Global ETD Search
Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.
Results
Showing 1 to 13 of 13 for “"Randomized Numerical Linear Algebra"”.
-
Randomized Numerical Linear Algebra for Large-scale Matrix Data
… is about computational tools based on randomized numerical linear algebra for handling larg-scale matrix data. Since large datasets have become commonly available in a wide variety of modern applications, there has been an increasing demand for numerical methods for storing, processing, …
-
Responsible Recklessness: Why High Performance Computing Must Embrace Randomized Numerical Linear Algebra
<p>Randomized Numerical Linear Algebra (RandNLA) is a relatively young branch of the field of Numerical Linear Algebra (NLA). It leverages randomization as the main approach to develop algorithms for computing solutions to classical linear algebra problems with performance superior to the …
-
Randomized Algorithms for Solving Singular Value Decomposition Problems with Matlab Toolbox
… thesis gives an overview of the state-of-the-art randomized linear algebra algorithms for singular value decomposition (SVD), including the presentation of existing pseudo-codes and theoretical error analysis. Our main focus is on presenting numerical experiments illustrating image restoration …
-
Large Eddy Simulation and Reduced Order Modeling to Contain the Computational Cost of Ocean Simulations
… wind-driven ocean dynamics. Its direct numerical simulation (DNS) is challenging due to the size of the typical computational domain and the need for high-resolution meshes to resolve the full spectrum of turbulent scales. This dissertation presents three approaches to overcome these …
-
Data-driven methods for compact modeling of stochastic processes
… et al. to the committor problem using randomized numerical linear algebra. Both projects outline practical data-driven methods for estimating quantities of interest in stochastic processes that are tunable with only a few hyperparameters. The success of these methods is demonstrated …
-
Speeding up nonnegative low-rank approximation: Parallelism and Randomization
… We scale down by applying techniques from Randomized Numerical Linear Algebra (RandNLA) to design efficient, randomized algorithms for NMF which allow us to down sample data in a principled way. These randomized methods achieve up to 7× speed up over nonrandomized methods and preserve …
-
Natural gradient algorithms for training deep neural networks
… of K-factors, obtainable in quadratic-time with randomized Numerical Linear Algebra (NLA). Our theory shows that at least for sufficiently large layers, we can also control the K-factors representation error, owing to the exponentially averaged (EA) construction paradigm. Numerical results show …
-
Subspace learning by randomized sketching
… we consider more general problems in numerical linear algebra where data has a natural partition into blocks. This is common in applications with distributed or decentralized data. We study the problems of sketched ridge regression and sketched matrix multiplication under this …
-
Paleo-reconstruction of Greenland ice sheet dynamics over the Holocene via Bayesian inversion
… on ice sheets. Despite recent advances in numerical and physical modeling of ice sheets, the ice sheet models differ substantially in the forecasts they make. This is because these forecasts, although they start from a representative state of the GrIS, use ad-hoc parameters and suffer from …
-
Randomized algorithms and theory for rank estimation and least squares
<p>This thesis is concerned with randomized algorithms for two matrix problems: numerical rank estimation, with extensions to low-rank approximation, and numerical solutions to least squares (LS) problems. As the dimensions of matrix problems in modern computational settings grow, classical …
-
Sub-Sampled Matrix Approximations
… are widely used to accelerate many numerical algorithms. Current methods sample row (or column) spaces to reduce their computational footprint and approximate a matrix A with an appropriate embedding of the data sampled. This work introduces a novel family of randomized iterative …
-
Randomized dimension reduction with statistical guarantees
… perspective, we design and analyze fast randomized low-rank decomposition algorithms for large matrices based on “matrix sketching”, which can be regarded as a dimension reduction strategy in the data space. These include the randomized pivoting-based interpolative and CUR decomposition …
-
Fast algorithms for minimally supervised learning
… theory, regret analysis, and scalable numerical linear algebra enables minimally supervised learning algorithms with provable statistical guarantees and high computational efficiency.