extensively on treatment planning optimization, radiation-induced biological effects, and imaging techniques for precise radiation delivery. Leading clinical trials in particle therapy Developing novel techniques for normal tissue protection Advancing understanding of radiation biology across different modalities Professor Debus has secured significant research funding for his work in radiation oncology and particle therapy. His research has contributed to clinical implementation of advanced treatment techniques at the Heidelberg Ion-Beam Therapy Center (HIT), one of the world's leading facilities for particle therapy. He supervises numerous doctoral students and postdoctoral researchers in the radiation oncology field. His laboratory and research team focus on translational research bridging basic radiobiology with clinical applications, with particular emphasis on optimizing treatment protocols for challenging tumor types and improving patient outcomes through precision radiation therapy.
Claudius Zibrowius is a Professor of Low-Dimensional Topology at Ruhr-University Bochum's Faculty of Mathematics. He leads the Topology Group, focusing on categorified knot invariants like knot Floer homology and Khovanov homology, leveraging Fukaya categories of surfaces. His work bridges Heegaard Floer theories and quantum invariant categorifications. Current group members include postdoc Dr. Chen Zhang and PhD student Luca Marchiori. Research highlights include ERC Starting Grant (ERC-2024) and DFG funding, as well as contributions to topics like Gordian distances, Conway spheres, and mutation invariance. Notable collaborations involve Artem Kotelskiy and Liam Watson. Teaching activities include Algebraic Topology I and outreach lectures. He co-organizes conferences like 'Categorification in Low Dimensional Topology' (2025). Awards include DFG Individual Grant (2022) and ERC Starting Grant (2024). His lab engages in computational projects like kht++, a C++ program for Khovanov invariants. Beyond academia, he sings in choirs and advocates for Palestinian rights, publicly opposing Israeli-Palestinian conflict complicity.
