@inproceedings{eac32744bf564ea685b236102e871af5,
title = "INEAT: Incomplete network alignment",
abstract = "Network alignment and network completion are two fundamental cornerstones behind many high-impact graph mining applications. The state-of-the-arts have been addressing these tasks in parallel. In this paper, we argue that network alignment and completion are inherently complementary with each other, and hence propose to jointly address them so that the two tasks can benefit from each other. We formulate it from the optimization perspective, and propose an effective algorithm iNEAT to solve it. The proposed method offers two distinctive advantages. First (Alignment accuracy), our method benefits from higher-quality input networks while mitigates the effect of incorrectly inferred links introduced by the completion task itself. Second (Alignment efficiency), thanks to the low-rank structure of the complete networks and alignment matrix, the alignment can be significantly accelerated. The extensive experiments demonstrate the performance of our algorithm.",
keywords = "Incomplete network alignment, Low-rank, Network completion",
author = "Si Zhang and Hanghang Tong and Jie Tang and Jiejun Xu and Wei Fan",
note = "Funding Information: Network Completion. Kim et al. propose an Expectation Maximization (EM) based method that fits the network structure under the Kronecker graph model and reestimates the parameters using a scalable Gibbs sampling approach [17]. Another work proposed by Masrour et al. leverages node similarity matrix [3]. In addition, inferring the missing edges can be considered as an adjacency matrix completion problem. One well-known matrix completion method is based on singular value thresholding (SVT) to minimize the nuclear norm [18]. To speed up, Toh and Yun propose an accelerated proximal gradient method to solve a nuclear norm regularized linear least squares problem [11]. In addition to matrix completion, tensor completion is very powerful in many areas [19]. Some work use tensor completion on multiple aligned networks [4]. Nonetheless, how input networks are aligned beforehand is not answered. 7. Conclusion In the era of big data, the multi-sourced and incomplete characteristics often co-exist in many real networks. However, the state-of-the-arts have been largely addressing them in parallel. In this paper, we propose to jointly address network alignment and network completion so that the two tasks can benefit from each other. We formulate incomplete network alignment problem as an optimization problem and propose a multiplicative update algorithm (INEAT). To our best knowledge, the proposed INEAT algorithm is the first network alignment algorithm with a provable linear complexity. The empirical evaluations demonstrate both the effectiveness in network alignment and network completion, and the efficiency. Future work includes extending our algorithm to handle attributed networks. 8. Acknowledgement This material is supported by the National Science Foundation under Grant No. IIS-1651203, IIS-1715385, IIS- Funding Information: run time and accuracy. ber of nodes. 1743040, and CNS-1629888, by DTRA under the grant number HDTRA1-16-0017, by Army Research Office under the contract number W911NF-16-1-0168, by National Institutes of Health under the grant number R01LM011986, National High-tech R&D Program (2015AA124102), National Natural Science Foundation of China (61561130160), a research fund supported by MSRA, the Royal Society-Newton Advanced Fellowship Award and by additional gifts from Huawei and Baidu. The content of the information in this document does not necessarily reflect the position or the policy of the Government, and no official endorsement should be inferred. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation here on. Publisher Copyright: {\textcopyright} 2017 IEEE.; 17th IEEE International Conference on Data Mining, ICDM 2017 ; Conference date: 18-11-2017 Through 21-11-2017",
year = "2017",
month = dec,
day = "15",
doi = "10.1109/ICDM.2017.160",
language = "English (US)",
series = "Proceedings - IEEE International Conference on Data Mining, ICDM",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "1189--1194",
editor = "George Karypis and Srinivas Alu and Vijay Raghavan and Xindong Wu and Lucio Miele",
booktitle = "Proceedings - 17th IEEE International Conference on Data Mining, ICDM 2017",
address = "United States",
}