Search
📃

Discovering Latent Representations of Relations for Interacting Systems

Citation
Dohae Lee, Young Jin Oh, and In-Kwon Lee, "Discovering Latent Representations of Relations for Interacting Systems ", IEEE Access, Vol. 9, 149089-149099, https://doi.org/10.1109/ACCESS.2021.3125335, November 2021
Abstract
Systems whose entities interact with each other are common. In recent years, there has been increasing interest in discovering the relationships between entities in interacting systems using graph neural networks. However, existing approaches are difficult to apply if the number of relations is unknown or if the relations are complex. We propose the DiScovering Latent Relation (DSLR) model, which is flexibly applicable even if the number of relations is unknown or many types of relations exist. The flexibility of our DSLR model comes from an encoder that represents the relation between two entities in a latent space and a decoder that can handle all types of relations. Experiments conducted on graph data with various relationships between entities show that the proposed method is suitable for analyzing dynamic graphs with an unknown number of complex relations. (Journal IF = 3.367 (2020), Ranking = 40.37% (Q2), Category = COMPUTER SCIENCE - INFORMATION SYSTMES, SCIE)