Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

FCS-HGNN: Flexible Multi-type Community Search in Heterogeneous Information Networks

Guoxin Chen, Fangda Guo, Yongqing Wang, Yanghao Liu, Peiying Yu, Huawei Shen, Xueqi Cheng

TL;DR

This work tackles the problem of flexible, query-driven multi-type community search in heterogeneous information networks (HINs). It introduces FCS-HGNN, a dual-encoder Graph Neural Network with an edge semantic attention mechanism that learns from both a heterogeneity-focused view and a query-centric view, thereby avoiding rigid meta-paths or user-defined constraints. An extended variant, LS-FCS-HGNN, adds neighbor sampling and a depth-based heuristic search to scale to large graphs, significantly boosting training and query efficiency. Experiments on five real-world HINs demonstrate consistent improvements over state-of-the-art baselines in both single-type and multi-type settings, validating the approach and its practicality for complex networked data.

Abstract

Community search is a personalized community discovery problem designed to identify densely connected subgraphs containing the query node. Recently, community search in heterogeneous information networks (HINs) has received considerable attention. Existing methods typically focus on modeling relationships in HINs through predefined meta-paths or user-specified relational constraints. However, metapath-based methods are primarily designed to identify single-type communities with nodes of the same type rather than multi-type communities involving nodes of different types. Constraint-based methods require users to have a good understanding of community patterns to define a suitable set of relational constraints, which increases the burden on users. In this paper, we propose FCS-HGNN, a novel method for flexibly identifying both single-type and multi-type communities in HINs. Specifically, FCS-HGNN extracts complementary information from different views and dynamically considers the contribution of each relation instead of treating them equally, thereby capturing more fine-grained heterogeneous information. Furthermore, to improve efficiency on large-scale graphs, we further propose LS-FCS-HGNN, which incorporates i) the neighbor sampling strategy to improve training efficiency, and ii) the depth-based heuristic search strategy to improve query efficiency. We conducted extensive experiments to demonstrate the superiority of our proposed methods over state-of-the-art methods, achieving average improvements of 14.3% and 11.1% on single-type and multi-type communities, respectively.

FCS-HGNN: Flexible Multi-type Community Search in Heterogeneous Information Networks

TL;DR

This work tackles the problem of flexible, query-driven multi-type community search in heterogeneous information networks (HINs). It introduces FCS-HGNN, a dual-encoder Graph Neural Network with an edge semantic attention mechanism that learns from both a heterogeneity-focused view and a query-centric view, thereby avoiding rigid meta-paths or user-defined constraints. An extended variant, LS-FCS-HGNN, adds neighbor sampling and a depth-based heuristic search to scale to large graphs, significantly boosting training and query efficiency. Experiments on five real-world HINs demonstrate consistent improvements over state-of-the-art baselines in both single-type and multi-type settings, validating the approach and its practicality for complex networked data.

Abstract

Community search is a personalized community discovery problem designed to identify densely connected subgraphs containing the query node. Recently, community search in heterogeneous information networks (HINs) has received considerable attention. Existing methods typically focus on modeling relationships in HINs through predefined meta-paths or user-specified relational constraints. However, metapath-based methods are primarily designed to identify single-type communities with nodes of the same type rather than multi-type communities involving nodes of different types. Constraint-based methods require users to have a good understanding of community patterns to define a suitable set of relational constraints, which increases the burden on users. In this paper, we propose FCS-HGNN, a novel method for flexibly identifying both single-type and multi-type communities in HINs. Specifically, FCS-HGNN extracts complementary information from different views and dynamically considers the contribution of each relation instead of treating them equally, thereby capturing more fine-grained heterogeneous information. Furthermore, to improve efficiency on large-scale graphs, we further propose LS-FCS-HGNN, which incorporates i) the neighbor sampling strategy to improve training efficiency, and ii) the depth-based heuristic search strategy to improve query efficiency. We conducted extensive experiments to demonstrate the superiority of our proposed methods over state-of-the-art methods, achieving average improvements of 14.3% and 11.1% on single-type and multi-type communities, respectively.
Paper Structure (36 sections, 2 theorems, 8 equations, 9 figures, 4 tables, 3 algorithms)

This paper contains 36 sections, 2 theorems, 8 equations, 9 figures, 4 tables, 3 algorithms.

Table of Contents

  1. Introduction
  2. RELATED WORK
  3. Preliminaries
  4. FCS-HGNN
  5. Heterogeneous & Query Information Preprocessing
  6. Type-shared Feature Projection
  7. Query-dependent Graph
  8. Information Integration
  9. Edge Semantic Attention Mechanism
  10. Heterogeneous Encoder
  11. Query Encoder
  12. Feature Fusion
  13. Depth-based Heuristic Search Algorithm
  14. Training and Online Query
  15. Neighbor Sampling
  16. ...and 21 more sections

Key Result

Theorem 1

Let $T$ denote the number of training iterations, $\vert V\vert$ represent the number of nodes, $\vert E\vert$ represent the number of edges, $\vert \mathcal{D}\vert$ be the quantity of the training set, and $F$ and $F'$ represent the feature dimensions before and after linear transformation, respec

Figures (9)

  • Figure 1: The blue and red circles represent single-type and multi-type community examples in the bibliographic network, respectively.
  • Figure 2: The overall framework of FCS-HGNN.
  • Figure 3: Efficiency comparison
  • Figure 4: Ablation Study of EA.
  • Figure 5: Ablation Study of Query Enc.
  • ...and 4 more figures

Theorems & Definitions (4)

  • Definition 1: Heterogeneous Information Network, HIN
  • Definition 2: Multi-type Community
  • Theorem 1
  • Theorem 2