Mask Combination of Multi-Layer Graphs for Global Structure Inference

  title={Mask Combination of Multi-Layer Graphs for Global Structure Inference},
  author={Eda Bayram and Dorina Thanou and Elif Vural and Pascal Frossard},
  journal={IEEE Transactions on Signal and Information Processing over Networks},
  • Eda Bayram, Dorina Thanou, +1 author P. Frossard
  • Published 22 October 2019
  • Computer Science, Engineering, Mathematics
  • IEEE Transactions on Signal and Information Processing over Networks
Structure inference is an important task for network data processing and analysis in data science. In recent years, quite a few approaches have been developed to learn the graph structure underlying a set of observations captured in a data space. Although real-world data is often acquired in settings where relationships are influenced by a priori known rules, such domain knowledge is still not well exploited in structure inference problems. In this paper, we identify the structure of signals… Expand
Joint Estimation and Inference for Data Integration Problems based on Multiple Multi-layered Gaussian Graphical Models
This work proposes a general statistical framework based on Gaussian graphical models for horizontal and vertical integration of information in such datasets, and develops a debiasing technique and asymptotic distributions of inter-layer directed edge weights that utilize already computed neighborhood selection coefficients for nodes in the upper layer. Expand
Node Attribute Completion in Knowledge Graphs with Multi-Relational Propagation
The approach, denoted as MRAP, imputes the values of missing attributes by propagating information across the multi-relational structure of a knowledge graph by employing regression functions for predicting one node attribute from another depending on the relationship between the nodes and the type of the attributes. Expand
Unveiling Anomalous Edges and Nominal Connectivity of Attributed Networks
This work deals with uncovering anomalous edges in attributed graphs using two distinct formulations with complementary strengths, which can be easily distributed, and hence efficient, and relies on decomposing the graph data matrix into low rank plus sparse components to markedly improve performance. Expand
Non-Bayesian Estimation Framework for Signal Recovery on Graphs
  • T. Routtenberg
  • Computer Science, Engineering
  • IEEE Transactions on Signal Processing
  • 2021
This paper formulate the graph signal recovery as a non-Bayesian estimation problem under a weighted mean-squared-error (WMSE) criterion, which is based on a quadratic form of the Laplacian matrix of the graph, and its trace WMSE is the Dirichlet energy of the estimation error w.r.t. the graph. Expand


Connecting the Dots: Identifying Network Structure via Graph Signal Processing
Challenges and opportunities for SP research in emerging topic areas at the crossroads of modeling, prediction, and control of complex behavior arising in networked systems that evolve over time are introduced. Expand
Graph Laplacian Mixture Model
This work proposes a novel generative model that represents a collection of distinct data which naturally live on different graphs that demonstrates promising performance in data clustering and multiple graph inference, and shows desirable properties in terms of interpretability and coping with high dimensionality on weather and traffic data. Expand
Learning Graphs From Data: A Signal Representation Perspective
The conceptual similarities and differences between classical and GSP-based graph-inference methods are emphasized and the potential advantage of the latter is highlighted in a number of theoretical and practical scenarios. Expand
Joint inference of networks from stationary graph signals
This work investigates the problem of jointly inferring the topology of a collection of networks from the observation of signals defined on their common set of nodes, paying particular attention to setups where the different graphs differ only on a few links. Expand
Network Topology Inference from Spectral Templates
The novel idea is to find a graph shift that, while being consistent with the provided spectral information, endows the network with certain desired properties such as sparsity, and develops efficient inference algorithms stemming from provably tight convex relaxations of natural nonconvex criteria. Expand
Characterization and Inference of Graph Diffusion Processes From Observations of Stationary Signals
This paper proposes a characterization of the space of valid graphs, in the sense that they can explain stationary signals, and illustrates how this characterization can be used for graph recovery. Expand
A Recurrent Graph Neural Network for Multi-relational Data
Key aspects of the novel GRNN architecture are the use of multi-relational graphs, the dynamic adaptation to the different relations via learnable weights, and the consideration of graph-based regularizers to promote smoothness and alleviate over-parametrization. Expand
Graph Heat Mixture Model Learning
This paper introduces a generative model for mixed signals following a heat diffusion process on multiple graphs and proposes an expectation-maximisation algorithm that can successfully separate signals into corresponding groups, and infer multiple graphs that govern their behaviour. Expand
Multi-view graph learning with adaptive label propagation
MVGL integrates latent factor extraction, graph sparsification, and label propagation into a unified framework, and seeks shared latent factors from multi-view data as view-independent data representations, and then constructs a sparse graph accordingly. Expand
Combinatorial Analysis of Multiple Networks
This work introduces a real-world layered network combining different kinds of online and offline relationships, and presents an innovative methodology and related analysis tools suggesting the existence of hidden motifs traversing and correlating different representation layers. Expand