Autoregressive Moving Average Graph Filtering
arXiv:1602.04436 · doi:10.1109/TSP.2016.2614793
Abstract
One of the cornerstones of the field of signal processing on graphs are graph filters, direct analogues of classical filters, but intended for signals defined on graphs. This work brings forth new insights on the distributed graph filtering problem. We design a family of autoregressive moving average (ARMA) recursions, which (i) are able to approximate any desired graph frequency response, and (ii) give exact solutions for tasks such as graph signal denoising and interpolation. The design philosophy, which allows us to design the ARMA coefficients independently from the underlying graph, renders the ARMA graph filters suitable in static and, particularly, time-varying settings. The latter occur when the graph signal and/or graph are changing over time. We show that in case of a time-varying graph signal our approach extends naturally to a two-dimensional filter, operating concurrently in the graph and regular time domains. We also derive sufficient conditions for filter stability when the graph and signal are time-varying. The analytical and numerical results presented in this paper illustrate that ARMA graph filters are practically appealing for static and time-varying settings, as predicted by theoretical derivations.
References in corpus (4)
Cited by in corpus (39)
- Graph Neural Networks with convolutional ARMA filters
- Graphs, Convolutions, and Neural Networks: From Graph Filters to Graph Neural Networks
- Graph Signal Processing: History, Development, Impact, and Outlook
- Joint Detection and Localization of Stealth False Data Injection Attacks in Smart Grids using Graph Neural Networks
- Signal Processing on Higher-Order Networks: Livin' on the Edge ... and Beyond
- Forecasting Time Series with VARMA Recursions on Graphs
- Adaptive Graph Signal Processing: Algorithms and Optimal Sampling Strategies
- Advances in Distributed Graph Filtering
- Filtering Random Graph Processes Over Random Time-Varying Graphs
- Grid-Graph Signal Processing (Grid-GSP): A Graph Signal Processing Framework for the Power Grid
- Spectral Projector-Based Graph Fourier Transforms
- A Directed Graph Fourier Transform with Spread Frequency Components
- Flow Smoothing and Denoising: Graph Signal Processing in the Edge-Space
- Fast Graph Fourier Transforms Based on Graph Symmetry and Bipartition
- Bayesian Estimation of Graph Signals
- Graphon Filters: Graph Signal Processing in the Limit
- Stochastic Graph Neural Networks
- Distributed Training of Graph Convolutional Networks
- Controllability of Bandlimited Graph Processes Over Random Time Varying Graphs
- Beyond Low-Pass Filters: Adaptive Feature Propagation on Graphs
- Joint Time-Vertex Fractional Fourier Transform
- Widely-Linear MMSE Estimation of Complex-Valued Graph Signals
- DCT and DST Filtering with Sparse Graph Operators
- Preconditioned Gradient Descent Algorithm for Inverse Filtering on Spatially Distributed Networks
- Online Graph-Based Change Point Detection in Multiband Image Sequences
- Designing Asymmetric Shift Operators for Decentralized Subspace Projection
- DFNets: Spectral CNNs for Graphs with Feedback-Looped Filters
- Polynomial graph filter of multiple shifts and distributed implementation of inverse filtering
- Node-Adaptive Regularization for Graph Signal Reconstruction
- Modelling Graph Errors: Towards Robust Graph Signal Processing
- Blind Community Detection from Low-rank Excitations of a Graph Filter
- BiGCN: A Bi-directional Low-Pass Filtering Graph Neural Network
- Learning Optimal Graph Filters for Clustering of Attributed Graphs
- Optimized Quantization in Distributed Graph Signal Filtering
- Online Distributed Learning over Graphs with Multitask Graph-Filter Models
- Distributed Network Privacy using Error Correcting Codes
- Graph Denoising with Framelet Regularizer
- Graphs for deep learning representations
- Estimating Network Processes via Blind Identification of Multiple Graph Filters