Multi‐scale graph capsule with influence attention for information cascades prediction

Chen, Xueqin; Zhang, Fengli; Zhang, Fan; Bonsangue, Marcello M.

doi:10.1002/int.22786

Cited by 30 publications

(22 citation statements)

References 52 publications

(112 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although the risk of data leak is mitigated, the vanilla FL suffers from the heterogeneity of the client data sets [14][15][16][17] and exhibits unfavorable performance on model accuracy and convergence rate. As stated by Chen et al, 11 Kairouz et al, 18 Zhang et al, 19 Wang et al, 20 and Chen et al, 21 (1) the local data sets cannot represent the overall data distribution, and the local distributions may be also different from each other; (2) the number of clients (e.g., mobile phone users) can be very large, and a large portion of clients are often offline or on unreliable connections; and (3) the amount of data on different clients may be highly imbalanced. This great disparity between the client data sets significantly degrades the performance of FL.…”

mentioning

confidence: 85%

CRFL: A novel federated learning scheme of client reputation assessment via local model inversion

Zheng

Huang

2022

Int J of Intelligent Sys

View full text Add to dashboard Cite

Federated learning (FL) is gradually becoming a key learning paradigm in Privacy-preserving Machine Learning (ML) systems. In FL, a large number of clients cooperate with a central server to learn a shared model without sharing their own data sets.However, since there is a great disparity between the client data sets, standard FL is often hard to tune and suffers from performance degradation due to the inharmony among local models. To this end, in this paper we propose a novel FL scheme, termed client reputation federated learning (CRFL), which dynamically assesses the reputation of the clients participating in FL. Our method leverages techniques from model explanation, and aims at precisely measure each client's impact to the global model. To be specific, we first calculate the saliency-weighted variance on pixelwise relevance scores as the quality factor of a single sample. Then we extract activation function values at the last hidden layer to compute the divergence factor of individual data set. Finally, the server integrates these two factors as an assessment of the client reputation. By leveraging such assessment, CRFL can dynamically adjust the weights of the clients in each aggregation round, thus leading to a significant improvement over the baseline method in terms of model accuracy and

show abstract

mentioning

confidence: 85%

CRFL: A novel federated learning scheme of client reputation assessment via local model inversion

Zheng

Huang

2022

Int J of Intelligent Sys

View full text Add to dashboard Cite

show abstract

“…MUCas [97] incorporated GCN, dynamic routing mechanism and AM to learn latent representations of cascade graphs to fully utilize the directional, high-order, positional, and dynamic scales of cascade information. Since the input is only the cascade graph before some specific observation time, MUCas designed a novel sampling method to generate subcascade graphs based on disjoint time intervals.…”

Section: E Gcn and Gat Based Modelsmentioning

confidence: 99%

Graph Representation Learning for Popularity Prediction Problem: A Survey

Chen,

Guo,

2022

Preprint

View full text Add to dashboard Cite

The online social platforms, like Twitter, Facebook, LinkedIn and WeChat, have grown really fast in last decade and have been one of the most effective platforms for people to communicate and share information with each other. Due to the "word of mouth" effects, information usually can spread rapidly on these social media platforms. Therefore, it is important to study the mechanisms driving the information diffusion and quantify the consequence of information spread. A lot of efforts have been focused on this problem to help us better understand and achieve higher performance in viral marketing and advertising. On the other hand, the development of neural networks has blossomed in the last few years, leading to a large number of graph representation learning (GRL) models. Compared with traditional models, GRL methods are often shown to be more effective. In this paper, we present a comprehensive review for recent works leveraging GRL methods for popularity prediction problem, and categorize related literatures into two big classes, according to their mainly used model and techniques: embedding-based methods and deep learning methods. Deep learning method is further classified into six small classes: convolutional neural networks, graph convolutional networks, graph attention networks, graph neural networks, recurrent neural networks, and reinforcement learning. We compare the performance of these different models and discuss their strengths and limitations. Finally, we outline the challenges and future chances for popularity prediction problem.Impact Statement-Social platforms have exploded in popularity in last decade, and information diffusion on social networks has attracted widespread attention in both industrials and academics. Companies need estimate whether marketing strategies will be successful before real implementation to save costs and earn more profits. On the other hand, it is vital to predict diffusion of misinformation before it spreads too far, and block its further diffusion. Therefore, it is important to understand underlying diffusion dynamics of cascade and its popularity trend. With recent achievements of graph representation learning in artificial intelligence and other fields, many works leveraged graph representation learning methods for information diffusion modeling and popularity prediction. This article presents a comprehensive review of recent works using graph representation learning methods for popularity prediction. Reviewing different methods can help researchers identify strengths and limitations of existing models, and find out challenges and directions for popularity prediction in future work.

show abstract

“…With the large-scale construction and application of new-generation digital infrastructures such as 5G, industrial Internet, big data centers, and cloud computing, more and more important information systems will carry core businesses and massive amounts of data that are closely related to national security and economic development. 1,2 More and more researchers have begun to focus on artificial intelligence (AI) technology applications, [3][4][5][6] network infrastructure optimization, 7,8 and software security analysis 9,10 for critical infrastructure. The development and research results of these works also bring some inspiration for the study of the security of critical infrastructure for us.…”

Section: Introductionmentioning

confidence: 99%

A comprehensive study of Mozi botnet

Qin

Zhang

et al. 2022

Int J of Intelligent Sys

View full text Add to dashboard Cite

With the trend of digital transformation of enterprises, the use of Internet of Things (IoT) devices is increasing. IoT devices that are not protected by security measures have gradually become targets of attackers. Attackers use weak passwords and software vulnerabilities in the device to invade the device and control it to become a node of the botnet. The Mozi botnet was discovered in December 2019, and its attention has increased day by day, and its influence once exceeded Mirai. After a preliminary reverse analysis of the Mozi samples, we have continued to track the development and changes of the Mozi botnet since February 2021. First, through the in-depth analysis of the communication principles of the Mozi botnet and the distributed sloppy hash table protocol, we have proposed an in-depth analysis of the Mozi botnet. The active detection method of Mozi, through daily and continuous tracking of the number of Mozi nodes, is infinitely close to the boundary of the Mozi network. On the basis of the collected detection data, we give our conclusions on Mozi's node size, global geographic distribution, 24-hour global activity, equipment composition, and Mozi botnet countermeasures.Through this study, we found that the security of IoT devices around the world is not optimistic, and there is an urgent need to increase the security protection

show abstract

Multi‐scale graph capsule with influence attention for information cascades prediction

Cited by 30 publications

References 52 publications

CRFL: A novel federated learning scheme of client reputation assessment via local model inversion

CRFL: A novel federated learning scheme of client reputation assessment via local model inversion

Graph Representation Learning for Popularity Prediction Problem: A Survey

A comprehensive study of Mozi botnet

Contact Info

Product

Resources

About