Influence Maximization in Temporal Social Networks with a Cold-Start Problem: A Supervised Approach

2025-04-15

AI Summary

This work addresses the influence maximization problem under cold-start conditions in temporal social networks. We propose a seed node identification and prediction framework centered on influential propagation paths (IPPs). Our method introduces a motif-based automatic IPP annotation scheme, constructs a tensorized Temporal Graph Network (TGN) tailored for multi-relational temporal graphs, and incorporates a historical-IPP-driven neighborhood enhancement mechanism to mitigate cold-start bias. Experimental results demonstrate significant improvements in seed prediction accuracy and training efficiency. Online A/B testing confirms that the proposed approach drives a 12.7% increase in network growth and improves influence coverage for cold-start users by 34.5%. The key contributions include: (i) a novel motif-guided IPP labeling strategy; (ii) a tensorized TGN architecture supporting heterogeneous temporal interactions; and (iii) a neighborhood augmentation paradigm explicitly leveraging historical IPPs to alleviate data sparsity in cold-start scenarios.

Technology Category

• Planning, Routing, and Scheduling: Temporal Planning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Machine Learning: Semi-Supervised Learning

Application Category

• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Content-based information diffusion

Abstract

Influence Maximization (IM) in temporal graphs focuses on identifying influential"seeds"that are pivotal for maximizing network expansion. We advocate defining these seeds through Influence Propagation Paths (IPPs), which is essential for scaling up the network. Our focus lies in efficiently labeling IPPs and accurately predicting these seeds, while addressing the often-overlooked cold-start issue prevalent in temporal networks. Our strategy introduces a motif-based labeling method and a tensorized Temporal Graph Network (TGN) tailored for multi-relational temporal graphs, bolstering prediction accuracy and computational efficiency. Moreover, we augment cold-start nodes with new neighbors from historical data sharing similar IPPs. The recommendation system within an online team-based gaming environment presents subtle impact on the social network, forming multi-relational (i.e., weak and strong) temporal graphs for our empirical IM study. We conduct offline experiments to assess prediction accuracy and model training efficiency, complemented by online A/B testing to validate practical network growth and the effectiveness in addressing the cold-start issue.

Problem

• Identify influential seeds in temporal networks for maximizing expansion
• Address cold-start issue in temporal social networks effectively
• Predict and label Influence Propagation Paths accurately and efficiently

Innovation

• Motif-based labeling for efficient IPP identification
• Tensorized TGN for multi-relational temporal graphs
• Cold-start mitigation via historical neighbor augmentation