Efficient Nonnegative Tucker Decompositions: Algorithms and Uniqueness

Abstract

Nonnegative Tucker decomposition (NTD) is a powerful tool for the extraction of nonnegative parts-based and physically meaningful latent components from high-dimensional tensor data while preserving the natural multilinear structure of data. However, as the data tensor often has multiple modes and is large-scale, existing NTD algorithms suffer from a very high computational complexity in terms of both storage and computation time, which has been one major obstacle for practical applications of NTD. To overcome these disadvantages, we show how low (multilinear) rank approximation (LRA) of tensors is able to significantly simplify the computation of the gradients of the cost function, upon which a family of efficient first-order NTD algorithms are developed. Besides dramatically reducing the storage complexity and running time, the new algorithms are quite flexible and robust to noise because any well-established LRA approaches can be applied. We also show how nonnegativity incorporating sparsity substantially improves the uniqueness property and partially alleviates the curse of dimensionality of the Tucker decompositions. Simulation results on synthetic and real-world data justify the validity and high efficiency of the proposed NTD algorithms.

appears in IEEE Transactions on Image Processing, 2015

References in corpus (5)

• Robust Low-rank Tensor Recovery: Models and Algorithms
• Linked Component Analysis from Matrices to High Order Tensors: Applications to Biomedical Data
• Factoring nonnegative matrices with linear programs
• Bayesian Sparse Tucker Models for Dimension Reduction and Tensor Completion
• Decomposition of Big Tensors With Low Multilinear Rank

Cited by in corpus (12)

• Tensor Ring Decomposition
• Linked Component Analysis from Matrices to High Order Tensors: Applications to Biomedical Data
• Tensor Networks for Big Data Analytics and Large-Scale Optimization Problems
• Unsupervised EHR-based Phenotyping via Matrix and Tensor Decompositions
• Sketching for low-rank nonnegative matrix approximation: Numerical study
• Randomized algorithms for the low multilinear rank approximations of tensors
• Uncovering audio patterns in music with Nonnegative Tucker Decomposition for structural segmentation
• Learning the Hierarchical Parts of Objects by Deep Non-Smooth Nonnegative Matrix Factorization
• D-Tracker: Modeling Interest Diffusion in Social Activity Tensor Data Streams
• Fast Hypergraph Regularized Nonnegative Tensor Ring Factorization Based on Low-Rank Approximation
• Jointly Learning Structured Analysis Discriminative Dictionary and Analysis Multiclass Classifier
• Best Nonnegative Rank-One Approximations of Tensors