期刊
出版社
ASSOC COMPUTING MACHINERY
DOI: 10.1145/3219819.3220068
关键词
network embedding; regular equivalence; recurrent neural network
类别
资金
- National Program on Key Basic Research Project [2015CB352300]
- National Natural Science Foundation of China [61772304, 61521002, 61531006, 61702296, U1611461]
- NSF [IIS-1526499, IIS-1763325, CNS-1626432]
- Tsinghua-Tencent Joint Laboratory for Internet Innovation Technology
- Young Elite Scientist Sponsorship Program by CAST
- NSFC [61672313]
Network embedding aims to preserve vertex similarity in an embedding space. Existing approaches usually define the similarity by direct links or common neighborhoods between nodes, i.e. structural equivalence. However, vertexes which reside in different parts of the network may have similar roles or positions, i.e. regular equivalence, which is largely ignored by the literature of network embedding. Regular equivalence is defined in a recursive way that two regularly equivalent vertexes have network neighbors which are also regularly equivalent. Accordingly, we propose a new approach named Deep Recursive Network Embedding (DRNE) to learn network embeddings with regular equivalence. More specifically, we propose a layer normalized LSTM to represent each node by aggregating the representations of their neighborhoods in a recursive way. We theoretically prove that some popular and typical centrality measures which are consistent with regular equivalence are optimal solutions of our model. This is also demonstrated by empirical results that the learned node representations can well predict the indexes of regular equivalence and related centrality scores. Furthermore, the learned node representations can be directly used for end applications like structural role classification in networks, and the experimental results show that our method can consistently outperform centrality-based methods and other state-of-the-art network embedding methods.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据