Semi-supervised Network Embedding with Differentiable Deep Quantisation

Abstract

Learning accurate low-dimensional embeddings for a network is a crucial task as it facilitates many downstream network analytics tasks. For large networks, the trained embeddings often require a significant amount of space to store, making storage and processing a challenge. Building on our previous work on semi-supervised network embedding, we develop d-SNEQ, a differentiable DNN-based quantisation method for network embedding. d-SNEQ incorporates a rank loss to equip the learned quantisation codes with rich high-order information and is able to substantially compress the size of trained embeddings, thus reducing storage footprint and accelerating retrieval speed. We also propose a new evaluation metric, path prediction, to fairly and more directly evaluate model performance on the preservation of high-order information. Our evaluation on four real-world networks of diverse characteristics shows that d-SNEQ outperforms a number of state-of-the-art embedding methods in link prediction, path prediction, node classification, and node recommendation while being far more space- and time-efficient.

References in corpus (9)

• FaceNet: A Unified Embedding for Face Recognition and Clustering
• DeepWalk: Online Learning of Social Representations
• LINE: Large-scale Information Network Embedding
• The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables
• Network Embedding as Matrix Factorization: Unifying DeepWalk, LINE, PTE, and node2vec
• Deep Gaussian Embedding of Graphs: Unsupervised Inductive Learning via Ranking
• Inner Product Similarity Search using Compositional Codes
• Hierarchical Graph Convolutional Networks for Semi-supervised Node Classification
• Semi-supervised Embedding in Attributed Networks with Outliers