Research
Below, you can find a description of several areas of research the SCOPA lab has contributed to. Code related to our research is deployed at https://github.com/scopagroup.
Machine Learning
We develop mathematical frameworks that combine neural networks with physical models for parameter estimation and uncertainty quantification in inverse problems. A key focus is on Bayesian inverse problems governed by systems of nonlinear ordinary differential equations. We have shown that neural networks can overcome challenges posed by strong nonlinearities and sharp gradients by learning reconstruction maps from observational data. Our approach enables simultaneous estimation of model parameters, noise parameters (including autocorrelated additive noise and noise modeled via stochastic differential equations), and the covariance matrix of the posterior distribution – all from a single forward network evaluation.
We have worked on lifted training methods for deep neural networks, which reformulate the nested optimization into higher-dimensional constrained optimization problems amenable to block-coordinate descent with accelerated and stochastic variants. This framework supports non-differentiable proximal activations, improved conditioning, and extends naturally to inverse problems in imaging.
We have proposed efficient clustering on Riemannian manifolds. We have developed a novel approach based on Frechet maps that embeds high-dimensional, non-Euclidean manifold data (such as symmetric positive definite matrices) into lower-dimensional Euclidean spaces, enabling standard k-means clustering with runtime reductions of up to two orders of magnitude compared to intrinsic manifold-based approaches.
Optimization Porblems Governed by Complex Dynamical Systems
We develop fast iterative methods for large-scale PDE-constrained optimization problems. Our work includes the design and analysis of efficient solvers for optimal control problems governed by hyperbolic transport equations, with applications in image registration, shape matching, and biophysical modeling. We study preconditioned Newton–Krylov methods, operator-splitting schemes, and novel acceleration techniques (including nonlinear GMRES variants) that exploit the structure of the underlying PDE constraints. We have developed fast methods for evaluating forward and adjoint operators in transport dominated problems. We also develop and studied tensorial reduced-order models for parametric coupled reaction-diffusion systems.
Diffeomorphic Image Registration
We develop scalable, GPU-accelerated algorithms for diffeomorphic image registration. Image registration is a nonlinear inverse problem: given two images of the same object or scene, we seek a spatial mapping that brings one image into alignment with the other. In diffeomorphic image registration, the admissible transformations are restricted to maps that are smooth, one-to-one, and have a smooth inverse. We formulate this as a variational problem governed by transport equations and solve it using an inexact, globalized Gauss–Newton–Krylov method.
Our software tool CLAIRE (Constrained Large Deformation Diffeomorphic Image Registration) implements these algorithms. CLAIRE features mixed-precision, hardware-accelerated computational kernels for optimal throughput on multi-node, multi-GPU architectures (default) as well as multi-core CPU systems. It can solve clinically relevant registration problems in under four seconds on a single GPU and scales to large-scale 3D imaging problems with billions of voxels. We have designed a novel alternating nonlinear GMRES acceleration method that achieves runtimes up to 5x faster than state-of-the-art Newton–Krylov methods while relying solely on first-order derivatives. We are also exploring novel formulations based on variational principle grid generation methods that construct non-folding grids with prescribed Jacobian determinants and provide accurate inverse transformations within the diffeomorphism group.
CLAIRE is released under the GNU General Public License and is available at github.com/andreasmang/claire. The deployment page is andreasmang.github.io/claire.
Diffeomorphic Shape Matching and Shape Analysis
We develop numerical algorithms for studying the variability of shapes and shape deformations through the lens of geodesic flows of diffeomorphisms. We formulate diffeomorphic shape matching as an ODE-constrained optimization problem, where the ODE constraint models the flow of diffeomorphisms and the velocity is modeled in a reproducing kernel Hilbert space. We use operator-splitting methods (Douglas–Rachford) for numerical optimization. Our algorithms enable rotation-invariant automatic classification of 3D surfaces using registration-based feature vectors, with data augmentation via diffeomorphic interpolation and random flows of smooth diffeomorphisms.
Parallel and High-Performance Computing
Our algorithms are designed to run efficiently on modern high-performance computing architectures. CLAIRE uses MPI for distributed-memory parallelism and CUDA for GPU acceleration, and has been demonstrated to scale on several supercomputing platforms including multi-node, multi-GPU systems. We have been awarded allocations on the Neocortex supercomputer at the Pittsburgh Supercomputing Center to explore Cerebras-accelerated deep neural networks for scientific computing applications.
Support
- 2025: NSF CBMS (DMS-2430460)
- 2025: UH DOR GEAR Award
- since 2022: NSF CAREER Award (Computational Mathematics) (DMS-2145845)
- 2023: Allocation on the Neocortex Supercomputer at the Pittsburgh Supercomputing Center
- 2020–2023: NSF Applied Mathematics (DMS-2009923)
- 2020–2023: NSF Computational Mathematics (DMS-2012825)
- 2019–2022: NSF CDS&E-MSS (DMS-1854853)
- 2019: NVIDIA Corporation GPU Grant Program (Accelerated Data Science Call)
- 2018: SIMONS Foundation Collaboration Grants for Mathematicians (Award 586055)
Talks & Posters
- A. Mang: A generalized alternating nonlinear GMRES acceleration method. Contributed talk at SIAM Texas-Louisiana Sectional Meetings (SIAM TX-LA; Session: High-Performance Solvers and Rapid PDE-Constrained Optimization), University of Texas at Austin, Austin, TX, US, 2025
- A. Mang: Numerical methods for PDE-based diffeomorphic image registration. Contributed talk at SIAM Annual Meeting (SIAMAN25; Session: Image Analysis and Learning with Variational Models and PDEs), Montreal, CA, 2025.
- A. Mang: CLAIRE: Constrained large deformation diffeomorphic image registration. Contributed talk at International Conference on Continuous Optimization (ICCOPT; Session: Recent Advances on PDE-constrained optimization packages and libraries), University of Southern California, Los Angeles, CA, US, 2025.
- A. Mang: Data- and model-driven approaches for solving inverse problems. Invited talk (hosts: D. Mishra (TAMU), M. Zhong (UH), X. Chen (TAMU), D. Casey (TAMU)) at the Scientific Machine Learning (SciML) Summer School 2025 at the Institute of Data Science, Texas A&M, College Station, TX, US, 2025.
- P. Amiri: Transport-Based Variational Bayesian Inference. Contributed talk at SIAM Conference on Computational Sciences and Engineering (SIAM CSE; Session: Decision Making for Coupled Systems), Fort Worth, TX, US, 2025.
- C. Jannatul: Efficient Numerical Methods for PDE-constrained Optimization Problems in Diffeomorphic Image Registration. Contributed talk at SIAM Conference on Computational Sciences and Engineering (SIAM CSE; Session: Methods for Image Processing and Numerical Modeling in Computational Medicine), Fort Worth, TX, US, 2025.
- A. Mang: Bayesian Inference for Large Scale Inverse Problems Governed by Hyperbolic Dynamical Systems. Contributed talk at SIAM Conference on Computational Sciences and Engineering (SIAM CSE; Session: Investigating Inverse Problems using Bayesian Inference: Challenges and Advances), Fort Worth, TX, US, 2025.
- A. Mang: Efficient numerical methods for inverse problems governed by transport equations. Contributed talk at 3rd IACM Digital Twins in Engineering Conference (DTE 2025) & 1st ECCOMAS Artificial Intelligence and Computational Methods in Applied Science (DTE & AICOMAS 25; Session: Inverse Problems and Data Assimilation for Digital Twins); Paris, FR, 2025.
- A. Mang: Fast iterative methods for large-scale initial value control problems. Contributed talk at SIAM Texas-Louisiana Sectional Meetings (SIAM TX-LA; Session: Recent Developments in Computational Inversion and Reduced Order Modelling), Baylor University, Waco, TX, US, 2024.
- A. Mang: Deep learning for Bayesian inverse problems governed by nonlinear ODEs. Contributed talk at SIAM Conference on Mathematics of Data Science (MDS24; Session: Recent Advances in Scientific Deep Learning); Atlanta, GA, US, 2024.
- A. Mang: Fast iterative methods for large-scale initial value control problems. Contributed talk at the Modeling and Optimization: Theory and Applications Conference (MOPTA; Session: Computational and Theoretical Methods for High-dimensional Optimization Problems), Lehigh University, Bethlehem, PA, USA, 2024.
- A. Mang: Efficient numerical schemes for uncertainty quantification in diffeomorphic image registration governed by transport equations. Contributed talk at the International Conference on Computational and Mathematical Biomedical Engineering (CMBE24; Session: Inverse Problems and Uncertainty Quantification in Biological and Medical Applications), Arlington, VA, USA, 2024.
- J. Rudi: Data representations for parameter estimation with deep learning models for a dynamical system. Contributed talk at International Conference on Computational and Mathematical Biomedical Engineering (CMBE24; Session: Inverse Problems and Uncertainty Quantification in Biological and Medical Applications), Fairfax, VA, US, 2024.
- A. Mang: CLAIRE: Scalable algorithms for diffeomorphic image registration. Contributed talk at the SIAM Conference on Imaging Sciences (IS24; Session: Model- and Data-Driven Approaches in Motion Analysis), Atlanta, US, 2024.
- J. Chhoa: Efficient Numerical Methods for Optimization Problems Governed by Transport Equations. Contributed talk at the SIAM Conference on Imaging Sciences (IS24; Session: Frontiers in Deep Image Reconstruction, Restoration Across Diverse Modalities), Atlanta, US, 2024.
- J. Y. Kim: Fast Iterative Solvers for PDE-constrained Optimization in Diffeomorphic Image Registration. Contributed talk at the SIAM Conference on Imaging Sciences (IS24; Session: Shapes, Manifolds and Geometry in Imaging), Atlanta, US, 2024.
- A. Mang: Fast iterative solvers for initial value control problems with application to diffeomorphic image registration. Contributed talk at the INFORMS Optimization Society Conference (IOS; Session: Optimization of Complex Physics-Based Systems), Houston, TX, 2024.
- A. Mang: CLAIRE: Scalable Algorithms for Diffeomorphic Image Registration. Contributed talk at the SIAM Conference on Uncertainty Quantification (UQ24; Session: Computational Tools for Large-Scale Inverse Problems and UQ), Trieste, IT, 2024.
- A. Mang: Efficient algorithms for inverse problems governed by dynamical systems. Invited talk (host: K. B. Nakshatrala) at the Department of Civil and Environmental Engineering, University of Houston, TX, 2023.
- A. Mang: Fast algorithms for nonlinear optimal control of geodesic flows of diffeomorphisms. Contributed talk at the U.S. National Congress on Computational Mechanics (USNCCM7; Session: Recent Advances in Large-Scale Optimal Engineering Design), Albuquerque, NM, 2023.
- A. Mang: Shape classification through the lens of geodesic flows of diffeomorphisms. Invited talk at workshop entitled “Leveraging Model- and Data-Driven Methods in Medical Imaging” at Banff International Research Station for Mathematical Innovation and Discover, CA, 2023.
- A. Mang: Scalable algorithms for inverse problems governed by dynamical systems. Invited talk at DSI’s webinar at the Hewlett Packard Enterprise Data Science Institute, University of Houston, Houston, TX, 2023.
- A. Mang: Deep neural networks for Bayesian inverse problems governed by nonlinear ODEs. Invited talk at workshop entitled Learning for Inverse Problems at Istituto Nazionale di Alta Matematica, Rome, IT, 2023.
- A. Mang: Fast algorithms for PDE-constrained optimization under uncertainty. Contributed talk at SIAM Conference on Optimization (OP23; Session: Challenges in Inverse Problems with Massive Data), Seattle, US, 2023.
- A. Mang: Fast algorithms for optimal control problems governed by geodesic flows of diffeomorphisms. Invited colloquium talk (host: J. Rudi) at the Department of Mathematics, Virginia Tech, Blacksburg, VA,US, 2023.
- A. Mang: Efficient numerical methods for optimal control problems governed by geodesic flows of diffeomorphisms. Invited talk (host: S. Foucart) at Center for Approximation and Mathematical Data Analytics, Texas A&M University, College Station, TX US, 2023.
- A. Mang: Fast algorithms for optimal control problems governed by geodesic flows of diffeomorphisms. Invited talk (host: S. Shontz) at Mathematical Methods and Interdisciplinary Computing Center (MMICC) at the University of Kansas, Lawrence, KS, US 2023.
- A. Mang: Numerical methods for PDE-constrained optimization problems governed by hyperbolic equations. Invited colloquium talk (host: Juan R. Romero) at Department of Mathematical Sciences, University of Puerto Rico, US, 2023.
- A. Mang: CLAIRE: Scalable multi-GPU algorithms for diffeomorphic image registration in 3D. Invited ACMD Seminar talk (host: Gunay Dogan) at National Institute of Standards and Technology, Gaithersburg, MD, US, 2023.
- A. Mang: Fast algorithms for inverse problems governed by transport equations. Contributed talk at AMS Sectional Meeting (Session: Recent Developments on Analysis and Computation for Inverse Problems for PDEs) in Atlanta, GA, US, 2023.
- A. Mang: Deep learning for Bayesian inverse problems governed by nonlinear ODEs. Contributed talk at SIAM Conference on Computational Science and Engineering (CSE23; Session: Uncertainty Quantification for Data-Intensive Inverse Problems and Learning) in Amsterdam, NL, 2023.
- J. Chhoa: CLAIRE: A framework for constrained large deformation diffeomorphic image registration. Invited talk at Texas Women In Mathematics Symposium in Austin, TX, 2023.
- J. Y. Kim: Numerical methods for Bayesian inference for inverse transport problems. Contributed talk at Joint Mathematics Meetings (JMM23) in Boston, MA, 2023.
- A. Mang: CLAIRE: Scalable Multi-GPU Algorithms for Diffeomorphic Image Registration in 3D. Contributed talk at Joint Mathematics Meetings (JMM23) in Boston, MA, 2023.
- A. Mang: Fast algorithms for nonlinear optimal control of geodesic flows of diffeomorphisms. Invited talk (host: Harbir Antil) at CMAI Colloquium at the Center for Mathematics and Artificial Intelligence, George Mason University, Fairfax, VA, US, 2022.
- A. Mang: Randomized algorithms for preconditioning and uncertainty quantification in inverse transport problems. Contributed talk at SIAM Conference on Mathematics of Data Science 2022 (Session: Randomized Methods in Large-Scale Inference and Data Problems), San Diego, CA, US, 2022.
- A. Mang: Fast algorithms for initial value control problems. Contributed talk at SIAM Conference on Imaging Sciences (IS22; Session: Partial Differential Equations and Control Problems); virtual conference, 2022.
- H. Dabirian: Automatic classification of shapes and shape deformations in 3D. Contributed talk at Joint Mathematics Meetings (JMM22); virtual conference, 2022.
- N. Himthani: CLAIRE: A scalable multi-GPU solver for diffeomorphic image registration in 3D. Contributed talk at SIAM TX-LA Annual Meeting (SIAM TX-LA21; Session: Mathematics and Computation in Biomedicine); South Padre Island, TX, 2021.
- M. Brunn: High-Speed Image Registration for Large-Scale Applications with CLAIRE. Invited talk (host: Barbara Gris) at Workshop on Registering Medical Images, Paris, FR, 2021.
- J. Y. Kim: Efficient numerical methods for initial value control problems. Contributed talk at SIAM Annual Meeting (AN21; Session: Fast Analysis Based Algorithms for Solution of Forward and Inverse Problems); virtual conference, 2021.
- A. Mang: Uncertainty quantification in diffeomorphic image registration. Contributed talk at SIAM Annual Meeting (AN21; Session: Uncertainty Quantification Strategies for Data-Driven, Large-Scale Problems); virtual conference, 2021.
- M. Brunn: Fast multi-GPU diffeomorphic image registration for large-scale applications. Contributed talk at US National Congress on Computational Mechanics (USNCCM16; Session: Imaging-Based Methods in Computational Medicine); virtual conference, 2021.
- A. Mang: CLAIRE: Scalable multi-GPU algorithms for diffeomorphic image registration in 3D. Contributed talk at SIAM Conference on Optimization (OPT21; Session: Large-Scale Optimization for Inverse Problems and Learning in Medical Imaging); virtual conference, 2021.
- A. Mang: Uncertainty quantification for inverse transport problems. Contributed talk at SIAM Conference on Computational Science and Engineering 2021 (CSE21; Session: Uncertainty Quantification for Data-Intensive Inverse Problems and Learning); virtual conference.
- A. Mang: Fast algorithms for nonlinear optimal control of geodesic flows of diffeomorphisms. Invited talk (host: Tan Bui-Thanh) at Oden Seminar at the Oden Institute for Computational Engineering and Sciences, University of Texas at Austin, TX, US, 2021; virtual seminar.
- N. Himthani: Multi-Node Multi-GPU Diffeomorphic Image Registration for Large-Scale Imaging Problems. Talk at Supercomputing 2020 (SC20); (virtual conference).
- A. Mang: Statistical analysis of shapes and shape deformations in 3D. Contributed talk at Joint Mathematics Meetings (JMM20; Session: AMS Special Session on Geometry in the Mathematics of Data Science); virtual conference.
- A. Mang: Classification of 3D shapes and shape deformations. Contributed talk at Annual Meeting of the SIAM Texas-Louisiana Section 2020 (SIAM TX-LA 20; Session: Scientific Machine Learning); virtual conference.
- A. Mang: Fast GPU-accelerated diffeomorphic image registration in 3D. Contributed talk at SIAM Imaging Science Conference 2020 (IS20; Session: Fast Algorithms for Inverse Problems and their Applications); virtual conference.
- A. Mang: Automatic classification of 3D shapes and shape deformations. Contributed talk at SIAM Conference on Mathematics of Data Science 2020 (MDS20; Session: Integration of Model-Based and Data-Based Methods with Medical Imaging); virtual conference.
- A. Mang: Estimating oncogenic parameters via biophysical brain tumor growth modeling. Invited talk at Annual Meeting of the Society for Neuro-Oncology 2019 (Session: Computational Neuro-Oncology), Phoenix, AZ, US.
- A. Mang: Fast GPU-accelerated diffeomorphic image registration in 3D. Contributed talk at SIAM TX-LA Sectional Meeting 2019 (Session: Recent Advances in Inverse Problems and Imaging), Southern Methodist University Dallas, TX.
- S. Subramanian: MRI-driven inverse problems for brain tumor growth models in personalized medicine. Contributed talk at SIAM TX-LA Sectional Meeting 2019 (Session: Recent Advances in Inverse Problems and Imaging), Southern Methodist University Dallas, TX.
- A. Mang: Fast algorithms for nonlinear optimal control for diffeomorphic registration. Invited talk (host: Roland Herzog) at RICAM’s Special Semester on Optimization (organized by E. Sachs, K. Kunisch; New Trends in PDE-Constrained Optimization), Johann Radon Institute for Computational and Applied Mathematics (RICAM), Linz, AT, 2019.
- A. Mang: Uncertainty quantification in non-linear optimal control problems for diffeomorphic registration. Contributed talk at AMS Sectional Meeting (Session: Uncertainty Quantification Strategies for Physics Applications), University of Wisconsin-Madison, Madison, WI, US, 2019.
- A. Mang: Fast algorithms for non-linear optimal control problems for diffeomorphic registration. Invited talk (host: Christoph Brune) at Department of Applied Mathematics (DAMUT colloquium), University of Twente, Enschede, NL, 2019.
- J. Herring: Fast ADMM-type algorithms for diffeomorphic shape matching. Contributed talk at International Congress on Industrial and Applied Mathematics 2019 (ICIAM; Session: Fast iterative methods for large-scale inverse problems in imaging), Valencia, ES.
- A. Mang: Fast diffeomorphic image registration in 3D. Contributed talk at International Congress on Industrial and Applied Mathematics 2019 (ICIAM; Session: Fast iterative methods for large-scale inverse problems in imaging), Valencia, ES.
- J. Herring: Fast algorithms for optimal control based diffeomorphic shape matching. Contributed talk at Applied Inverse Problems (AIP) Conference 2019 (Session: Numerical methods for optimal control problems in imaging), Grenoble, FR.
- A. Mang: Fast algorithms for initial value control problems in image registration. Contributed talk at Applied Inverse Problems (AIP) Conference 2019 (Session: Analysis and Fast Numerical Methods for Inverse Problems and their Applications), Grenoble, FR.
- A. Mang: Diffeomorphic shape matching: Fast algorithms for non-linear optimal control problems. Invited talk (host: Mathilde Mougeot) at Éléments de mathématique pour l’intelligence artificielle, École Normale Supérieure, Paris-Saclay, Cachan, FR, 2019.
- A. Mang: Optimal control of PDEs: Application to brain tumor modeling. Contributed talk at AMS Sectional Meeting 2018 (Session: Validation and Verification Strategies in Multiphysics Problems), University of Arkansas, Fayetteville, AR, US.
- A. Mang: Fast solvers for inverse transport problems. Contributed talk at SIAM Annual Meeting 2018 (Session: Inverse Problems), Portland, OR, US.
- A. Mang: CLAIRE: A parallel solver for constrained diffeomorphic image registration. Invited talk (host: Johannes Kast) at Mint Medical GmbH, Heidelberg, DE, 2018.
- A. Gholami: A framework for scalable biophysics-based image analysis, Supercomputing 17, Denver, CO, US.
- K. Scheufele: Coupling brain-tumor biophysical models and diffeomorphic image registration. Contributed talk at SIAM Conference on Imaging Sciences 2018 (IS18; Session: Imaging, Modeling, Visualization and Biomedical Computing), Bologna, IT.
- K. Scheufele: Block-Newton iterative solvers for joint inverse tumor growth and image registration. Contributed talk at Copper Mountain Conference on Iterative Methods 2018 (Session: Imaging), Copper Mountain, CO, US.
- A. Mang: Parallel algorithms for hyperbolic PDE-constrained optimization problems. Contributed talk at International Workshop on Parallel Matrix Algorithms and Applications 2018 (PMAA18; Session: Krylov and regularization methods for large scale inverse problems), ETH Zuerich, Zuerich, CH.
- A. Mang: CLAIRE: A parallel solver for constrained large deformation diffeomorphic image registration. Invited talk (host: Miriam Mehl) at Department of Computer Science at University of Stuttgart, Stuttgart DE, 2018.
- A. Mang: CLAIRE: A parallel solver for constrained large deformation diffeomorphic image registration. Contributed talk at SIAM Conference on Imaging Sciences 2018 (Session: Diffeomorphic image registration: Numerics, Applications, and Theory), Bologna, IT.
- A. Mang: CLAIRE: A distributed-memory solver for constrained diffeomorphic image registration. Invited talk (host: Jesse Chan) at Computational and Applied Mathematics Department, Rice University, Houston, TX, US, 2018.
- A. Mang: Computational mathematics meets medicine: Formulations, numerics, and parallel computing. Invited talk (host: James Nagy) Emory University, Numerical Analysis and Scientific Computing Seminar, Department of Mathematics & Computer Science, Atlanta, GA, US, 2018.
- A. Mang: Preconditioners for the reduced space Hessian in hyperbolic optimal control problems. Contributed talk at International Conference on Preconditioning Techniques for Scientific and Industrial Applications 2017 (Session: Preconditioning methods in large-scale ill-posed inverse problems), Vancouver, BC, CA.
- A. Mang: A distributed-memory Newton–Krylov solver for inverse transport problems. Contributed talk at US National Congress on Computational Mechanics 2017 (USNCCM17; Session: Advances in Computational Methods for Inverse Problems), Montreal, QC, CA.
- A. Mang: A distributed-memory Newton–Krylov solver for constrained diffeomorphic image registration. Contributed talk at Applied Inverse Problems (AIP17) Conference 2017, Hangzhou, CN.
- A. Mang: Parallel algorithms for optimal control based diffeomorphic image registration. Contributed talk at Houston Imaging Sciences Symposium 2017, Houston, TX, US.
- A. Mang: Parallel algorithms for PDE-constrained optimization problems with hyperbolic constraints. Contributed talk at SIAM Conference on Computational Science and Engineering 2017 (Session: Fast Solvers for Large-Scale Inverse Problems in Imaging), Atlanta, GA, US.
Poster Presentations
- P. Amiri: Bayesian inference on SPD manifolds: Geometry-aware learning of posterior covariances, SIAM TX LA Sectional Meeting 2025, University of Texas at Austin, Austin, TX.
- I. Asikul: Efficient numerical methods for multispecies tumor growth simulations, SIAM TX LA Sectional Meeting 2025, University of Texas at Austin, Austin, TX.
- M. Konduri: *Model-constrained deep learning for parameter estimation in semi-linear parabolic PDEs∗, SIAM TX LA Sectional Meeting 2025, University of Texas at Austin, Austin, TX.
- I. Asikul: Efficient numerical methods for multispecies tumor growth simulations, ChAMELEON Summer School 2025, University of Houston, Houston, TX, US.
- A. Nair: Exploration of the workings of neural networks, Undergraduate Research Day, University of Houston 2025, Houston, TX, US.
- G. Villalobos: Neural networks for inference in optimal control governed by the FitzHugh–Nagumo model, SIAM Conference on Mathematics of Data Science 2024, Atlanta, GA, US.
- J. Shi: Efficient clustering on Riemannian manifolds using Fréchet embeddings, SIAM Conference on Mathematics of Data Science 2024, Atlanta, GA, US.
- M. Konduri: DNNs for Parameter Identification in Semi-Linear Parabolic PDEs, SIAM TX LA Sectional Meeting 2024, University of Baylor, Waco, TX, US.
- B. Gutierrez: Stochastic Newton–MCMC for Bayesian inference, Undergraduate Research Day 2023, University of Houston, Houston, TX, US.
- B. Gutierrez: Stochastic Newton–MCMC for Bayesian inference, National Diversity in STEM Conference 2023, Phoenix, AZ, US.
- J. Kim: Fast evaluation of PDE operators for optimization and uncertainty quantification in problems governed by transport equations, SIAM TX LA Sectional Meeting 2022, University of Houston, Houston, TX.
- G. Villalobos: Inference for the Fitzhugh-Nagumo Model through ANNs, SIAM TX LA Sectional Meeting 2022, University of Houston, Houston, TX.
- R. Sultamuratov: Automatic classification of deformable shapes, SIAM TX LA Sectional Meeting 2022, University of Houston, Houston, TX.
- Y. Syed: Fast evaluation of kernel distances, Undergraduate Research Day 2020, University of Houston.
- A. H. A. Syed: Optimization and optimal control in machine learning, Undergraduate Research Day 2020, University of Houston.
- H. Rosso: Regularization schemes for linear inverse problems, Undergraduate Research Day 2020, University of Houston.
- M. Brunn: Fast 3D diffeomorphic image registration on GPUs. Research Poster at ACM/IEEE Conference on Supercomputing 2019, Colorado, Denver, CO, US.
- F. Huber: Efficient algorithms for geodesic shooting in diffeomorphic image registration. International Congress on Industrial and Applied Mathematics 2019, Valencia, ES.
- N. Himthani: GPU-accelerated interpolation for 3D image registration. Research Poster at ACM/IEEE Conference on Supercomputing 2018, Dallas, TX, US.
- B. Gonzalez: Fast and stable algorithms for deep learning. Undergraduate Research Day 2018, University of Houston, Houston, TX, US.