Newton-based optimization for Kullback-Leibler nonnegative tensor factorizations
From MaRDI portal
Abstract: Tensor factorizations with nonnegative constraints have found application in analyzing data from cyber traffic, social networks, and other areas. We consider application data best described as being generated by a Poisson process (e.g., count data), which leads to sparse tensors that can be modeled by sparse factor matrices. In this paper we investigate efficient techniques for computing an appropriate canonical polyadic tensor factorization based on the Kullback-Leibler divergence function. We propose novel subproblem solvers within the standard alternating block variable approach. Our new methods exploit structure and reformulate the optimization problem as small independent subproblems. We employ bound-constrained Newton and quasi-Newton methods. We compare our algorithms against other codes, demonstrating superior speed for high accuracy results and the ability to quickly find sparse solutions.
Recommendations
- On Tensors, Sparsity, and Nonnegative Factorizations
- Nonnegative tensor factorizations using an alternating direction method
- Nonnegative tensor factorization as an alternative Csiszar-Tusnady procedure: algorithms, convergence, probabilistic interpretations and novel probabilistic tensor latent variable analysis algorithms
- Tensor Methods for Large, Sparse Unconstrained Optimization
- Nonnegative matrix factorization with constrained second-order optimization
Cites work
- A comparison of algorithms for fitting the PARAFAC model
- A Limited Memory Algorithm for Bound Constrained Optimization
- Analysis of individual differences in multidimensional scaling via an \(n\)-way generalization of ``Eckart-Young decomposition
- Global Convergence of a Class of Trust Region Algorithms for Optimization with Simple Bounds
- Learning the parts of objects by non-negative matrix factorization
- Nonnegative Matrix Factorization Based on Alternating Nonnegativity Constrained Least Squares and Active Set Method
- Nonnegative matrix factorization with constrained second-order optimization
- Nonnegative tensor factorization as an alternative Csiszar-Tusnady procedure: algorithms, convergence, probabilistic interpretations and novel probabilistic tensor latent variable analysis algorithms
- On Tensors, Sparsity, and Nonnegative Factorizations
- On the complexity of nonnegative matrix factorization
- On the Goldstein-Levitin-Polyak gradient projection method
- Positive tensor factorization
- Projected Gradient Methods for Nonnegative Matrix Factorization
- Projected Newton Methods for Optimization Problems with Simple Constraints
- Sparse non-negative tensor factorization using columnwise coordinate descent
- Tackling box-constrained optimization via a new projected quasi-Newton approach
- Tensor Decompositions and Applications
Cited in
(15)- A unified global convergence analysis of multiplicative update rules for nonnegative matrix factorization
- Nonnegative tensor factorizations using an alternating direction method
- Best sparse rank-1 approximation to higher-order tensors via a truncated exponential induced regularizer
- Multiplicative algorithms for symmetric nonnegative tensor factorizations and its applications
- Variational auto-encoder based Bayesian Poisson tensor factorization for sparse and imbalanced count data
- Nonnegative tensor factorization as an alternative Csiszar-Tusnady procedure: algorithms, convergence, probabilistic interpretations and novel probabilistic tensor latent variable analysis algorithms
- Stochastic gradients for large-scale tensor decomposition
- Generalized canonical polyadic tensor decomposition
- Computing the gradient in optimization algorithms for the CP decomposition in constant memory through tensor blocking
- Rank decomposition and symmetric rank decomposition over arbitrary fields
- WINTENDED: WINdowed TENsor decomposition for densification event detection in time-evolving networks
- Sketch-based multiplicative updating algorithms for symmetric nonnegative tensor factorizations with applications to face image clustering
- Taming numerical imprecision by adapting the KL divergence to negative probabilities
- A composite optimization algorithm for Poisson tensor completions without nonnegative constraints based on the generalized Gauss-Newton method
- Tensor decompositions for count data that leverage stochastic and deterministic optimization
This page was built for publication: Newton-based optimization for Kullback-Leibler nonnegative tensor factorizations
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3458827)
