Σεμινάριο CEID & Social Hour 18.06.2021@3MM(ZOOM): "Distributed Community Detection Methodologies for Large Scale Social Networks", Πισπιρίγκος Γεώργιος, Υποψήφιος Διδάκτορας, ΤΜΗΥΠ

Εικόνα pintela
Κατηγορία: 

Τίτλος: Distributed Community Detection Methodologies for Large Scale Social Networks

Ομιλητής: Πισπιρίγκος Γεώργιος, Υποψήφιος Διδάκτορας, ΤΜΗΥΠ, ΠΠ

Ημερομηνία-χώρος: Παρασκευή 18  Ιουνίου, 3-5μμ

Zoom link: https://upatras-gr.zoom.us/j/92569423452?pwd=SnZ3T0FmNEs0TCtiNjZXNjl6UVBuZz09

Περίληψη:  Nowadays, the amount of digitally available information has  tremendously grown, with real-world data graphs outreaching the billion  vertices. Focusing on social media, which usage is universally spread,  information networks can be deemed as the predominant data structure due  to their handily compact representation. Hence, graph analysis  techniques have been broadly adopted due to their inherent efficiency in handling hierarchical data. Specifically, one of the most thoroughly studied graph partitioning problems is community detection that aim to form groups of vertices according to a well-defined similarity measure due  toits apparent abstraction. With extensive application in a wide range of scientific sectors -e.g. bio-informatics, sociology, nonlinear dynamics, digital marketing, computer science, etc.;- where the concept of retrieving the subjacent hierarchical structure from complex information networks is a major concern, the significance of community detection is strongly underlined. Despite the impressive amount of research that has yet been published to tackle this NP-hard class problem, the existing algorithms have practically been proven not only inefficient but mainly unscalable in case of contemporary social media graphs, for which polynomial solutions are prohibitive. In this regard, the purpose of this seminar is the presentation of distributed community detection methodologies that are capable of efficiently extracting the subjacent community hierarchy of any given social graph in spite of its size and density.

This presentation is based on the following research publications:

Makris, C.; Pettas, D.; Pispirigos, G. Distributed Community Prediction for Social Graphs Based on Louvain Algorithm. In IFIP International Conference on Artificial Intelligence Applications and Innovations; Springer: Cham, Switzerland, 2019; pp. 500–511. DOI: 10.1007/978-3-030-19823-7_42

Makris, C.; Pispirigos, G.; Rizos, I.O. A Distributed Bagging Ensemble Methodology for Community Prediction in Social Networks. Information 2020, 11, 199. https://doi.org/10.3390/info11040199

Makris, C.; Pispirigos, G. Stacked Community Prediction: A Distributed Stacking-Based Community Extraction Methodology for Large Scale Social Networks. Big Data Cogn. Comput. 2021, 5, 14. https://doi.org/10.3390/bdcc5010014

Konstantinos Georgiou, Christos Makris, Georgios Pispirigos: A Distributed Hybrid Community Detection Methodology for Social Networks. Algorithms 12(8): 175 (2019). https://doi.org/10.3390/a12080175

Eleanna Kafeza, Andreas Kanavos, Christos Makris, Georgios Pispirigos, Pantelis Vikatos: T-PCCE: Twitter Personality based Communicative Communities Extraction System for Big Data. IEEE Trans. Knowl. Data Eng. 32(8): 1625-1638 (2020). DOI: 10.1109/TKDE.2019.2906197

Makris, C.; Pispirigos, G.; Simos, M.A. Text Semantic Annotation: A Distributed Methodology Based on Community Coherence. Algorithms 2020, 13, 160. https://doi.org/10.3390/a13070160

Σχετικά με τον ομιλητή:  Georgios Pispirigos is an experienced big data software engineer, born in Patras, Achaia, Greece in 1986. He received  his Diploma and his master degree in Computer Science and Engineering in 2011 and 2015 respectively, from the department of Computer Engineering and Informatics (CEID), University of Patras, where he currently is a PhD student under Associate Prof. Christos Makris supervision. His research interests span the broad areas of distributed processing, information retrieval, data mining, graph mining and machine learning.

 

Startup Growth Lite is a free theme, contributed to the Drupal Community by More than Themes.