影響最大化問題(Influence maximization problem)指的是在一個給定的網路上，希望能找出一組一定數量的的種子點(seed nodes)，能夠將訊息最廣泛傳播。在現今大型的社群網路中，我們希望能找到兼顧效率及精確度的影響最大化演算法。傳統的貪婪演算法(greedy algorithm)有高度的準確度，但缺點是需要非常冗長的時間，且需要在擁有完整的網路下才能做運算。另一方面，啟發式演算法雖然擁有高效率，但有精確度在一些情形下可能會非常低的缺點。本篇論文提出了一種方法，它結合了圖神經網路(graph neural network)及深度強化式學習(deep reinforcement learning)。這個方法擁有泛用性(generalizability )，可以將它在較小的圖上學到的內容應用在更大的圖上，並迅速得到好的結果。

Influence maximization is the problem that discusses how to select a set of seed nodes to propagate the information most widely. Nowadays, for large-scale social networks, we want to find an influence maximization algorithm with both efficiency and effectiveness. Traditional greedy algorithm has high accuracy but low efficiency and must have a complete graph to work. On the other hand, though heuristic methods guarantee time efficiency, they suffer from the risk of low accuracy under some conditions. In this paper, we propose an approach that combines graph neural network and deep reinforcement learning. The approach has generalizability that can apply what it learned on a smaller network to get a good answer quickly on a larger network.

中文摘要 i
Abstract ii
Acknowledgements iii
List of Figures vi
List of Tables vii
1 Introduction 1
2 Related work 3
2.1 Independent Cascade Model . . . . . . . . . . . . . . . 3
2.2 Related research . . . . . . . . . . . . . . . . . . . . . . 6
3 Proposed Scheme 9
3.1 Background of the method . . . . . . . . . . . . . . . . 9
3.1.1 Graph embedding . . . . . . . . . . . . . . . . . 9
3.1.2 Deep reinforcement learning . . . . . . . . . . . 10
3.2 System of the IM problem in RL . . . . . . . . . . . . . 14
3.3 Algorithm representation . . . . . . . . . . . . . . . . . 15
4 Experiments 24
4.1 Experiment Setup . . . . . . . . . . . . . . . . . . . . . 24
4.2 Experiment results . . . . . . . . . . . . . . . . . . . . 26
4.2.1 Experiments on generated data . . . . . . . . . . 27
4.2.2 Experiments on real-world data . . . . . . . . . 31
5 Conclusions 34
Bibliography 34

[1] P. Domingos and M. Richardson, "Mining the network value of customers." In Proceedings of the 7th ACM SIGKDD Conference on
Knowledge Discovery and Data Mining, 2001.

[2] D. Kempe, J. M. Kleinberg, and É. Tardos, "Maximizing the spread of influence through a social network." In Proceedings of the 9th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2003

[3] W. Chen, Y. Wang, and S. Yang, "Efficient influence maximization in social networks." In Proceedings of the 15th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2009.

[4] M. Kimura and K. Saito, “Tractable models for information diffusion in social networks.” Knowledge Discovery in Databases: PKDD, 2006.

[5] Y Li, J Fan, Y Wang, K Tan, "Influence maximization on social graphs: A survey." IEEE Transactions on Knowledge and Data
Engineering, 2018

[6] J. Leskovec, A. Krause, C. Guestrin, C. Faloutsos, J. VanBriesen, and N. Glance, “Cost-effective outbreak detection in networks.” 13th ACM SIGKDD Int. Conf. Knowl. Discovery Data Mining, 2007

[7] A. Goyal, W. Lu, and L. V. Lakshmanan, “Celf++: Optimizing the greedy algorithm for influence maximization in social networks” WWW, 2011, pp. 47–48

[8] C. Zhou, P. Zhang, W. Zang, and L. Guo, “On the upper bounds of spread for greedy algorithms in social network influence maximization.” IEEE Transactions on Knowledge and Data Engineering, 2015

[9] Y.Wang, G. Cong, G. Song, and K. Xie, “Community-based, greedy algorithm for mining top-k influential nodes in mobile social networks.” 16th ACM SIGKDD Int. Conf. Knowl. Discovery Data Mining, 2010

[10] M Heidari, M Asadpour, H Faili, "SMG Fast scalable greedy algorithm for influence maximization in social networks." Physica A: Statistical Mechanics and its Applications, Volume 420, 15 February 2015

[11] M. Kimura, K. Saito, and R. Nakano, “Extracting influential nodes for information diffusion on a social network.” In AAAI, pages 1371-1376, 2007

[12] K. Jung,W. Heo, andW. Chen, “IRIE: Scalable and robust influence maximization in social networks.” 2012 IEEE 12th International Conference on Data Mining, 2012, pp. 918-923

[13] WX Lu, C Zhou, J Wu, "Big social network influence maximization via recursively estimating influence spread." Knowledge-Based Systems Volume, 113, 1 December 2016

[14] A Zareie, A Sheikhahmadi, K Khamforoosh, "Influence maximization in social networks based on TOPSIS." Expert Systems with Applications Volume, 108, 15 October 2018

[15] F Lu, W Zhang, L Shao, X Jiang, P Xu, H Jin, "Scalable influence maximization under independent cascade model." Journal of Network and Computer Applications, Volume 86, 15 May 2017

[16] Q Liqing, G Chunmei, Z Shuang, T Xiangbo, Zhang Mingjv, "TSIM: A Two-Stage Selection Algorithm for Influence Maximization in Social Networks." IEEE Access, 2020

[17] Z.-L. Luo,W.-D. Cai, Y.-J. Li, D. Peng, "in Recent Progress in Data Engineering and Internet Technology, 2012, pp. 485–490" LNEE, 2012

[18] G Tong, W Wu, S Tang, DZ Du, "Adaptive Influence Maximization in Dynamic Social Networks." IEEE/ACM Transactions on Networking, Volume 25, February 2017

[19] N Ohsaka, T Akiba, Y Yoshida, K Kawarabayashi, "Fast and Accurate Influence Maximization on Large Networks with Pruned Monte Carlo Simulations." AAAI Publications, Twenty-Eighth AAAI Conference on Artificial Intelligence, 2014, Québec, Canada

[20] E Cohen, D Delling, T Pajor, RF Werneck, "Sketch-based influence maximization and computation: scaling up with guarantees." Proceedings of the 23rd ACM International Conference on Conference on Information and Knowledge Management, 2014, New York, United States

[21] S Cheng, H Shen, J Huang, W Chen, X Cheng, "IMRank influence maximization via finding self-consistent ranking." Proceedings of the 37th international ACM SIGIR conference on Research & development in information retrieval, 2014

[22] Q He, XWang, Z Lei,MHuang, Y Cai, L Ma, "TIFIM: A two-stage iterative framework for influence maximization in social networks." Applied Mathematics and Computation , ELSEVIER, 2019

[23] Q Jiang, G Song, C Gao, Y Wang, W Si, K Xie, "Simulated Annealing Based Influence Maximization in Social Networks." Twenty-Fifth AAAI Conference on Artificial Intelligence, 2011, San Francisco, United States

[24] C Tsai, Y Yang, M Chiang, "A Genetic NewGreedy Algorithm for Influence Maximization in Social Network." IEEE International Conference on Systems, Man, and Cybernetics, 2015, City University, Hong Kong

[25] K Zhang, H Du, MW Feldman, "Maximizing influence in a social network: Improved results using a genetic algorithm." Physica A: Statistical Mechanics and its Applications, 2017

[26] S Peng, A Yang, L Cao, S Yu, D Xie, "Social influence modeling using information theory in mobile social networks." Information Sciences, 2017 – Elsevier

[27] S Cheng, H Shen, J Huang, G Zhang, XCheng, "StaticGreedy: Solving the Scalability-Accuracy Dilemma in Influence Maximization." Proceedings of the 22nd ACM international conference on Information & Knowledge Management, 2014, San Francisco, United States

[28] C. Borgs, M. Brautbar, J. Chayes, and B. Lucier, "Maximizing social influence in nearly optimal time." Proc. 25th Annu. ACM-SIAM Symp Discrete Algorithms, 2014

[29] Y. Tang, X. Xiao, and Y. Shi, "Influence maximization: Near-optimal time complexity meets practical efficiency." Proceedings of the 2014 ACM SIGMOD International Conference on Management of Data, Utah, United States

[30] G Tong, W Wu, DZ Du, "Distributed Rumor Blocking With Multiple Positive Cascades." IEEE Transactions on Computational Social Systems, 2018, vol. 5

[31] M Li, F Liu, "Game Theory-Based Network Rumor Spreading Model." International Conference on Network and Information Systems for Computers (ICNISC), 2016, Wuhan, China

[32] S Li, Y Zhu, D Li, D Kim, H Huang, "Rumor Restriction in Online Social Networks." IEEE 32nd International Performance Computing and Communications Conference (IPCCC), 2013, San Diego, CA, USA

[33] J Guo, P Zhang, C Zhou, Y Cao, L Guo, "Personalized Influence Maximization on Social Networks." Proceedings of the 22nd ACM international conference on Information & Knowledge Management, San Francisco, United States

[34] Y Li, D Zhang, "Realtime Targeted Influence Maximization for Online Advertisements." Proceedings of the VLDB Endowment, 2015, Kohala Coast, Hawaii

[35] T Zhou, J Cao, B Liu, S Xu, Z Zhu, J Luo, "Location-Based Influence Maximization in Social Networks." Proceedings of the 24th ACM International on Conference on Information and Knowledge
Management, Melbourne, Australia

[36] G Li, S Chen, J Feng, K Tan, W Li, "Efficient location-aware influence maximization." Proceedings of the 2014 ACM SIGMOD International Conference, 2014

[37] Dai, Hanjun, Khalil, Elias B, Zhang, Yuyu, Dilkina, Bistra, and Song, Le. "Learning combinatorial optimization algorithms over graphs." Proceedings of NIPS (2017), pp. 6348-6358

[38] Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A. A., Veness, J., Bellemare, M. G., Graves, A., Riedmiller, M., Fidjeland, A. K., Ostrovski, G., et al. "Human-level control through deep reinforcement learning." Nature, 518(7540): 529–533, 2015.

[39] Keyulu Xu,Weihua Hu, Jure Leskovec, and Stefanie Jegelka. "How powerful are graph neural networks?" In 7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6-9, 2019

[40] R. Albert, Barabasi AL. "Statistical Mechanics Of Complex Networks." Rev. Mod. Phys. 74, 47 (2002).

[41] Network Repository Datasets [Online], Available: http://networkrepository.com/index.php

[42] SNAP Datasets: Stanford Large Network Dataset Collection. [Online], Available: http://snap.stanford.edu/data

[43] Diederik Kingma and Jimmy Ba. " Adam: A method for stochastic optimization." International Conference on Learning Representations, 2015, San Diego, United State