Abstract
Social network alignment, identifying social accounts of the same individual across different social networks, shows fundamental importance in a wide spectrum of applications, such as link prediction and information diffusion. Individuals more often than not join in multiple social networks, and it is in fact much too expensive or even impossible to acquiring supervision for guiding the alignment. To the best of our knowledge, few method in the literature can align multiple social networks without supervision. In this article, we propose to study the problem of unsupervised multiple social network alignment. To address this problem, we propose a novel unsupervised model of joint Matrix factorization with a diagonal Cone under orthogonal Constraint, referred to as MC2. Its core idea is to embed and align multiple social networks in the common subspace via an unsupervised approach. Specifically, in MC2 model, we first design a matrix optimization to infer the common subspace from different social networks. To address the nonconvex optimization, we then design an efficient alternating algorithm by leveraging its inherent functional property. Through extensive experiments on real-world datasets, we demonstrate that the proposed MC2 model significantly outperforms the state-of-the-art methods.
- [1] Ishaan Gulrajani, Faruk Ahmed, Martín Arjovsky, Vincent Dumoulin, and Aaron C. Courville. 2017. Improved training of Wasserstein GANs. In Advance in NeurIPS, 5767–5777.Google Scholar
- [2] . 2016. Bass: A bootstrapping approach for aligning heterogenous social networks. In Proceeding of the ECML-PKDD. 459–475.Google ScholarDigital Library
- [3] . 2018. Exploiting structural and temporal evolution in dynamic link prediction. In Proceeding of the CIKM. 427–436.Google ScholarDigital Library
- [4] . 2019. Cross-Network embedding for multi-network alignment. In Proceeding of the Web Conf. (WWW).Google ScholarDigital Library
- [5] . 2006. Orthogonal nonnegative matrix tri-factorizations for clustering. In Proceeding of the SIGKDD.Google Scholar
- [6] Ngoc Diep Ho. 2008. Nonnegative matrix factorization: Algorithms and applications, Doctoral thesis, Universié catholique de Louvain, 1–165.Google Scholar
- [7] . 2020. Adaptive network alignment with unsupervised and multi-order convolutional networks. In Proceedings of ICDE. IEEE, 85–96.Google Scholar
- [8] . 2013. Inferring anchor links across multiple heterogeneous social networks. In Proceeding of the CIKM. 179–188.Google ScholarDigital Library
- [9] . 1999. Learning the parts of objects by non-negative matrix factorization. Nature 401, 6755 (1999), 788–791.Google Scholar
- [10] . 2000. Algorithms for non-negative matrix factorization. In Advances in the (NIPS’20), 535–541.Google Scholar
- [11] . 2018. Distribution distance minimization for unsupervised user identity linkage. In Proceeding of the CIKM. 447–456.Google ScholarDigital Library
- [12] . 2019. Adversarial learning for weakly-supervised social network alignment. In Proceeding of the AAAI.Google ScholarDigital Library
- [13] . 2009. IsoRankN: Spectral methods for global alignment of multiple protein networks. Bioinformatics 25, 12 (2009), i253–i258.Google ScholarDigital Library
- [14] . 2016. Aligning users across social networks using network embedding. In Proceeding of the IJCAI. 1774–1780.Google Scholar
- [15] . 2014. Hydra: Large-scale social identity linkage via heterogeneous behavior modeling. In Proceeding of the SIGMOD. 51–62.Google ScholarDigital Library
- [16] . 2017. Cross-Domain recommendation: An embedding and mapping approach. In Proceeding of the IJCAI. 2464–2470.Google ScholarCross Ref
- [17] . 2016. Predict anchor links across social networks via an embedding approach. In Proceeding of the IJCAI. 1823–1829.Google Scholar
- [18] . 2016. User identity linkage by latent user space modelling. In Proceeding of the SIGKDD. 1775–1784.Google ScholarDigital Library
- [19] Yu. E. Nesterov. 1983. A method for solving the convex programming problem with convergence rate \(O(1/k^2)\). Proceedings of the USSR Academy of Sciences (Mathematics), Vol. 3, 543–547.Google Scholar
- [20] . 2016. Asymmetric transitivity preserving graph embedding. In Proceeding of the SIGKDD. 1105–1114.Google ScholarDigital Library
- [21] . 2016. Motives for online friending and following: The dark side of social network site connections. Social Media+ Society 2, 3 (2016), 1–13.Google Scholar
- [22] . 2014. Deepwalk: Online learning of social representations. In Proceedings of the KDD. 701–710.Google ScholarDigital Library
- [23] . 2017. User identity linkage across online social networks: A review. SIGKDD Explorations 18, 2 (2017), 5–17.Google ScholarDigital Library
- [24] . 2018. MASTER: Across multiple social networks, integrate attribute and structure embedding for reconciliation. In Proceeding of the IJCAI. 3863–3869.Google ScholarCross Ref
- [25] . 2020. Perfect: A hyperbolic embedding for joint user and community alignment. In Proceedings of the 20th ICDM. IEEE, 501–510.Google ScholarCross Ref
- [26] . 2014. Mapping users across networks by manifold alignment on hypergraph. In Proceeding of the AAAI. 159–165.Google ScholarCross Ref
- [27] . 2015. LINE: Large-scale information network embedding. In Proceeding of the WWW. 1067–1077.Google ScholarDigital Library
- [28] . 2012. It is all about being popular: The effects of need for popularity on social network site use. Cyberpsychology Behavior & Social Networking 15, 1 (2012), 37–42.Google ScholarCross Ref
- [29] . 2020. Hyperbolic user identity linkage across social networks. In Proceedings of IEEE Global Communications Conference. IEEE, 1–6.Google ScholarDigital Library
- [30] . 2018. User identity linkage across social networks via linked heterogeneous network embedding. World Wide Web - Internet and Web Information Systems 22, 6 (2018), 2611–2632.Google Scholar
- [31] . 2014. Learning the consistent behavior of common users for target node prediction across social networks. In Proceeding of the ICML. 298–306.Google Scholar
- [32] . 2018. Unsupervised user identity linkage via factoid embedding. In Proceeding of the ICDM. 1338–1343.Google ScholarCross Ref
- [33] . 2009. Connecting corresponding identities across communities. In Proceeding of the ICWSM. 354–357.Google ScholarCross Ref
- [34] . 2013. Connecting users across social media sites: A behavioral-modeling approach. In Proceeding of the SIGKDD. 41–49.Google ScholarDigital Library
- [35] . 2015. Influence maximization across partially aligned heterogenous social networks. In Proceeding of the PAKDD. 58–69.Google ScholarCross Ref
- [36] . 2018. MEgo2Vec: Embedding matched ego networks for user alignment across social networks. In Proceeding of the CIKM. 327–336.Google ScholarDigital Library
- [37] . 2015. Multiple anonymized social networks alignment. In Proceeding of the ICDM. 855–864.Google ScholarDigital Library
- [38] . 2016. Pct: Partial co-alignment of social networks. In Proceeding of the WWW. 749–759.Google ScholarDigital Library
- [39] . 2019. Multilevel network alignment. In Proceeding of the Web Conference.Google ScholarDigital Library
- [40] . 2015. COSNET: Connecting heterogeneous social networks with local and global consistency. In Proceeding of the SIGKDD. 1485–1494.Google ScholarDigital Library
- [41] . 2022. Hierarchical graph transformer with adaptive node sampling. In Advance in NeurIPS, 1–13.Google Scholar
- [42] . 2022. ProtGNN: Towards self-explaining graph neural networks. In Proceedings of the 36th AAAI. AAAI Press, 9127–9135.Google ScholarCross Ref
- [43] . 2017. Efficient incremental dynamic link prediction algorithms in social network. Knowledge-Based Systems 132 (2017), 226–235.Google ScholarCross Ref
- [44] . 2018. CoLink: An unsupervised framework for user identity linkage. In Proceeding of the AAAI. 5714–5721.Google ScholarCross Ref
- [45] . 2018. DeepLink: A deep learning approach for user identity linkage. In Proceeding of the INFOCOM. 1313–1321.Google ScholarDigital Library
Index Terms
- MC2: Unsupervised Multiple Social Network Alignment
Recommendations
Social network alignment: a bi-layer graph attention neural networks based method
AbstractThe task of social network alignment is to identify the user nodes which are active in multiple social networks simultaneously, thus the information from multiple social networks can be integrated to conduct some downstream tasks. Existing network ...
REBORN: Transfer learning based social network alignment
AbstractSocial network alignment (SNA) aims to identify the same person across different social networks. Existing SNA methods can be categorized as supervised, semi-supervised and unsupervised. However, for supervised and semi-supervised methods, ...
Multiple graph unsupervised feature selection
Feature selection improves the quality of the model by filtering out the noisy or redundant part. In the unsupervised scenarios, the selection is challenging due to the unavailability of the labels. To overcome that, the graphs which can unfold the ...
Comments