Reinforcement Learning-based Black-Box Evasion Attacks to Link Prediction in Dynamic Graphs

Link prediction in dynamic graphs (LPDG) is an important research problem that has diverse applications such as online recommendations, studies on disease contagion, organizational studies, etc. Various LPDG methods based on graph embedding and graph neural networks have been recently proposed and achieved state-of-the-art performance. In this paper, we study the vulnerability of LPDG methods and propose the first practical black-box evasion attack. Specifically, given a trained LPDG model, our attack aims to perturb the graph structure, without knowing to model parameters, model architecture, etc., such that the LPDG model makes as many wrong predicted links as possible. We design our attack based on a stochastic policy-based RL algorithm. Moreover, we evaluate our attack on three real-world graph datasets from different application domains. Experimental results show that our attack is both effective and efficient.

[1]  Purnamrita Sarkar,et al.  Nonparametric Link Prediction in Dynamic Networks , 2012, ICML.

[2]  Dan Wang,et al.  Adversarial Network Embedding , 2017, AAAI.

[3]  Yan Liu,et al.  DynGEM: Deep Embedding Method for Dynamic Graphs , 2018, ArXiv.

[4]  Ling Chen,et al.  Link prediction in dynamic social networks by integrating different types of information , 2014, Applied Intelligence.

[5]  Ryan A. Rossi,et al.  Continuous-Time Dynamic Network Embeddings , 2018, WWW.

[6]  Charu C. Aggarwal,et al.  NetWalk: A Flexible Deep Embedding Approach for Anomaly Detection in Dynamic Networks , 2018, KDD.

[7]  Wenbing Huang,et al.  A Restricted Black-Box Adversarial Framework Towards Attacking Graph Embedding Models , 2019, AAAI.

[8]  Daniel Lowd,et al.  Convex Adversarial Collective Classification , 2013, ICML.

[9]  Zhanxing Zhu,et al.  Spatio-temporal Graph Convolutional Neural Network: A Deep Learning Framework for Traffic Forecasting , 2017, IJCAI.

[10]  Jure Leskovec,et al.  node2vec: Scalable Feature Learning for Networks , 2016, KDD.

[11]  Yueting Zhuang,et al.  Dynamic Network Embedding by Modeling Triadic Closure Process , 2018, AAAI.

[12]  Stephan Günnemann,et al.  Adversarial Attacks on Node Embeddings via Graph Poisoning , 2018, ICML.

[13]  Sijia Liu,et al.  Topology Attack and Defense for Graph Neural Networks: An Optimization Perspective , 2019, IJCAI.

[14]  Mingzhe Wang,et al.  LINE: Large-scale Information Network Embedding , 2015, WWW.

[15]  Liming Zhu,et al.  Adversarial Examples on Graph Data: Deep Insights into Attack and Defense , 2019 .

[16]  Max Welling,et al.  Semi-Supervised Classification with Graph Convolutional Networks , 2016, ICLR.

[17]  Yu Wang,et al.  Graph regularized nonnegative matrix factorization for temporal link prediction in dynamic networks , 2018 .

[18]  Stephan Gunnemann,et al.  Adversarial Attacks on Graph Neural Networks via Meta Learning , 2019, ICLR.

[19]  Palash Goyal,et al.  dyngraph2vec: Capturing Network Dynamics using Dynamic Graph Representation Learning , 2018, Knowl. Based Syst..

[20]  Binghui Wang,et al.  Attacking Graph-based Classification via Manipulating the Graph Structure , 2019, CCS.

[21]  Yu Liu,et al.  T-GCN: A Temporal Graph Convolutional Network for Traffic Prediction , 2018, IEEE Transactions on Intelligent Transportation Systems.

[22]  Steven Skiena,et al.  DeepWalk: online learning of social representations , 2014, KDD.

[23]  Suhang Wang,et al.  Adversarial Attacks on Graph Neural Networks via Node Injections: A Hierarchical Reinforcement Learning Approach , 2020, WWW.

[24]  Li Sun,et al.  Efficient incremental dynamic link prediction algorithms in social network , 2017, Knowl. Based Syst..

[25]  Justin Zhan,et al.  Deep Learning for Link Prediction in Dynamic Networks Using Weak Estimators , 2018, IEEE Access.

[26]  Le Song,et al.  Adversarial Attack on Graph Structured Data , 2018, ICML.

[27]  Qi Xuan,et al.  Fast Gradient Attack on Network Embedding , 2018, ArXiv.

[28]  Huan Liu,et al.  Attributed Network Embedding for Learning in a Dynamic Environment , 2017, CIKM.

[29]  Hui Li,et al.  A Deep Learning Approach to Link Prediction in Dynamic Networks , 2014, SDM.

[30]  Pietro Liò,et al.  Graph Attention Networks , 2017, ICLR.

[31]  Alessandro Rozza,et al.  Dynamic Graph Convolutional Networks , 2017, Pattern Recognit..

[32]  Jure Leskovec,et al.  How Powerful are Graph Neural Networks? , 2018, ICLR.

[33]  Junjie Wu,et al.  Embedding Temporal Network via Neighborhood Formation , 2018, KDD.

[34]  Jie Chen,et al.  EvolveGCN: Evolving Graph Convolutional Networks for Dynamic Graphs , 2020, AAAI.

[35]  Yizheng Chen,et al.  Practical Attacks Against Graph-based Clustering , 2017, CCS.

[36]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[37]  Min Du,et al.  Time-aware Gradient Attack on Dynamic Network Link Prediction , 2019, ArXiv.

[38]  Stephan Günnemann,et al.  Adversarial Attacks on Neural Networks for Graph Data , 2018, KDD.

[39]  Le Song,et al.  Know-Evolve: Deep Temporal Reasoning for Dynamic Knowledge Graphs , 2017, ICML.

[40]  Binghui Wang,et al.  Backdoor Attacks to Graph Neural Networks , 2020, ArXiv.

[41]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[42]  Mingjie Sun,et al.  Data Poisoning Attack against Unsupervised Node Embedding Methods , 2018, ArXiv.

[43]  Henry Zhu,et al.  Soft Actor-Critic Algorithms and Applications , 2018, ArXiv.

[44]  Qi Xuan,et al.  E-LSTM-D: A Deep Learning Framework for Dynamic Network Link Prediction , 2019, IEEE Transactions on Systems, Man, and Cybernetics: Systems.