Visual Explanations of Differentiable Greedy Model Predictions on the Influence Maximization Problem
Social networks have become important objects of study in recent years. Social media marketing has, for example, greatly benefited from the vast literature developed in the past two decades. The study of social networks has taken advantage of recent advances in machine learning to process these imme...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-09-01
|
| Series: | Big Data and Cognitive Computing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2504-2289/7/3/149 |
| _version_ | 1797581216869777408 |
|---|---|
| author | Mario Michelessa Christophe Hurter Brian Y. Lim Jamie Ng Suat Ling Bogdan Cautis Carol Anne Hargreaves |
| author_facet | Mario Michelessa Christophe Hurter Brian Y. Lim Jamie Ng Suat Ling Bogdan Cautis Carol Anne Hargreaves |
| author_sort | Mario Michelessa |
| collection | DOAJ |
| description | Social networks have become important objects of study in recent years. Social media marketing has, for example, greatly benefited from the vast literature developed in the past two decades. The study of social networks has taken advantage of recent advances in machine learning to process these immense amounts of data. Automatic emotional labeling of content on social media has, for example, been made possible by the recent progress in natural language processing. In this work, we are interested in the influence maximization problem, which consists of finding the most influential nodes in the social network. The problem is classically carried out using classical performance metrics such as accuracy or recall, which is not the end goal of the influence maximization problem. Our work presents an end-to-end learning model, SGREEDYNN, for the selection of the most influential nodes in a social network, given a history of information diffusion. In addition, this work proposes data visualization techniques to interpret the augmenting performances of our method compared to classical training. The results of this method are confirmed by visualizing the final influence of the selected nodes on network instances with edge bundling techniques. Edge bundling is a visual aggregation technique that makes patterns emerge. It has been shown to be an interesting asset for decision-making. By using edge bundling, we observe that our method chooses more diverse and high-degree nodes compared to the classical training. |
| first_indexed | 2024-03-10T23:02:11Z |
| format | Article |
| id | doaj.art-2d660292843c4928bdbe8f6669195c0e |
| institution | Directory Open Access Journal |
| issn | 2504-2289 |
| language | English |
| last_indexed | 2024-03-10T23:02:11Z |
| publishDate | 2023-09-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Big Data and Cognitive Computing |
| spelling | doaj.art-2d660292843c4928bdbe8f6669195c0e2023-11-19T09:34:25ZengMDPI AGBig Data and Cognitive Computing2504-22892023-09-017314910.3390/bdcc7030149Visual Explanations of Differentiable Greedy Model Predictions on the Influence Maximization ProblemMario Michelessa0Christophe Hurter1Brian Y. Lim2Jamie Ng Suat Ling3Bogdan Cautis4Carol Anne Hargreaves5Department of Statistics and Data Science, Faculty of Science, National University of Singapore, Singapore 117546, SingaporeENAC, Université de Toulouse, 31400 Toulouse, FranceDepartment of Statistics and Data Science, Faculty of Science, National University of Singapore, Singapore 117546, SingaporeInstitute for Infocomm Research, A*STAR, Singapore 138632, SingaporeDepartment of Computer Science, University of Paris-Sud, 91405 Orsay, FranceDepartment of Statistics and Data Science, Faculty of Science, National University of Singapore, Singapore 117546, SingaporeSocial networks have become important objects of study in recent years. Social media marketing has, for example, greatly benefited from the vast literature developed in the past two decades. The study of social networks has taken advantage of recent advances in machine learning to process these immense amounts of data. Automatic emotional labeling of content on social media has, for example, been made possible by the recent progress in natural language processing. In this work, we are interested in the influence maximization problem, which consists of finding the most influential nodes in the social network. The problem is classically carried out using classical performance metrics such as accuracy or recall, which is not the end goal of the influence maximization problem. Our work presents an end-to-end learning model, SGREEDYNN, for the selection of the most influential nodes in a social network, given a history of information diffusion. In addition, this work proposes data visualization techniques to interpret the augmenting performances of our method compared to classical training. The results of this method are confirmed by visualizing the final influence of the selected nodes on network instances with edge bundling techniques. Edge bundling is a visual aggregation technique that makes patterns emerge. It has been shown to be an interesting asset for decision-making. By using edge bundling, we observe that our method chooses more diverse and high-degree nodes compared to the classical training.https://www.mdpi.com/2504-2289/7/3/149influence maximizationend-to-end learningdecision-focused learninggraph visualizationedge bundlingdifferentiable greedy |
| spellingShingle | Mario Michelessa Christophe Hurter Brian Y. Lim Jamie Ng Suat Ling Bogdan Cautis Carol Anne Hargreaves Visual Explanations of Differentiable Greedy Model Predictions on the Influence Maximization Problem Big Data and Cognitive Computing influence maximization end-to-end learning decision-focused learning graph visualization edge bundling differentiable greedy |
| title | Visual Explanations of Differentiable Greedy Model Predictions on the Influence Maximization Problem |
| title_full | Visual Explanations of Differentiable Greedy Model Predictions on the Influence Maximization Problem |
| title_fullStr | Visual Explanations of Differentiable Greedy Model Predictions on the Influence Maximization Problem |
| title_full_unstemmed | Visual Explanations of Differentiable Greedy Model Predictions on the Influence Maximization Problem |
| title_short | Visual Explanations of Differentiable Greedy Model Predictions on the Influence Maximization Problem |
| title_sort | visual explanations of differentiable greedy model predictions on the influence maximization problem |
| topic | influence maximization end-to-end learning decision-focused learning graph visualization edge bundling differentiable greedy |
| url | https://www.mdpi.com/2504-2289/7/3/149 |
| work_keys_str_mv | AT mariomichelessa visualexplanationsofdifferentiablegreedymodelpredictionsontheinfluencemaximizationproblem AT christophehurter visualexplanationsofdifferentiablegreedymodelpredictionsontheinfluencemaximizationproblem AT brianylim visualexplanationsofdifferentiablegreedymodelpredictionsontheinfluencemaximizationproblem AT jamiengsuatling visualexplanationsofdifferentiablegreedymodelpredictionsontheinfluencemaximizationproblem AT bogdancautis visualexplanationsofdifferentiablegreedymodelpredictionsontheinfluencemaximizationproblem AT carolannehargreaves visualexplanationsofdifferentiablegreedymodelpredictionsontheinfluencemaximizationproblem |