Find Academic Supervisors

Kiwon UmView profile

Assistant Professor

• Physics-based Simulations
• Deep Learning
• Perceptual Evaluation of Simulations
• +3 more

Kiwon Um is a tenured Assistant Professor in the Computer Graphics group at Télécom Paris, France, since October 2019. He focuses on physics-based simulations and data-driven approaches using deep learning, with an emphasis on human visual perception in computer graphics and engineering applications. Education: Ph.D. in Computer Science and Engineering from Korea University His research explores effective simulation of natural phenomena through refined data utilization and develops reliable data acquisition methods for machine learning. He also investigates perceptual evaluation of simulations to advance numerical method understanding. Recent work trends include: (1) turbulence modeling with machine learning integration, (2) elastic material simulation stability, (3) fluid dynamics optimization, and (4) differentiable physics frameworks. His publications range from 2008 to 2025, covering topics like SPH solvers, porous shell simulations, and numerical method validation.

• Computational Statistics
• Machine Learning
• Monte Carlo Methods
• +4 more

Alain Durmus is a Professor at École Polytechnique, affiliated with the applied mathematics department (CMAP). His research focuses on computational statistics, machine learning, and stochastic methods, including Monte Carlo algorithms, Bayesian inference, and optimization. He explores topics such as Markov chain Monte Carlo (MCMC), stochastic approximation, and generative models. His work emphasizes theoretical guarantees for algorithms like Langevin Monte Carlo and Hamiltonian Monte Carlo, with applications to high-dimensional Bayesian inference and inverse problems. Key contributions include hypocoercivity analysis of piecewise deterministic MCMC processes, convergence guarantees for stochastic gradient methods, and the development of efficient sampling techniques. He has also contributed to Bayesian imaging and federated learning through works like the QLSD algorithm. Awarded the Best Student Paper Award at ICASSP 2020 for his work on the Sliced-Wasserstein distance. His teaching spans mathematical statistics, stochastic methods, and probability at École Polytechnique and ENS Paris-Saclay. He has also contributed to conferences and workshops on topics ranging from MCMC convergence to optimization in machine learning.

Pierre MarquisView profile

Professor

• Artificial Intelligence
• Knowledge Representation
• Automated Reasoning
• +6 more

Pierre Marquis is a distinguished Professor of Computer Science at Université d'Artois , affiliated with the Centre de Recherche en Informatique de Lens (CRIL-CNRS, UMR 8188) . Since December 2024, he has served as the vice-president for research and doctoral studies at Université d'Artois. His research focuses on Artificial Intelligence , particularly knowledge representation , automated reasoning , inconsistency handling , and knowledge compilation , with recent emphasis on Explainable AI (XAI) . Research Interests: Marquis's work spans foundational AI topics including abduction , induction , belief revision , and preference modeling . He has pioneered knowledge compilation techniques to optimize AI tasks and developed frameworks for reasoning under inconsistency through paraconsistent logics and argumentation. His EXPEKTATION chair (2020-2026) under France's national AI program drives his current focus on interpretable machine learning models. Scientific Awards: 2025: CNRS Silver Medal 2022: AAIA Fellow 2017: Senior Member of Institut Universitaire de France (IUF) 2009: EurAI (ECCAI) Fellow Doctoral Students: Mentoring Clément Lens (critical patient monitoring systems) and Mehdi Sabiri (data-knowledge integration for AI explanations). Collaborating with students like Louenas Bounia (formal XAI models) and Romain Wallon (pseudo-Boolean constraints). Grants & Projects: Leads the EXPEKTATION research chair (2020-2026) and participates in ANR PING/ACK (2019-2023), ANR THEMIS (2021-2025), CNRS IRP MAKC (2020-2024), and H2020 TAILOR (2020-2024). Previously led PIA4 MAIA (2023-2032) and Pint (2022-2023). Labs & Teams: Active in CRIL-CNRS, contributing to PyXAI (Python XAI library) and d4 (model counting), while mentoring teams on consensus belief merging and dynamic constraint processing .

• Fluid Dynamics
• Turbulence
• Magnetohydrodynamics
• +3 more

Wouter Bos is a Research Director at CNRS working at the Laboratory of Fluid Mechanics and Acoustics (LMFA) at École Centrale de Lyon, France. He leads research within the Turbulence & Instabilities team, focusing on fundamental aspects of fluid dynamics with applications spanning from plasma physics to epidemiology. His academic journey reflects a deep engagement with theoretical and computational fluid mechanics, particularly in turbulence phenomena. Dr. Bos's research interests center on fluid dynamics, with particular emphasis on turbulence in various contexts including two-dimensional flows, magnetohydrodynamics, plasma physics, and statistical mechanics of fluids. His work explores fundamental questions about energy transfer, coherent structures, and statistical properties of turbulent flows. He has made significant contributions to understanding turbulence without vortex stretching, two-dimensional turbulence, and the application of fluid dynamics principles to epidemiological modeling. Analysis of his recent publications reveals a strong focus on theoretical and computational approaches to turbulence. His work spans from fundamental questions about equilibrium states in two-dimensional turbulence to practical applications in plasma confinement and epidemic modeling. The publications demonstrate consistent innovation in turbulence theory, with particular attention to statistical mechanics approaches, spectral analysis, and the development of reduced-order models for complex fluid phenomena. His research shows growing interdisciplinary connections, especially between fluid dynamics and epidemiology as evidenced by his work on modeling the spread of infectious diseases. Dr. Bos actively supervises doctoral students and postdoctoral researchers, with recent students including Tong Wu, Ryo Araki, Smiron Varghese, Wesley Agoua, and Bruce (Xi Yuan) Yin. He participates in collaborative research projects such as the ANR CM2E project (2021-2025) on Characteristic Mapping Method for the Euler Equations, working with researchers from Aix-Marseille University and McGill University. His laboratory work involves both theoretical analysis and computational simulations, with applications ranging from fundamental fluid mechanics to practical problems in energy research (particularly related to ITER and fusion plasma physics) and public health. The interdisciplinary nature of his research demonstrates the broad applicability of fluid dynamics principles across seemingly disparate scientific domains.

Jamal AtifView profile

Professor

• Artificial Intelligence
• Responsible AI
• Privacy Preservation in Machine Learning
• +4 more

Jamal Atif is a Professor at Paris-Dauphine University and holds multiple significant leadership positions including Project Manager for 'Data Science and Artificial Intelligence' at the Institute of Information Sciences and their Interactions (INS2I) of the CNRS, Deputy Scientific Director of 3IA PRAIRIE, Head of the MILES team/project at LAMSADE (UMR CNRS-Université Paris-Dauphine), Co-leader of the Transverse Artificial Intelligence Program at PSL University, and Director of the Dauphine Numérique program. Professor Atif's primary research focuses on the foundations of responsible artificial intelligence, with specific expertise in privacy preservation in machine learning, robustness of deep learning algorithms to malicious attacks, causality, and explainability. His work bridges theoretical foundations with practical applications in security and reliability of AI systems. He has developed innovative approaches to address adversarial vulnerabilities in machine learning models and has made significant contributions to privacy-preserving techniques in data analysis. His publication record demonstrates a consistent focus on robust and trustworthy AI systems, with recent work exploring differential privacy in clustering, adversarial robustness, and explainable AI. The research spans theoretical foundations in logic and knowledge representation to practical applications in finance, healthcare, and computer vision. His publications appear in top-tier venues including Machine Learning journal, Neural Information Processing Systems, and International Joint Conferences on Artificial Intelligence. Scientific Awards: Recipient of two awards from the North American Society of Radiology for his thesis work Professor Atif has co-supervised or is currently supervising around fifteen doctoral students, demonstrating his commitment to mentoring the next generation of AI researchers. His leadership extends to directing major institutional programs including Dauphine Numérique and the Transverse Artificial Intelligence Program at PSL University, where he shapes strategic research directions in AI. He leads the MILES team/project at LAMSADE, which focuses on foundational aspects of machine learning and artificial intelligence. The team's research spans theoretical aspects of learning algorithms to practical applications requiring robust and reliable AI systems, with particular emphasis on security and privacy considerations in modern machine learning deployments.

Sonia FlissView profile

Professor

• Wave Propagation
• Mathematical Analysis
• Numerical Analysis
• +5 more

Sonia Fliss is a Full Professor at ENSTA Paris , affiliated with the Department of Applied Mathematics and the POEMS laboratory (UMR CNRS-INRIA-ENSTA) . She teaches applied mathematics courses on partial differential equations (PDEs) , finite element methods , and periodic homogenization to undergraduate and graduate students. Doctor in Applied Mathematics (2009) Authorized to supervise research (2019) Research Interests : Sonia Fliss specializes in the modeling and numerical analysis of wave propagation in periodic, quasi-periodic, and random media . Her work includes transparent boundary conditions , guided waves , and asymptotic methods for acoustic, electromagnetic, and elastic wave phenomena . Recent Publications highlight her contributions to the Half-Space Matching Method , edge states in honeycomb structures , and scattering problems in unbounded domains . Her numerical techniques address multi-scale waveguides and time-harmonic propagation . Laboratory : As a member of the POEMS team, she collaborates on interdisciplinary projects involving mathematical analysis , computational physics , and engineering applications in domains like defence, energy, and transport .

Sébastien TordeuxView profile

Associate Professor

• Numerical Analysis
• Wave Propagation
• Electromagnetic Theory
• +3 more

Sébastien Tordeux is an Associate Professor at the University of Pau, affiliated with the Magique 3D research team in the Faculty of Sciences. His research focuses on numerical methods for wave propagation, particularly Trefftz methods, asymptotic expansions, and electromagnetic/acoustic scattering. He holds a PhD in Applied Mathematics from the University of Versailles Saint-Quentin and has held positions at INSA-Toulouse and ETH Zurich. He has advised multiple PhD students and organized conferences such as the 2017 conference in honor of Abderrahmane Bendali. His work emphasizes high-order numerical modeling and efficient solvers for wave equations in complex media. Education: PhD in Applied Mathematics (2004, Université de Versailles), DEA M2SAP (2001), ENSTA Engineer (1998-2001). Professional roles include Chair of Excellence in Numerical Analysis (INRIA-UPPA 2010-2015), and leadership in Master’s program administration and national academic committees (CNU 26). Research interests span Trefftz methods for time-harmonic models, numerical solvers reducing pollution effects, and asymptotic modeling for small obstacle scattering. Recent contributions include quasi-Trefftz methods for electromagnetic systems and iterative approaches for 3D wave simulations. He has supervised PhD students working on topics like antenna patch models, perforated plate acoustics, and multiscale electromagnetic diffraction. His collaborative projects involve institutions like ONERA and TU Berlin, focusing on high-accuracy wave modeling and computational methods.

Maryam Mehri DehnaviView profile

Associate Professor

• High-Performance Computing
• Cloud Computing
• Machine Learning
• +3 more

Maryam Mehri Dehnavi is an Associate Professor in the Department of Computer Science at the University of Toronto and a Principal Research Scientist at NVIDIA. She holds the Canada Research Chair in Parallel and Distributed Computing and leads the ParaMathics research group. Research focuses on high-performance computing , machine learning , sparse matrix optimizations , and compiler design for heterogeneous systems. Her work develops domain-specific languages , scalable numerical libraries , and auto-vectorization techniques for cloud and GPU platforms. Recent publications address LLM compression , sparse code translation , GPU kernel synchronization , and control flow optimization . Scientific recognition: Ontario Early Researcher Award (2021), NSF CRII Grant, NSERC New Frontiers in Research Fund. Current students: Mushegh Shahinyan , Martin Phan , Maryam Haghifam , and others. Former advisees: Kazem Cheshmi (NJIT), Zachary Blanco (MIT Lincoln Lab), Yuanxi Li (Amazon).

• Numerical Analysis
• Partial Differential Equations (PDEs)
• Finite Element Methods
• +3 more

Ari Rappaport is a postdoctoral researcher at the Unité de Mathématiques Appliquées (UMA) of ENSTA Paris, part of the Propagation des Ondes, Étude Mathématique et Simulation (POEMS) laboratory. His research focuses on numerical methods for partial differential equations, particularly hybridizable discontinuous Galerkin (HDG) methods for time-harmonic wave propagation and Maxwell equations. He also works on adaptive regularization, error estimation, and high-performance computing implementations. His recent work includes contributions to iterative solvers, finite element methods, and parallel computing strategies. Education and affiliations: Postdoctoral researcher at ENSTA Paris (since 2023?), with involvement in projects like CIEDS ElectroMath (AID/DGA). Collaborates with institutions like Inria and Université de Lille. Research interests emphasize computational electromagnetics, numerical analysis of PDEs, and the development of robust numerical schemes for complex physical models. Key contributions include advancements in HDG methods for Maxwell equations, adaptive error estimation frameworks, and scalable parallel implementations. His work bridges theoretical analysis with practical computational tools for engineering and scientific applications.

Romain RaveauxView profile

Associate Professor

• Image Analysis
• Machine Learning
• Structural Pattern Recognition
• +5 more

Romain Raveaux is an Associate Professor at the LIFAT Computer Science Laboratory, University of Tours, affiliated with Polytech Tours. His research focuses on Image Analysis, Machine Learning, Structural Pattern Recognition, Graph Matching, Graph Neural Networks, Discrete Optimization, Reinforcement Learning, and Transfer Learning . Email: romain.raveaux@gmail.com , romain.raveaux@laposte.net Address: 64 av. Jean Portalis, Tours, France, 37200 Phone: +33 (0)2 47 36 14 27 Research Interests Graph Matching and Neural Networks Discrete Optimization for Pattern Recognition Transfer Learning in Graph-Based Models Historical Document Analysis Scientific Trends His recent work bridges Graph Neural Networks with Mixed-Integer Programming , focusing on Image Semantic Segmentation and Graph Cycle Detection . Earlier studies emphasize Genetic Algorithms for graph classification and Graph Edit Distance optimization in pattern recognition.

Christophe KervazoView profile

Assistant Professor

• Sparse Blind Source Separation
• Nonnegative Matrix Factorization
• Hyperspectral Imaging
• +4 more

Christophe Kervazo is an Assistant Professor (Maître de Conférences) at Télécom Paris, France, affiliated with the IMAGES group under the Image, Data, Signal (IDS) department . His research focuses on sparse blind source separation, nonnegative matrix factorization, hyperspectral imaging, and optimization techniques for remote sensing and biomedical applications. Education: Engineering degree from Supélec (2015), Master of Science from Georgia Institute of Technology (2016), PhD in Signal and Image Processing from Université Paris Saclay (2019). His work spans deep learning for inverse problems (including deep unrolling techniques), remote sensing (hyperspectral imaging and SAR), and uncertainty quantification . Recent publications address synthetic data training for medical imaging, distributed sparse BSS, and nonlinear component separation. Collaborators include institutions like CEA Saclay, Université de Mons, and ONERA. Current students include PhD candidates working on topics such as digital breast tomosynthesis, hyperspectral unmixing, and SAR image reconstruction. Former students and interns have contributed to projects involving plug-and-play methods, unrolling algorithms, and implicit regularization. He is involved in teaching and research projects, including collaborations with Airbus and ONERA on hyperspectral imaging and spectral band optimization. His lab, LTCI (Information Processing and Communication Laboratory), supports interdisciplinary work in signal and image processing.

Woosuk LeeView profile

Researcher

• Program Analysis
• Program Synthesis
• Computer Science
• +1 more

Woosuk Lee is a researcher at Hanyang University in South Korea, specializing in Program Analysis and Program Synthesis . He has been actively involved in academic activities since 2015, contributing to major conferences like OOPSLA, PLDI, POPL, and ESEC/FSE. Current affiliation: Hanyang University Research interests: Program Analysis, Program Synthesis His research focuses on automating program generation through inductive methods and abstract interpretation. Key contributions include: Advancing iterative forward-backward abstract interpretation for program synthesis (2023) Structurally recursive functional program synthesis (2023) Homomorphic encryption circuit optimization via term rewriting (2020) Data-driven feedback systems for programming education (2021) He has served on program committees for PLDI (2024), POPL (2022), and SPLASH (2025), while also chairing workshops at PLDI 2026. His work bridges theoretical computer science with practical software engineering applications.

Patrick Le TallecView profile

Professor

• Computational Mechanics
• Nonlinear Mechanics
• Domain Decomposition Methods
• +5 more

Patrick Le Tallec is a Professor of Mechanics at École Polytechnique in France, where he currently serves as Dean of the Bachelor Program and is a member of the M3DISIM project. His distinguished academic career spans multiple institutions including Université Paris Dauphine, INRIA (French National Institute for Research in Digital Science and Technology), and international universities such as Stanford University, University of Wisconsin, and Shanghai Jiao Tong University. He has held leadership positions including Vice President for Education and Head of the Laboratory of Solid Mechanics at École Polytechnique. His educational background includes: Graduate from École Polytechnique Ph.D. in Engineering Mechanics from The University of Texas at Austin (1980) Thèse d'Etat in Applied Mathematics from Université Pierre et Marie Curie in Paris (1981) Professor Le Tallec's research focuses on computational mechanics and applied mathematics with expertise in nonlinear mechanics, domain decomposition methods, and multiscale modeling. His work bridges theoretical mathematics with practical engineering applications, particularly in material science and fluid-structure interactions. He has developed advanced numerical methods for elasticity, viscoelasticity, and fluid dynamics with applications in industrial manufacturing and biomedical engineering. His recent publications demonstrate progression from foundational numerical methods to sophisticated multiscale approaches addressing complex engineering challenges in material science. The research shows particular emphasis on rubber mechanics, fatigue analysis, and computational methods for nonlinear structures, reflecting his ongoing commitment to solving real-world engineering problems through mathematical innovation. His scientific honors include: CISI award in Scientific Computing Prize Blaise Pascal of the French Academy of Sciences Chevalier des Palmes Académiques Chevalier de la Légion d'Honneur Officier de l'Ordre National du Mérite Professor Le Tallec has directed over 40 Ph.D. students from 10 different nationalities, demonstrating significant impact in academic mentoring. His research has been supported through extensive collaborations with industrial partners including Michelin, PSA Group, and Dassault Aviation, as well as scientific advisory roles at the French Alternative Energies and Atomic Energy Commission. He has served as president of the French Society of Applied and Industrial Mathematics and held editorial positions with leading journals in his field. His laboratory work centers around computational mechanics research, particularly through the M3DISIM project at École Polytechnique. His research team brings together mathematicians, engineers, and computer scientists to develop innovative solutions for complex problems in material science and structural mechanics, with applications ranging from industrial tire manufacturing to biomedical engineering.

• Microelectronics
• Radiation Effects on Electronics
• Semiconductor Reliability
• +3 more

Jean-Luc Autran is an Exceptional University Professor (PRCE2) at Aix-Marseille University, affiliated with the Department of Detection, Radiation and Reliability (DETECT) within the Faculty of Sciences. Since July 2023, he has been temporarily assigned to the University of Rennes for managerial roles. His research focuses on radiation effects in microelectronics, particularly soft errors caused by atmospheric neutrons, muons, protons, and terrestrial radiation in nanoscale devices. Key Research Areas: Microelectronics Reliability, Single-Event Effects, Atmospheric Radiation, Neutron Interactions, Muon Physics, Radiation-Hardened Design. Scientific Leadership: As an Honorary Member of the Institut Universitaire de France (since 2003), he leads multidisciplinary efforts from radiation metrology to multi-physics circuit simulation via tools like GEANT4, SRIM, and NGSPICE. His work spans 25+ years, evolving from quantum transport in nano-MOSFETs (1998-2008) to atmospheric radiation effects (2005-present). Recent Contributions: 2025 studies on ultrawide-bandgap semiconductors, deep learning-based radiation simulations, and JET Tokamak neutron experiments. He pioneers a multi-scale Single-Event Effect simulation framework for decananometer CMOS, integrating particle physics, device modeling, and system-level error rate prediction. Scientific Awards: Honorary Member, Institut Universitaire de France (2003 class). Educational Impact: Teaches graduate courses in quantum mechanics simulation, nanoelectronics reliability, and radiation detection at Aix-Marseille University, with lectures delivered in English for international audiences.

Axel ModaveView profile

Research Fellow

• Numerical simulation
• Wave propagation
• Finite element methods
• +2 more

Axel Modave is a CNRS Research Fellow at the POEMS laboratory (CNRS, Inria, ENSTA) and the Applied Mathematics Unit (UMA) of ENSTA Paris in Palaiseau, France. He obtained his PhD from the University of Liège (Belgium) in 2013 and his HDR (post-doctoral degree) from IP Paris in 2024. His academic career includes postdoctoral positions at UCLouvain (Belgium), Rice University (USA), and VirginiaTech (USA). Current Affiliation: CNRS Research Fellow, POEMS Laboratory & Applied Mathematics Unit (UMA), ENSTA Paris Education: PhD (2013, ULiège), HDR (2024, IP Paris) Research Focus: Numerical simulation of wave propagation (acoustic, electromagnetic, elastic), high-order finite element methods, domain decomposition techniques, and high-performance computing. Email: axel.modave@ensta-paris.fr | axel.modave@ensta.fr His research explores advanced numerical methods for wave propagation problems, including the ANR WavesDG project on wave-specific DG finite element methods for time-harmonic problems. His recent publications focus on hybridizable discontinuous Galerkin methods for electromagnetic and acoustic wave problems, domain decomposition preconditioners, and GPU-accelerated simulations. He teaches courses on optimization and high-performance computing.