Sparse Decomposition

Sparse decomposition techniques aim to represent data as a sum of low-rank and sparse components, effectively separating structured background information from sparse anomalies or details. Current research focuses on refining existing models like low-rank plus sparse (L+S) decomposition, incorporating additional regularizations (e.g., smoothness constraints) to improve accuracy and efficiency, and exploring novel algorithms such as deep unfolding networks and reinforcement learning for optimized decomposition. These advancements have significant implications across diverse fields, including anomaly detection in industrial imaging, improved dynamic MRI reconstruction, and enhanced audio processing, by enabling more efficient and accurate data analysis.

Papers

October 25, 2024

Sparse Decomposition of Graph Neural Networks
Yaochen Hu, Mai Zeng, Ge Zhang, Pavel Rumiantsev, Liheng Ma, Yingxue Zhang, Mark Coates
Graph Neural Network Node Representation Graph Neural Network Model Sparse Decomposition

September 9, 2024

A Novel Representation of Periodic Pattern and Its Application to Untrained Anomaly Detection
Peng Ye, Chengyu Tao, Juan Du
Application Proficiency Unsupervised Anomaly Detection Indoor Wall Frame Installation Pixel Level Anomaly Novel Representation Periodic Activation Function Sparse Decomposition

January 30, 2024

Dynamic MRI reconstruction using low-rank plus sparse decomposition with smoothness regularization
Chee-Ming Ting, Fuad Noman, Raphaël C. -W. Phan, Hernando Ombao
dMRI Reconstruction Dynamic Magnetic Resonance Dynamic MR Sparse Decomposition Smoothness Regularization

December 13, 2023

Improving Spatial Resolution of First-order Ambisonics Using Sparse MDCT Representation
Denis Likhachov, Nick Petrovsky, Elias Azarov
Spatial Resolution Order Ambisonics Sparse Decomposition Frequency Component Frequency Decomposition

October 15, 2023

Alpha Elimination: Using Deep Reinforcement Learning to Reduce Fill-In during Sparse Matrix Decomposition
Arpan Dasgupta, Pawan Kumar
Deep Reinforcement Learning Sparse Matrix Dense Environment Sparse Decomposition Elimination Voting

September 13, 2023

Efficient Learning of PDEs via Taylor Expansion and Sparse Decomposition into Value and Fourier Domains
Md Nasim, Yexiang Xue
Partial Differential Equation Efficient Learning Compressed Model PDE Model Taylor Expansion Fourier Domain Sparse Decomposition

February 21, 2023

Fusion of Global and Local Knowledge for Personalized Federated Learning
Tiansheng Huang, Li Shen, Yan Sun, Weiwei Lin, Dacheng Tao
World Event Hybrid Fusion Personalized Federated Learning Local Feature Sparse Decomposition

November 25, 2022

Learning Large Causal Structures from Inverse Covariance Matrix via Sparse Matrix Decomposition
Shuyu Dong, Kento Uemura, Akito Fujii, Shuang Chang, Yusuke Koyanagi, Koji Maruhashi, Michèle Sebag
Causal Structure Functional Magnetic Resonance Imaging Structural Equation Structural Causal Matrix Decomposition Sparse Decomposition

November 6, 2022

A Deep-Unfolded Spatiotemporal RPCA Network For L+S Decomposition
Shoaib Imran, Muhammad Tahir, Zubair Khalid, Momin Uppal
Low Rank Dense Clutter Robust Principal Component Analysis Sparse Decomposition

July 8, 2022

Signed Network Embedding with Application to Simultaneous Detection of Communities and Anomalies
Haoran Zhang, Junhui Wang
Application Proficiency Community Detection Unlabeled Anomaly Network Inference Small Community Network Model Synthetic Network Signed Network Sparse Decomposition Simultaneous Identification

July 7, 2022

Highlight Specular Reflection Separation based on Tensor Low-rank and Sparse Decomposition Using Polarimetric Cues
Moein Shakeri, Hong Zhang
Low Rank Tensor Specular Reflection Polarization Image Sparse Decomposition Polarimetric Cue Specular Highlight Removal

January 19, 2022

Compressed Smooth Sparse Decomposition
Shancong Mou, Jianjun Shi
Sparse Signal Kronecker Product Sparse Decomposition

Sparse Decomposition

Papers

Sparse Decomposition of Graph Neural Networks

A Novel Representation of Periodic Pattern and Its Application to Untrained Anomaly Detection

Dynamic MRI reconstruction using low-rank plus sparse decomposition with smoothness regularization

Improving Spatial Resolution of First-order Ambisonics Using Sparse MDCT Representation

Alpha Elimination: Using Deep Reinforcement Learning to Reduce Fill-In during Sparse Matrix Decomposition

Efficient Learning of PDEs via Taylor Expansion and Sparse Decomposition into Value and Fourier Domains

Fusion of Global and Local Knowledge for Personalized Federated Learning

Learning Large Causal Structures from Inverse Covariance Matrix via Sparse Matrix Decomposition

A Deep-Unfolded Spatiotemporal RPCA Network For L+S Decomposition

Signed Network Embedding with Application to Simultaneous Detection of Communities and Anomalies

Highlight Specular Reflection Separation based on Tensor Low-rank and Sparse Decomposition Using Polarimetric Cues

Compressed Smooth Sparse Decomposition