HyperS2V: A Framework for Structural Representation of Nodes in Hyper Networks

Shu Liu,Cameron Lai,Fujio Toriumi

2023-11-08

Abstract:In contrast to regular (simple) networks, hyper networks possess the ability to depict more complex relationships among nodes and store extensive information. Such networks are commonly found in real-world applications, such as in social interactions. Learning embedded representations for nodes involves a process that translates network structures into more simplified spaces, thereby enabling the application of machine learning approaches designed for vector data to be extended to network data. Nevertheless, there remains a need to delve into methods for learning embedded representations that prioritize structural aspects. This research introduces HyperS2V, a node embedding approach that centers on the structural similarity within hyper networks. Initially, we establish the concept of hyper-degrees to capture the structural properties of nodes within hyper networks. Subsequently, a novel function is formulated to measure the structural similarity between different hyper-degree values. Lastly, we generate structural embeddings utilizing a multi-scale random walk framework. Moreover, a series of experiments, both intrinsic and extrinsic, are performed on both toy and real networks. The results underscore the superior performance of HyperS2V in terms of both interpretability and applicability to downstream tasks.

Social and Information Networks,Machine Learning

What problem does this paper attempt to address?

The paper aims to address the lack of structure-based node embedding methods in hyper networks. Specifically:

- **Research Background**: In simple networks, edges represent binary relationships between nodes, while in hyper networks, hyperedges can represent complex relationships among multiple nodes. Most existing node embedding methods are primarily designed for simple networks and usually focus on node proximity rather than structural similarity. For hyper networks, although there are some proximity-based embedding methods (such as HHE and HGE), there is no method specifically designed to capture node structural features.

- **Research Objective**: This paper proposes a new method called HyperS2V, which aims to generate node embeddings by capturing structural information in hyper networks. Specifically, the method defines "hyper-degree" to represent the degree information of nodes and proposes a novel distance function to measure structural differences between different hyper-degrees. Structural embeddings are generated through a multi-scale random walk framework, and a series of experiments validate its superior performance in interpretability and downstream tasks.

- **Summary of Contributions**:
1. Proposed HyperS2V, a structure-based hyper network node embedding method that captures structural similarity between nodes through hyper-degree and distance functions.
2. Demonstrated HyperS2V's excellent interpretability and adaptability on toy networks and real networks, especially in tasks such as hyperedge prediction.
3. Showcased an application case on the "Les Misérables" dataset, using structural embeddings to identify meaningful clusters among characters, thereby better understanding structural similarity in the network.