RiWalk: Fast Structural Node Embedding via Role Identification

Abstract

Nodes performing different functions in a network have different roles, and these roles can be gleaned from the structure of the network. Learning latent representations for the roles of nodes helps to understand the network and to transfer knowledge across networks. However, most existing structural embedding approaches suffer from high computation and space cost or rely on heuristic feature engineering.Here we propose RiWalk, a flexible paradigm for learning structural node representations. It decouples the structural embedding problem into a role identification procedure and a network embedding procedure. Through role identification, rooted kernels with structural dependencies kept are built to better integrate network embedding methods. To demonstrate the effectiveness of RiWalk, we develop two different role identification methods named RiWalk-SP and RiWalk-WL respectively and employ random walk based network embedding methods.Experiments on within-network classification tasks show that our proposed algorithms achieve comparable performance with other baselines while being an order of magnitude more efficient. Besides, we also conduct across-network role classification tasks. The results show potential of structural embeddings in transfer learning. RiWalk is also scalable, making it capable of capturing structural roles in massive networks.

Role-Aware Information Spread in Online Social Networks

“…The definition of roles in dMMSB implies that users with similar roles share common features and relation patterns, even without direct relationships [92]. More recently, network embedding methods [29,121,[125][126][127] have gained popularity in studying graph structure, achieving state-of-the-art performance in downstream tasks such as node classification. These methods represent network data as vectors in a latent space to preserve the topological structure and properties of the original network.…”

“…SPINE [137] incorporates structural features of both local and global node proximity to learn embeddings. RiWalk [127] assumes that nodes with different functionalities have different roles in a network, and the structure of the network can be learned using random-walk-based node embedding. These methods reconstruct the edges between nodes based on the structural similarities so that the context nodes obtained by random walks are structurally similar to the central nodes.…”

DP-GCN: Node Classification Based on Both Connectivity and Topology Structure Convolutions for Risky Seller Detection

“…In our implementation, we evaluate a few ways to generate topology structure features, e.g., Struc2vec [28], Graphwave [6], Role2vec [1] and Riwalk [20]. Based on the features, we discover the initial latent topology roles with 𝐾-means or regression classifier…”

“…It then computes random walk transition probabilities based on topology structure similarities, and embeds nodes using feature-based random walk. Recently, Riwalk [20] is proposed to model node topology structure by its degree and relative position. It can better integrate graph kernels with node embedding methods to leverage recent advancements in node embedding.…”

“…Statistics of four open and two synthetic networks Riwalk[20] models node topology structure by degree and relative position. Node2vec[26] is an extension of the skip-gram embedding model by capturing breadth-first and depth-first search based neighborhood relationship.…”