Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

$\text{H}^2\text{TNE}$: Temporal Heterogeneous Information Network Embedding in Hyperbolic Spaces

Qijie Bai, Jiawen Guo, Haiwei Zhang, Changli Nie, Lin Zhang, Xiaojie Yuan

TL;DR

This work addresses the limitations of Euclidean embeddings for temporal heterogeneous information networks by introducing H^2TNE, a hyperbolic temporal HIN embedding framework. It combines a temporally and heterogeneously double-constrained random walk to capture evolving topology and semantics, with embeddings learned in the Poincaré ball using hyperbolic proximity and negative sampling. The method is augmented with Riemannian-gradient optimization and incremental update strategies, achieving superior performance on temporal link prediction and node classification, especially for sparse, hierarchical networks. Overall, it demonstrates that modeling temporal HINs in hyperbolic space yields both representational and practical benefits for dynamic, heterogeneous graphs.

Abstract

Temporal heterogeneous information network (temporal HIN) embedding, aiming to represent various types of nodes of different timestamps into low dimensional spaces while preserving structural and semantic information, is of vital importance in diverse real-life tasks. Researchers have made great efforts on temporal HIN embedding in Euclidean spaces and got some considerable achievements. However, there is always a fundamental conflict that many real-world networks show hierarchical property and power-law distribution, and are not isometric of Euclidean spaces. Recently, representation learning in hyperbolic spaces has been proved to be valid for data with hierarchical and power-law structure. Inspired by this character, we propose a hyperbolic heterogeneous temporal network embedding ($\text{H}^2\text{TNE}$) model for temporal HINs. Specifically, we leverage a temporally and heterogeneously double-constrained random walk strategy to capture the structural and semantic information, and then calculate the embedding by exploiting hyperbolic distance in proximity measurement. Experimental results show that our method has superior performance on temporal link prediction and node classification compared with SOTA models.

$\text{H}^2\text{TNE}$: Temporal Heterogeneous Information Network Embedding in Hyperbolic Spaces

TL;DR

This work addresses the limitations of Euclidean embeddings for temporal heterogeneous information networks by introducing H^2TNE, a hyperbolic temporal HIN embedding framework. It combines a temporally and heterogeneously double-constrained random walk to capture evolving topology and semantics, with embeddings learned in the Poincaré ball using hyperbolic proximity and negative sampling. The method is augmented with Riemannian-gradient optimization and incremental update strategies, achieving superior performance on temporal link prediction and node classification, especially for sparse, hierarchical networks. Overall, it demonstrates that modeling temporal HINs in hyperbolic space yields both representational and practical benefits for dynamic, heterogeneous graphs.

Abstract

Temporal heterogeneous information network (temporal HIN) embedding, aiming to represent various types of nodes of different timestamps into low dimensional spaces while preserving structural and semantic information, is of vital importance in diverse real-life tasks. Researchers have made great efforts on temporal HIN embedding in Euclidean spaces and got some considerable achievements. However, there is always a fundamental conflict that many real-world networks show hierarchical property and power-law distribution, and are not isometric of Euclidean spaces. Recently, representation learning in hyperbolic spaces has been proved to be valid for data with hierarchical and power-law structure. Inspired by this character, we propose a hyperbolic heterogeneous temporal network embedding () model for temporal HINs. Specifically, we leverage a temporally and heterogeneously double-constrained random walk strategy to capture the structural and semantic information, and then calculate the embedding by exploiting hyperbolic distance in proximity measurement. Experimental results show that our method has superior performance on temporal link prediction and node classification compared with SOTA models.
Paper Structure (4 sections, 1 figure)

This paper contains 4 sections, 1 figure.

Table of Contents

  1. Introduction
  2. Related Work
  3. Preliminaries
  4. Temporal HIN Embedding

Figures (1)

  • Figure 1: The degree distributions of two real-world networks Tokyo and DBLP. The coordinate axes are logarithmic.

Theorems & Definitions (3)

  • definition thmcounterdefinition: HINs
  • definition thmcounterdefinition: Temporal Networks
  • definition thmcounterdefinition: Temporal HINs