A0654
Title: Representation learning of dynamic networks
Authors: Haixu Wang - University of Calgary (Canada) [presenting]
Abstract: A novel representation learning model is introduced for dynamic networks, capturing evolving relationships within a population. Framing the problem within functional data analysis, dynamic networks are represented as matrix-valued functions and embedded into a lower-dimensional functional space. This space preserves network topology while enabling attribute learning, community detection, and link prediction. The model accommodates asymmetric embeddings to distinguish nodes' regulatory and receiving roles, ensuring continuity over time. Unlike discrete-time methods, this approach leverages a functional representation to infer network structures at unobserved time points. The model is validated through simulations and real-world applications, demonstrating superior link prediction accuracy compared to existing approaches. A statistical framework is provided, balancing representation learning capacity with interpretability, offering insights into dynamic network structures.
