Debunking the Myths of Influence Maximization

Arora, Akhil; Galhotra, Sainyam; Ranu, Sayan

doi:10.1145/3218967

Cited by 49 publications

(159 citation statements)

References 0 publications

Supporting

Mentioning

155

Contrasting

Order By: Relevance

“…Clearly presented by Arora et al [1] there are three main categories of IM algorithms: greedy, sampling and approximation.…”

Section: Algorithmsmentioning

confidence: 99%

“…Conclusions: Greedy solutions require hours or days of processing due to the repeated computation of σ M C (S). They perform poorly for seed sets larger than 50 and do not scale to large datasets [1]. Heuristics don't offer theoretical guarantee and often lack in precision.…”

Section: Algorithmsmentioning

confidence: 99%

“…We choose IMM [25] since it is the best compromise between computation speed, scalability and accuracy, especially on large datasets. The code for IMM is provided by [1] Stage I: Pre-processing. First, IMM is run in its entirety and adds k users to its seed set S IM M (k = 10, 000 and optimal seed = 0.1).…”

Section: Experimental Processmentioning

confidence: 99%

“…However, as far as we know, no existing IM algorithm can work within a real-time bidding environment and satisfy it's requirements. Indeed, whereas existing algorithms take hours to find seed sets up to 200 seeds in a large social network [1], they do not scale up or take into account real time streams of users.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

RTIM: A Real-Time Influence Maximization Strategy

Dupuis

Mouza

Travers

et al. 2019

Web Information Systems Engineering – WISE 2019

View full text Add to dashboard Cite

Influence Maximization (IM) consists in finding in a network the top-k influencers who will maximize the diffusion of information. However, the exponential growth of online advertisement is due to Real-Time Bidding (RTB) which targets users on webpages. It requires complex ad placement decisions in real-time to face a high-speed stream of users. In order to stay relevant, the IM problem should be updated to answer RTB needs. While traditional IM generates a static set of influencers, they do not fit with an RTB environment which requires dynamic influence targeting. This paper proposes RTIM, the first IM algorithm capable of targeting users in a RTB environment. We also analyze influence scores of users in several social networks and provide a thorough experimental process to compare static versus dynamic IM solutions.

show abstract

“…Clearly presented by Arora et al [1] there are three main categories of IM algorithms: greedy, sampling and approximation.…”

Section: Algorithmsmentioning

confidence: 99%

Section: Algorithmsmentioning

confidence: 99%

Section: Experimental Processmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

RTIM: A Real-Time Influence Maximization Strategy

Dupuis

Mouza

Travers

et al. 2019

Web Information Systems Engineering – WISE 2019

View full text Add to dashboard Cite

show abstract

“…(Note that ASTI is the version with a batch size of 1.) AdaptIM is modified from the AdaptIM-1 method proposed in [2] [10,33], we set the approximation parameter ε = 0.5 for the five adaptive algorithms. For those parameters in ATEUC, we use the values recommended in [24].…”

Section: Methodsmentioning

confidence: 99%

Efficient adaptive approximation algorithms in online social networks

Huang¹

View full text Add to dashboard Cite

Online social networks (OSNs) have witnessed their prosperous developments in recent years. Based on the established relations among individuals in OSNs, ideas and opinions can be spread via a word-of-mouth effect. To exploit this effect, many companies have taken OSNs as the major advertising channels to promote their products, which has motivated substantial research on viral marketing. Among them, influence maximization, seed minimization, and profit maximization have been extensively studied in the past decades. These three problems are mainly studied in the non-adaptive setting in the literature, i.e., seed user selection phase does not involve any observations on influence propagation in reality. In this thesis, we extend profit maximization into a generalized version as target profit maximization and investigate these three problems in the adaptive setting, formulated as adaptive influence maximization, adaptive seed minimization, and adaptive target profit maximization respectively. Firstly, we study the adaptive influence maximization (IM) problem. Given a social network G and an integer k, the influence maximization problem asks for a seed set S of k nodes from G to maximize the expected number of nodes activated via an influence propagation. In the adaptive setting, the k seed nodes are selected in batches of equal size b, such that the i-th batch is identified after the actual influence results of the former i − 1 batches are observed. We propose the first practical algorithm for the adaptive IM problem that could provide the worst-case approximation guarantee of 1 − e α(ε−1) with high probability, where ε ∈ (0, 1) is a user-specified parameter and α = 1−(1−1/b) b is set by the batch size b. Our approach is based on a novel general framework AdaptGreedy which could be instantiated by any existing non-adaptive IM algorithms with expected approximation guarantee. In this regard, we design a novel non-adaptive IM algorithm called EPIC which not only provides strong expected approximation guarantee, but also i First and foremost, I would like to express my sincerest gratitude towards my supervisors, Prof. Xiaokui Xiao and Prof. Aixin Sun. It is their guidance of patience on my PhD research that makes the accomplishment of this thesis possible. In particular, Prof. Xiao's rigorous attitude, the pursuit of high standards, and unlimited enthusiasm for research have impressed and shaped me deeply. Prof. Sun, who comes from the hometown of Confucius in China, well interprets the charms of being a professor. It is his continuous help and care that support me forward. Their positive influence on me would be the priceless treasure for my future career and life. Then, I would like to thank my collaborators, especially Dr. Jing Tang. We have collaborated for two years. I learn so much from his great intelligence and experience on research such that I could conquer the challenges in my research.

show abstract