Tensor decompositions for learning latent variable models
From MaRDI portal
Abstract: This work considers a computationally and statistically efficient parameter estimation method for a wide class of latent variable models---including Gaussian mixture models, hidden Markov models, and latent Dirichlet allocation---which exploits a certain tensor structure in their low-order observable moments (typically, of second- and third-order). Specifically, parameter estimation is reduced to the problem of extracting a certain (orthogonal) decomposition of a symmetric tensor derived from the moments; this decomposition can be viewed as a natural generalization of the singular value decomposition for matrices. Although tensor decompositions are generally intractable to compute, the decomposition of these specially structured tensors can be efficiently obtained by a variety of approaches, including power iterations and maximization approaches (similar to the case of matrices). A detailed analysis of a robust tensor power method is provided, establishing an analogue of Wedin's perturbation theorem for the singular vectors of matrices. This implies a robust and computationally tractable estimation approach for several popular latent variable models.
Recommendations
Cited in
(only showing first 100 items - show all)- Convergence rate analysis for the higher order power method in best rank one approximations of tensors
- A new method of moments for latent variable models
- A note on the simultaneous Waring rank of monomials
- Nonparametric estimation of non-exchangeable latent-variable models
- On the number of Waring decompositions for a generic polynomial vector
- Orthogonal and unitary tensor decomposition from an algebraic perspective
- Using negative curvature in solving nonlinear programs
- Convergence rates of latent topic models under relaxed identifiability conditions
- Reducing uncertainty of dynamic heterogeneous information networks: a fusing reconstructing approach
- Efficient semiparametric estimation and model selection for multidimensional mixtures
- Spectral methods for learning discrete latent tree models
- On a geometric method for the identifiability of forms
- On the subdifferential of symmetric convex functions of the spectrum for symmetric and orthogonally decomposable tensors
- On polynomial time methods for exact low-rank tensor completion
- Least squares moment identification of binary regression mixture models
- Robust high-dimensional factor models with applications to statistical machine learning
- Structuring data with block term decomposition: decomposition of joint tensors and variational block term decomposition as a parametrized mixture distribution model
- Moment identifiability of homoscedastic Gaussian mixtures
- Sparse random tensors: concentration, regularization and applications
- Approximate minimization of weighted tree automata
- Riemannian Newton optimization methods for the symmetric tensor approximation problem
- Asymptotic behavior of large Gaussian correlated Wishart matrices
- Smoothed tensor quantile regression estimation for longitudinal data
- An optimal statistical and computational framework for generalized tensor estimation
- Low tubal rank tensor recovery using the Bürer-Monteiro factorisation approach. Application to optical coherence tomography
- Inference for low-rank tensors -- no need to debias
- Tensor decomposition for learning Gaussian mixtures from moments
- Learning diagonal Gaussian mixture models and incomplete tensor decompositions
- Subgroup-effects models for the analysis of personal treatment effects
- Smoothed analysis for tensor methods in unsupervised learning
- The set of orthogonal tensor trains
- On the identifiability of ternary forms
- Statistical limits of spiked tensor models
- Some inequalities on the spectral radius of nonnegative tensors
- Optimal estimation of Gaussian mixtures via denoised method of moments
- Anomaly detection in scientific data using joint statistical moments
- Bounds on strongly orthogonal ranks of tensors
- An inexact augmented Lagrangian method for computing strongly orthogonal decompositions of tensors
- Three-way clustering of multi-tissue multi-individual gene expression data using semi-nonnegative tensor decomposition
- Tensor train spectral method for learning of hidden Markov models (HMM)
- An operator theoretic approach to nonparametric mixture models
- The Hitchhiker guide to: secant varieties and tensor decomposition
- A spectral algorithm for latent Dirichlet allocation
- Condition numbers for the tensor rank decomposition
- Successive partial-symmetric rank-one algorithms for almost unitarily decomposable conjugate partial-symmetric tensors
- Identifiability for a class of symmetric tensors
- Statistically optimal and computationally efficient low rank tensor completion from noisy entries
- Statistical thresholds for tensor PCA
- Three decompositions of symmetric tensors have similar condition numbers
- Detecting approximate replicate components of a high-dimensional random vector with latent structure
- Mixed membership Gaussians
- The average condition number of most tensor rank decomposition problems is infinite
- Lower bounds on the rank and symmetric rank of real tensors
- When are overcomplete topic models identifiable? Uniqueness of tensor Tucker decompositions with structured sparsity
- Online tensor methods for learning latent variable models
- Estimating multivariate latent-structure models
- Tensor Decompositions for Learning Latent Variable Models (A Survey for ALT)
- On generic identifiability of symmetric tensors of subgeneric rank
- Eigenconfigurations of tensors
- Fast moment estimation for generalized latent Dirichlet models
- Newton correction methods for computing real eigenpairs of symmetric tensors
- Resonator Networks, 2: Factorization Performance and Capacity Compared to Optimization-Based Methods
- Successive Rank-One Approximations for Nearly Orthogonally Decomposable Symmetric Tensors
- Training Gaussian mixture models at scale via coresets
- The search problem in mixture models
- Greedy approaches to symmetric orthogonal tensor decomposition
- Low rank symmetric tensor approximations
- Singular vectors of orthogonally decomposable tensors
- The condition number of join decompositions
- Learning Tensors From Partial Binary Measurements
- Exploiting Efficient Representations in Large-Scale Tensor Decompositions
- STORE: sparse tensor response regression and neuroimaging analysis
- Spectral learning of weighted automata. A forward-backward perspective
- scientific article; zbMATH DE number 7255154 (Why is no real title available?)
- A Model-Based Embedding Technique for Segmenting Customers
- Recovering structured probability matrices
- Predictive learning on hidden tree-structured Ising models
- Sparse tensor additive regression
- Statistical query lower bounds for tensor PCA
- Comparison of Accuracy and Scalability of Gauss--Newton and Alternating Least Squares for CANDECOMC/PARAFAC Decomposition
- Tensor-structured sketching for constrained least squares
- ISLET: fast and optimal low-rank tensor regression via importance sketching
- Minimality and uniqueness for decompositions of specific ternary forms
- Ergodicity Coefficients for Higher-Order Stochastic Processes
- Overcomplete order-3 tensor decomposition, blind deconvolution, and Gaussian mixture models
- Nonconvex Low-Rank Tensor Completion from Noisy Data
- Low-rank, Orthogonally Decomposable Tensor Regression With Application to Visual Stimulus Decoding of fMRI Data
- Robustness and Tractability for Non-convex M-estimators
- Learning Markov models via low-rank optimization
- Approximate matrix and tensor diagonalization by unitary transformations: convergence of Jacobi-type algorithms
- Operator norm inequalities between tensor unfoldings on the partition lattice
- Multiresolution low-rank tensor formats
- Estimating Higher-Order Moments Using Symmetric Tensor Decomposition
- Learning from binary multiway data: probabilistic tensor decomposition and its statistical optimality
- scientific article; zbMATH DE number 7306902 (Why is no real title available?)
- Sequence classification using third-order moments
- A sharp blockwise tensor perturbation bound for orthogonal iteration
- Computing tensor Z-eigenvectors with dynamical systems
- Optimal sparse singular value decomposition for high-dimensional high-order data
- Non-convex matrix completion and related problems via strong duality
This page was built for publication: Tensor decompositions for learning latent variable models
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2934101)
