Graph Summarization with Controlled Utility Loss

Hajiabadi, Mahdi; Singh, Jasbir; Srinivasan, Venkatesh; Thomo, Alex

doi:10.1145/3447548.3467359

Cited by 12 publications

(2 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Instead, existing work customarily group nodes based on structural features of the network. For instance, graph summarization groups nodes with similar neighborhoods to obtain a compressed representation of network structure [2][3][4][5][6][7][8], and community detection partitions nodes into communities characterized by high interconnections among the nodes within each community [9][10][11][12][13]. Although some influence-based approaches and diffusion-aware approaches to community detection conduct a group-level influence analysis, their final goal is still to detect node communities [14][15][16][17][18].…”

Section: Introductionmentioning

confidence: 99%

Diffusion Pattern Mining

Yan,

Yang,

Yin

et al. 2024

Preprint

View full text Add to dashboard Cite

In a diffusion network, some nodes exhibit similar diffusion patterns as they have analogous influence reachabilities to the other nodes. When these nodes are selected as initially infected nodes, they tend to yield similar infection results. Mining diffusion patterns of nodes is of practical significance in various applications, such as online marketing and epidemic prevention. Nonetheless, few existing work has effectively addressed this problem. In this work, we investigate how to find out which nodes in a diffusion network share similar diffusion pattern based only on historical infection results. Towards this, we first reconstruct the structure of influence relationships in the network, and then infer the infection propagation probability on each influence relationship, based on which the influence reachability of each node can be estimated. We present a diffusion pattern similarity metric to quantify the similarity of influence reachabilities, and group nodes that share similar influence reachabilities via hierarchical clustering. Extensive experimental results on both synthetic and real-world networks verify the effectiveness and efficiency of our approach.

show abstract

Section: Introductionmentioning

confidence: 99%

Diffusion Pattern Mining

Yan,

Yang,

Yin

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…For example, Merchant et al proposed a novel algorithm called SpecSumm [3] for graph summarization via node aggregation. Additionally, Hajiabadi et al have developed a utility-driven graph summarization method G-SCIS [4] that produces optimal compression with no loss of utility. They focused primarily on summarizing static graphs, which were incapable of handling dynamic graphs formed by graph streams.…”

Section: Introductionmentioning

confidence: 99%

Dichotomy Graph Sketch: Summarizing Graph Streams with High Accuracy Based on Deep Learning

Li,

Zhang

et al. 2023

Applied Sciences

View full text Add to dashboard Cite

In many applications, data streams are indispensable to describe the relationships between nodes in networks, such as social networks, computer networks, and hyperlink networks. Fundamentally, a graph stream is a dynamic representation of a graph, which is usually composed of a sequence of edges, where each edge is represented by two endpoints and a weight. As a result of its large volume and highly dynamic nature, several graph sketches were proposed for the purposes of summarizing large-scale graph streams and enabling fast query processing. By using a compact data structure with hash functions, the graph sketches sequentially store the edges. Nevertheless, the existing graph sketches suffer from low performance on graph query tasks as a result of unpredictable collisions between heavy edges and light edges. To store heavy edges and light edges, this paper introduces a novel learning-based Dichotomy Graph Sketch (DGS) mechanism that uses two separate graph sketches, a heavy sketch and a light sketch. During a graph stream session, DGS obtains heavy edges and light edges, and uses these edges as training samples for a deep neural network (DNN) based binary classifier. The DNN-based classifier is then used to determine whether the upcoming edges are heavy or not. We will store the edges that are classified as heavy edges in the heavy sketch, and those that are classified as light edges in the light sketch. By combining the learnable classifier and Dichotomy Graph Sketches, the proposed mechanism resolves the hashing collision problem in conventional graph sketches and significantly improves graph query accuracy. The DGS algorithm outperforms the state-of-the-art graph sketches in a variety of graph query tasks based on extensive experiments that were conducted on four real-world graph stream datasets.

show abstract

Scalable Summarization for Knowledge Graphs with Controlled Utility Loss

Wang,

Wang

2023

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Graph Summarization with Controlled Utility Loss

Cited by 12 publications

References 18 publications

Diffusion Pattern Mining

Diffusion Pattern Mining

Dichotomy Graph Sketch: Summarizing Graph Streams with High Accuracy Based on Deep Learning

Scalable Summarization for Knowledge Graphs with Controlled Utility Loss

Contact Info

Product

Resources

About