A numerical evaluation of the accuracy of influence maximization algorithms

Kingi, Hautahi; Wang, Li-An Daniel; Shafer, Tom; Huynh, Minh; Trinh, Mike; Heuser, Aaron; Rochester, G. D.; Paredes, Antonio

doi:10.1007/s13278-020-00680-5

Cited by 5 publications

(3 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We investigate two adversarial objectives of a socialbot: influencing people while evading socialbot detection. While the first one can be modeled as an IM task on graph networks, traditional IM algorithms-e.g., [3,25,27], assume that the number of seed nodes is relatively small and all nodes are equally acquirable, all of which are not applicable in the socialbot context as previously described. There have been also a few works-e.g., [33,49], that utilizes RL to IM task.…”

Section: Adversarial Socialbot Learningmentioning

confidence: 99%

“…Simultaneously, it also needs to systematically constrain its online behaviors such that it will not easily expose itself to socialbot detectors. Although the IM problem has been widely studied by several works [3,24,25,27], they only focus on maximizing the network influence given a fixed and static budget # of seed nodes (that is relatively small) and they assume that every node is equally acquirable. However, these assumptions are not practical in our context.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Socialbots on Fire: Modeling Adversarial Behaviors of Socialbots via Multi-Agent Hierarchical Reinforcement Learning

Le¹,

Tran-Thanh²,

Lee³

2021

Preprint

View full text Add to dashboard Cite

Socialbots are software-driven user accounts on social platforms, acting autonomously (mimicking human behavior), with the aims to influence the opinions of other users or spread targeted misinformation for particular goals. As socialbots undermine the ecosystem of social platforms, they are often considered harmful. As such, there have been several computational efforts to auto-detect the socialbots. However, to our best knowledge, the adversarial nature of these socialbots has not yet been studied. This begs a question "can adversaries, controlling socialbots, exploit AI techniques to their advantage?" To this question, we successfully demonstrate that indeed it is possible for adversaries to exploit computational learning mechanism such as reinforcement learning (RL) to maximize the influence of socialbots while avoiding being detected. We first formulate the adversarial socialbot learning as a cooperative game between two functional hierarchical RL agents. While one agent curates a sequence of activities that can avoid the detection, the other agent aims to maximize network influence by selectively connecting with right users. Our proposed policy networks train with a vast amount of synthetic graphs and generalize better than baselines on unseen real-life graphs both in terms of maximizing network influence (up to +18%) and sustainable stealthiness (up to +40% undetectability) under a strong bot detector (with 90% detection accuracy). During inference, the complexity of our approach scales linearly, independent of a network's structure and the virality of news. This makes our approach a practical adversarial attack when deployed in a real-life setting.

show abstract

Section: Adversarial Socialbot Learningmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Socialbots on Fire: Modeling Adversarial Behaviors of Socialbots via Multi-Agent Hierarchical Reinforcement Learning

Le¹,

Tran-Thanh²,

Lee³

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…In the same paper, the authors proposed a greedy heuristic providing a good guaranteed solution to the problem. Since then, a great number heuristics and approximations have been proposed to tackle this problem (Kempe et al 2005;Chen et al 2010;Jung et al 2012;Liu et al 2014;Srivastava et al 2015;Kingi et al 2020;Tang et al 2018;Hajdu et al 2018). The survey paper of Li et al gives an excellent summary of the field (Li et al 2018).…”

Section: Introductionmentioning

confidence: 99%

Evaluating the role of community detection in improving influence maximization heuristics

Hajdú

Krész

Bóta

2021

Soc. Netw. Anal. Min.

View full text Add to dashboard Cite

Both community detection and influence maximization are well-researched fields of network science. Here, we investigate how several popular community detection algorithms can be used as part of a heuristic approach to influence maximization. The heuristic is based on the community value, a node-based metric defined on the outputs of overlapping community detection algorithms. This metric is used to select nodes as high influence candidates for expanding the set of influential nodes. Our aim in this paper is twofold. First, we evaluate the performance of eight frequently used overlapping community detection algorithms on this specific task to show how much improvement can be gained compared to the originally proposed method of Kempe et al. Second, selecting the community detection algorithm(s) with the best performance, we propose a variant of the influence maximization heuristic with significantly reduced runtime, at the cost of slightly reduced quality of the output. We use both artificial benchmarks and real-life networks to evaluate the performance of our approach.

show abstract

An improved influence maximization method for social networks based on genetic algorithm

Lotf

Azgomi

Dishabi

2022

Physica A: Statistical Mechanics and its Applications

View full text Add to dashboard Cite

A numerical evaluation of the accuracy of influence maximization algorithms

Cited by 5 publications

References 38 publications

Socialbots on Fire: Modeling Adversarial Behaviors of Socialbots via Multi-Agent Hierarchical Reinforcement Learning

Socialbots on Fire: Modeling Adversarial Behaviors of Socialbots via Multi-Agent Hierarchical Reinforcement Learning

Evaluating the role of community detection in improving influence maximization heuristics

An improved influence maximization method for social networks based on genetic algorithm

Contact Info

Product

Resources

About