Abstract:Graphons are limit objects of sequences of graphs and are used to analyze the behavior of large graphs. Recently, graphon signal processing has been developed to study signal processing on large graphs. A major limitation of this approach is that any sparse sequence of graphs inevitably converges to the zero graphon, rendering the resulting signal processing theory trivial and inadequate for sparse graph sequences. To overcome this limitation, we propose a new signal processing framework that leverages the concept of generalized graphons and introduces the stretched cut distance as a measure to compare these graphons. Our framework focuses on the sampling of graph sequences from generalized graphons and explores the convergence properties of associated operators, spectra, and signals. Our signal processing framework provides a comprehensive approach to analyzing and processing signals on graph sequences, even if they are sparse. Finally, we discuss the practical implications of our theory for real-world large networks through numerical experiments.

What problem does this paper attempt to address?

### Problems Addressed by the Paper This paper aims to address the limitations of traditional graph theory methods (such as graph signal processing) when dealing with sparse graph sequences. Specifically, traditional graph theory methods converge to zero graph objects when processing sparse graph sequences, making signal processing theory trivial and insufficient. To overcome this limitation, the authors propose a new signal processing framework that utilizes the concept of generalized graphons and introduces the stretched cut distance as a metric for comparing these graph objects. ### Main Contributions 1. **Proposed Generalized Graphons and Stretched Cut Distance**: The authors define generalized graphons and introduce the stretched cut distance to handle sparse graph sequences. 2. **Signal Processing Framework**: Based on generalized graphons and stretched cut distance, a general signal processing framework is proposed. 3. **Sampling Scheme**: An easily implementable sampling scheme for sparse graph sequences is proposed, and the convergence of the sampled graphs is proven. 4. **Filter Construction and Convergence**: Methods for constructing different filters are described, and their convergence is discussed. 5. **Experimental Validation**: Theoretical predictions are validated through numerical experiments, particularly on two commonly used large graph datasets, where the proposed model outperforms classical graph theory models. ### Background and Motivation - **Graph Signal Processing (GSP)**: Tools for sampling, reconstructing, and filtering signals on graphs. - **Graph Neural Networks (GNNs)**: Provide deep learning architectures for graph-related tasks by introducing non-linearity. - **Graphons**: Limit objects of graph sequences used to analyze the behavior of large graphs. Classical graphons are suitable for dense graph sequences but have limitations when dealing with sparse graph sequences. ### Key Concepts - **Generalized Graphons**: Non-negative, bounded, symmetric functions defined on the non-compact domain \( \mathbb{R}_+^2 \). - **Stretched Cut Distance**: A metric for comparing generalized graphons. - **Stretched Graphons**: Transform sparse graphons into dense graphons by stretching. ### Experiments and Applications - **Generation of Sparse Graph Sequences**: Sparse graph sequences are generated through a sampling algorithm, and their convergence under the stretched cut distance is verified. - **Signal Processing**: Explores how to generate finite graphs and their associated graph signals from given generalized graphons and signals. - **Spectral Convergence**: Studies the spectral convergence properties of sampled graphs, validating theoretical predictions. ### Conclusion By introducing generalized graphons and the stretched cut distance, the paper provides a comprehensive framework for analyzing and processing signals on sparse graph sequences. This framework is applicable not only to sparse graph sequences but also to non-sparse graph sequences and performs well in practical applications.

Modeling Sparse Graph Sequences and Signals Using Generalized Graphons

Generalized Graphon Process: Convergence of Graph Frequencies in Stretched Cut Distance

Graphon Filters: Graph Signal Processing in the Limit

A Hilbert Space Theory of Generalized Graph Signal Processing

To further understand graph signals

On Random Sampling of Diffused Graph Signals with Sparse Inputs on Vertex Domain

A Graph Diffusion Lms Strategy For Adaptive Graph Signal Processing

Signal processing on large networks with group symmetries

Sampling and Uniqueness Sets in Graphon Signal Processing

On Sampling of Bandlimited Graph Signals

Graph Signal Processing for Geometric Data and Beyond: Theory and Applications

A graphon-signal analysis of graph neural networks

Graphons of Line Graphs

Unified Fourier bases for GSP on stochastic block model graphs

Spectral Convergence of Complexon Shift Operators

A Unified Approach to Scalable Spectral Sparsification of Directed Graphs

Spectral Representations of Graphons in Very Large Network Systems Control

Graphon Particle Systems, Part I: Spatio-Temporal Approximation and Law of Large Numbers

Learning Sparse Graphons and the Generalized Kesten-Stigum Threshold

Graph Laplacian Learning with Exponential Family Noise

Community-Aware Graph Signal Processing