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 7 of 7 for “"spectral initialization"”.
-
A non-convex framework for structured non-stationary covariance recovery theory and application
… an optimization scheme with a carefully tailored spectral initialization, combined with iteratively re ned alternating projected gradient descent. We prove a linear convergence rate for the proposed descent scheme and establish sample complexity guarantees for the estimator. As a motivating …
-
On Initializations for NMF
… (2024), the authors proposed a sketching based initialization for unconstrained two block matrix factorization (MF) with provable guarantee. The purpose of this thesis is to examine the limits of extending these two proof frameworks in the context of Non-negative Matrix Factorization (NMF), as …
-
Nonlinear Estimation with Applications to Wireless Communications, Imaging, and Machine Learning
… signals from correlated magnitude measurements), spectral initialization (computing accurate initializers for nonconvex optimization problems), and neural networks (initializing weights in neural networks). For each application, we devise new algorithms that benefit from one or more of the …
-
Learning from Multiple and Heterogeneous Datasets
… community structures. We propose a spectral initialization + maximum a posteriori based refinement algorithm that enjoys minimax optimality. This algorithm serves as a key step in MARIO’s quality control steps. The third part is about minimax optimal estimation of a latent …
-
COMPUTATIONALLY EFFICIENT AND ROBUST METHODS FOR LARGE-SCALE OPTIMIZATION AND SCIENTIFIC COMPUTING
… which is a more suitable error proxy for spectral algorithms used in certain graph problems. We obtain improved convergence rates and propose novel stopping criteria that enable substantial computational savings. We next turn to the problem of computing eigenvectors in distributed …
-
Co-clustering algorithms : extensions and applications
… challenges such as data transformation, cluster initialization, scalability, and so on. Therefore, this thesis focuses on developing sophisticated co-clustering methodologies to maturity and its ultimate goal is to promote co-clustering as an invaluable and indispensable unsupervised analysis …
-
Provable surrogate gradient-based optimization for unsupervised learning
… with missing entries. In each case, we provide a spectral initialization subroutine that gives a coarse estimate of the true dictionary. Then, starting from this estimate, we prove that the surrogate descent algorithm linearly converges to the true dictionary. We analyze the algorithm and …