Change Point Detection in Dynamic Graphs with Decoder-only Latent Space Model

2024-04-06

AI Summary

This paper addresses unsupervised change-point detection in dynamic graph time series. Methodologically, it proposes a generative model featuring a decoder-exclusive latent space—departing from conventional encoder-decoder architectures in favor of a pure decoder design. It introduces a data-driven empirical Bayes prior to model graph structural evolution and jointly incorporates grouped fused Lasso regularization with the Alternating Direction Method of Multipliers (ADMM) for adaptive latent-variable sensitivity to change points. Posterior inference employs maximum approximate likelihood estimation coupled with Langevin dynamics to enhance robustness. Experiments demonstrate that the method significantly outperforms existing baselines on synthetic data and successfully identifies change points highly aligned with major real-world events in applications including social networks and urban traffic systems—validating its effectiveness and practical utility.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Reasoning under Uncertainty: Graphical Models
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Web Mining and Content Analysis: Models for Web evolution
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics

Abstract

This manuscript studies the unsupervised change point detection problem in time series of graphs using a decoder-only latent space model. The proposed framework consists of learnable prior distributions for low-dimensional graph representations and of a decoder that bridges the observed graphs and latent representations. The prior distributions of the latent spaces are learned from the observed data as empirical Bayes to assist change point detection. Specifically, the model parameters are estimated via maximum approximate likelihood, with a Group Fused Lasso regularization imposed on the prior parameters. The augmented Lagrangian is solved via Alternating Direction Method of Multipliers, and Langevin Dynamics are recruited for posterior inference. Simulation studies show good performance of the latent space model in supporting change point detection and real data experiments yield change points that align with significant events.

Problem

• Detects change points in dynamic graphs unsupervised
• Uses decoder-only latent space model for detection
• Learns prior distributions to assist detection

Innovation

• Decoder-only latent space model for graphs
• Maximum likelihood with Group Fused Lasso
• ADMM and Langevin Dynamics for inference