Scalable and Effective Bipartite Network Embedding

Given a bipartite graph G consisting of inter-set weighted edges connecting the nodes in two disjoint sets U and V, bipartite network embedding (BNE) maps each node u(i) is an element of U and v(j) is an element of V to compact embedding vectors that capture the hidden topological features surrounding the nodes, to facilitate downstream tasks. Effective BNE should preserve not only the direct connections between nodes but also the multi-hop relationships formed alternately by the two types of nodes in G, which can incur prohibitive overheads, especially on massive bipartite graphs with millions of nodes and billions of edges. Existing solutions are hardly scalable to massive bipartite graphs, and often produce low-quality results. This paper proposes GEBE, a generic BNE framework achieving state-of-the-art performance on massive bipartite graphs, via four main algorithmic designs. First, we present two generic measures to capture the multi-hop similarity/proximity between homogeneous/heterogeneous nodes respectively, and the measures can be instantiated with three popular probability distributions, including Poisson, Geometric, and Uniform distributions. Second, GEBE formulates a novel and unified BNE objective to preserve the two measures of all possible node pairs. Third, GEBE includes several efficiency designs to get high-quality embeddings on massive graphs. Finally, we observe that GEBE achieves the best performance when instantiating MHS and MHP using a Poisson distribution, and thus, we further develop GEBE(p) based on Poisson-instantiated MHS and MHP, with non-trivial efficiency optimizations. Extensive experiments, comparing 15 competitors on 10 real datasets, demonstrate that our solutions, especially GEBE(p), obtain superior result utility than all competitors for top-N recommendation and link prediction, while being up to orders of magnitude faster.