Abstract:We consider the dictionary learning problem, where the aim is to model the given data as a linear combination of a few columns of a matrix known as a dictionary, where the sparse weights forming the linear combination are known as coefficients. Since the dictionary and coefficients, parameterizing the linear model are unknown, the corresponding optimization is inherently non-convex. This was a major challenge until recently, when provable algorithms for dictionary learning were proposed. Yet, these provide guarantees only on the recovery of the dictionary, without explicit recovery guarantees on the coefficients. Moreover, any estimation error in the dictionary adversely impacts the ability to successfully localize and estimate the coefficients. This potentially limits the utility of existing provable dictionary learning methods in applications where coefficient recovery is of interest. To this end, we develop NOODL: a simple Neurally plausible alternating Optimization-based Online Dictionary Learning algorithm, which recovers both the dictionary and coefficients exactly at a geometric rate, when initialized appropriately. Our algorithm, NOODL, is also scalable and amenable for large scale distributed implementations in neural architectures, by which we mean that it only involves simple linear and non-linear operations. Finally, we corroborate these theoretical results via experimental evaluation of the proposed algorithm with the current state-of-the-art techniques. Keywords: dictionary learning, provable dictionary learning, online dictionary learning, non-convex, sparse coding, support recovery, iterative hard thresholding, matrix factorization, neural architectures, neural networks, noodl, sparse representations, sparse signal processing.

Online Dictionary Learning for Local Coordinate Coding with Locality Coding Adaptors

Online Discriminative Dictionary Learning for Image Classification Based on Block-Coordinate Descent Method

Large-scale Dictionary Learning for Local Coordinate Coding.

Joint CRF and Locality-Consistent Dictionary Learning for Semantic Segmentation.

Local Laplacian Coding from Theoretical Analysis of Local Coding Schemes for Locally Linear Classification.

Locality-Constrained Transfer Coding for Heterogeneous Domain Adaptation

Discriminative analysis dictionary learning with adaptively ordinal locality preserving

Locality-Adaptive Structured Dictionary Learning for Cross-Domain Recognition

Adversarial Learning with Local Coordinate Coding.

Slice-Based Online Convolutional Dictionary Learning

Laplacian Sparse Coding Dictionary for Image Set Based Collaborative Representation

Locality Constrained Analysis Dictionary Learning via K-SVD Algorithm

Kernel Regularized Nonlinear Dictionary Learning for Sparse Coding

Multiscale Online Dictionary Learning for Quality Scalable Video Coding

A Locality-Constrained and Label Embedding Dictionary Learning Algorithm for Image Classification.

Topic Level Sampling Towards Optimized Locality Sensitive Vocabulary Coding

NOODL: Provable Online Dictionary Learning and Sparse Coding

Two-layers local coordinate coding

Multi-level Discriminative Dictionary Learning with Application to Large Scale Image Classification.

An augmented Lagrangian multi-scale dictionary learning algorithm

Blockwise Coordinate Descent Schemes for Efficient and Effective Dictionary Learning.